AI captured 53% of global VC funding in H1 2025, yet only 5% of pilot AI projects actually reach production and generate real benefits. This gap reveals everything you need to know about the current state of AI adoption.

While headlines promise revolution, the actual transformation is happening in very specific, often boring parts, especially in game development. Some tools have crossed into production-ready territory. Most haven’t. So let's look at AI applications that are currently effective, and what the next year will realistically bring.

With twenty percent of Steam games now disclosing AI usage, here’s what happened over the summer.

In The News

90% of developers now use AI in workflows, per Google Cloud

At Devcom 2025, Google announced that 90% of game developers use some form of AI, and 97% believe it’s transforming the industry. Applications range from automating repetitive tasks to accelerating playtesting and designing dynamic levels. The gap between “use some form of AI” and “successfully implemented AI” remains wide.

NetEase posts record Q2 revenue with heavy AI integration

NetEase’s Q2 results were driven by titles like MARVEL Mystic Mayhem and FragPunk, with AI used for development via their LLM Confucius and ad optimization. The company sees AI-designed games as a core strategy. It’s one of the first major publishers to publicly tie financial performance directly to AI implementation.

Steam: Nearly 1 in 5 new games use generative AI

Steam disclosed approximately 20% of 2025 releases used generative AI for content or character generation, an exponential increase from earlier in the year. Actual usage may be higher, as not all developers disclose their tools.

Unity beats Q2 expectations with AI-powered ad platform

Unity’s Q2 performance exceeded expectations largely due to its AI-powered advertising platform. The platform uses AI for targeting optimization and creative generation, showing that some of the clearest AI wins come from adjacent services rather than core production.

Inworld AI launches first AI runtime for consumer applications

The first AI runtime engineered to scale consumer applications from 10 to 10M users with minimal code changes. The platform automates AI operations and addresses the critical gap where most projects stall between prototype and production. Early results: Wishroll scaled to 1 million users in 19 days with 95% cost reduction, while Bible Chat reduced voice costs by 85%. The company has raised over $120M and works with Xbox, Disney, NVIDIA, and NBCUniversal. Pricing is usage-based with no upfront costs.

Ludus AI launches automated Blueprint generation for Unreal Engine

Open beta began in August 2025 with the first system to automatically generate Blueprints in Unreal Engine, reducing creation time from days to minutes.

The platform has hit 10,000 generated 3D models and analyzed 10,000 Blueprints across more than 20,000 users, positioning itself as a comprehensive Unreal Engine assistant. It handles Blueprint creation and analysis, 3D model generation, code writing and debugging, and project documentation. The tool offers self-hosted deployment for complete privacy and data control.

Q3 brings a landmark spike in gaming M&A

Drake Star’s Q3 2025 report revealed record M&A activity led by the largest leveraged buyout in gaming history: a Saudi PIF, Silver Lake, and Affinity Partners consortium acquired Electronic Arts for $55 billion. Overall gaming M&A reached 50 deals, the highest in four quarters. Private financing rebounded for the first time in six quarters with 115 rounds, led by Lingokids ($120M) and Good Job Games ($60M). Western public gaming equities surged, with the Drake Star Western Gaming Index rising 31.3% year-to-date, led by Roblox (+136.9%), Unity (+77.9%), and Everplay (+75.0%). Drake Star expects AI and tools sectors to remain hot, with PE momentum showing no signs of slowing.

Investments

AI companies in gaming and developer tooling captured significant investor attention in Q2 and Q3 2025. Notable funding rounds during this period include:

The Latest Research

lmgame-Bench: How Good are LLMs at Playing Games? introduces a benchmark suite for evaluating LLMs across platformer, puzzle, and narrative games through a unified API. Results showed RL training on a single game transferred to unseen games and planning tasks, though LLMs still struggle with vision perception and prompt sensitivity in complex environments.

Agents Play Thousands of 3D Video Games presents PORTAL, a framework where AI agents learn to play thousands of 3D video games using language-guided, embodied interaction. Agents demonstrate generalizable skills and quick learning across games, though bottlenecks remain in environments requiring social collaboration and emergent complex tasks.

AXIOM: Learning to Play Games in Minutes achieves seven times faster learning than leading DeepMind models by focusing on object-centric modeling rather than pixel memorization. It outperforms larger models with just 0.95M parameters versus 420M, showing high adaptability in the Gameworld 10k benchmark. Limitations appear in games demanding intricate narrative logic or hidden information.

Narrative Co-Creation frameworks explore how AI models collaborate with humans to co-create gaming narratives, with new assessment methods for creativity, quality, and engagement of AI-generated stories. Multiple papers introduce benchmarking paradigms for strategic dialogue, interactive reasoning, and embodied gameplay.

General Game Playing advances demonstrate new agent frameworks playing wide ranges of games using code-based world modeling, highlighting progress in generalization beyond fixed rule-based systems. There’s a push to leverage RL with LLMs, combining declarative and procedural knowledge for improved reasoning in multi-step, interactive challenges.

Playstyle and Artificial Intelligence examines how AI can model, adapt to, and shape player playstyles by analyzing in-game behaviors. The framework proposes adaptive AI that tailors difficulty and narrative to individual preferences, though accurately capturing nuanced player intent and ensuring user satisfaction remain challenges.

Impressive progress, but it also reveals a critical bottleneck: LLMs excel at language but struggle with spatial reasoning and physical interaction. This limitation is driving the industry’s pivot to an entirely different paradigm.

The Hottest Trends and Tools

World models: From prediction to spatial understanding

some researchers believe the real breakthrough won’t come from better language models. It’ll come from AI that learns by watching, not reading.

World models represent a fundamental shift from predicting the next word to understanding physical space. And here’s the interesting part: unlike the LLM race where OpenAI and Anthropic are miles ahead, world models are young enough that scrappy, well-funded startups actually have a shot.

The tech is moving fast. Google DeepMind’s Genie 3 (August 2025) can generate entire 3D worlds from text prompts at 24 FPS in 720p, keeping everything visually consistent for up to a minute with realistic physics.

Meta’s V-JEPA 2 (June 2025) showed something remarkable. After watching 1M+ hours of video and just 62 hours of actual robot footage, it could handle new manipulation tasks with 65-80% success rates and run 30x faster than NVIDIA’s Cosmos. Tencent’s Yan took things further with infinite simulation length and the ability to turn any image into an interactive video.

The efficiency gains are getting crazy. Lucid v1 figured out how to compress each video frame down to just 15 tokens instead of 500+. That’s a 600x reduction. Real-time AI-generated worlds running at 25 FPS on a consumer gaming GPU, no datacenter required.

Other developments include Wayve’s GAIA-2 for autonomous driving and NVIDIA’s Cosmos, which has been downloaded over 2M times since January 2025 and deployed by Uber for AV data curation, Amazon for manufacturing robotics, and Boston Dynamics for simulation.

The technology has attracted significant capital. World Labs raised $230M at a $1B+ valuation, while Decart reached a $3.1B valuation, demonstrating real-time world generation with models like Oasis.

xAI, Elon Musk’s AI venture, is also entering the space by hiring Nvidia experts like Zeeshan Patel and Ethan He to accelerate development of advanced world models capable of designing and navigating physical spaces.

This aligns with Musk’s announcement that xAI’s game studio plans to release a “great AI-generated game” before the end of 2026, potentially leveraging these models for dynamic, AI-driven video game generation.

Local LLMs moving from experiment to production

World models promise the future, but they require massive compute and cloud infrastructure. Meanwhile, a quieter revolution is happening on-device: local LLMs are already solving practical problems for developers today, no datacenter required.

Take Echoes of Ir, which showcased companions powered by local LLMs, complete with memory-like behavior and betrayal mechanics, running entirely on-device without any cloud latency.

Or Verbal Verdict, which shipped on Steam with a local LLM stack driving its interrogation-based gameplay. It’s one of the first commercial titles to prove that real-time, on-device dialogue works in production. Then there’s Adam Lyttle, who built a music learning app using Apple Intelligence that handles real-time note monitoring, adjusts difficulty on the fly, and announces progress, all running locally on the device.

Behind the scenes, developers are increasingly turning to tools like Ollama and LocalAI to integrate models such as Gemma and Llama into Unity and Godot. The result? Offline NPC systems that work without an internet connection. At GDC 2025, sessions on co-creative storytelling highlighted these architectures. The hardware barrier is falling fast. Smaller 8B models now run smoothly on consumer GPUs with just 12-24GB of VRAM.

The next frontier is voice. NeuTTS Air, released in October 2025, brings production-ready text-to-speech to the edge with its 748-million-parameter model running in just 400-600MB of RAM.

Built on a Qwen 0.5B backbone and compressed to around 500MB in quantized format, it achieves real-time synthesis on mid-range CPUs. No GPU required. The model clones voices from 3-15 seconds of reference audio with 200ms latency and 85-95% speaker similarity, making dynamic voiceover for procedural dialogue or adaptive tutorials suddenly viable offline. Open-sourced under Apache 2.0, it’s already being deployed on devices ranging from laptops to Raspberry Pi boards.

I think this shift to edge AI addresses three problems cloud-based solutions simply can’t solve. First, there’s zero latency for real-time interactions like NPC dialogue or adaptive difficulty. Second, there are no recurring API costs that balloon with your player count. And third, you get complete data privacy by design.

As a result, novel interaction patterns are emerging. Vibe-coded games where AI responds instantaneously to player whims, or reinforcement learning-enhanced NPCs that evolve without network delays. That kind of thing wasn’t really possible before.

From runtime to production: AI transforming asset creation

Tools like Tencent’s Hunyuan-Game are transforming how games are built. Launched in June 2025, it renders sketches in roughly one second and three-view diagrams in minutes instead of 12+ hours. Production speeds are up around 300%.

Trained on millions of gaming datasets, it handles mobile environments, characters, and UI with impressive precision. Real-time sketching instantly visualizes ideas. Automated character views give you 360° previews ready for rigging. The workflow goes straight from inspiration to design. By September 2025, version 2.0 added image-to-video generation, custom LoRA training for style consistency, and one-click refinement tools.

Hunyuan3D Studio goes further: text-to-3D generation, one-minute UV unwrapping, PBR editing, and auto-rigging. What used to take days now takes minutes.

The open-sourced Hunyuan-GameCraft is where it gets really interesting. It creates interactive video from single scenes with user-controlled camera movements and environmental effects like rain or clouds, all running on an RTX 4090.

With photorealistic textures and direct exports to Unity, Unreal, and Blender, high-quality asset creation is actually accessible now.

Vibe coding proves itself in unexpected places

The NHS Clinical Game Jam in September 2025 showed just how transformative these tools can be beyond traditional game studios. Healthcare professionals with zero game development experience used Rosebud’s AI tools to prototype playable games in a single weekend, including a winning mental health garden game that teaches emotional regulation through play.

These clinicians went from concept to prototype without touching traditional code. They focused entirely on design and therapeutic purpose instead of wrestling with toolchains.

Gambo.ai takes a similar approach. Users can build complete, playable games from a single text prompt in minutes. It automatically generates everything: characters, maps, animations, music, sound effects, even built-in ads for day one monetization.

This no-code workflow eliminates the usual barriers like scripting or manual asset organization. Creators focus on ideas rather than technical execution. Some users have recreated classics like Doodle Jump or built complex demos inspired by games like Silksong in hours or days, with AI handling everything from 2D/3D visuals to interactive elements.

I think this is a glimpse of where creation is headed. Not just faster workflows for existing developers, but genuinely new creators entering the space who would’ve been locked out before. The barrier to entry is collapsing.

Conclusion

AI in game development is past the “wait and see” phase. Studios experimenting now are finding real uses. Those sitting on the sidelines are starting to fall behind. Mobile leads adoption because players care less about AI usage and optimization matters more. PC and console developers have to be more careful. Their audiences are way more sensitive to AI. Some of the hype is overblown. But the technology solves real problems, especially around exploding budgets and production risk.

Small teams can suddenly punch above their weight. The 8,000+ games on Steam already using AI show it’s working when it’s used to help creators, not replace them. Meanwhile, bigger publishers often struggle with internal politics around adoption. That might actually give indies an edge.

Like every tech wave before, AI will unlock entirely new types of games and experiences we haven’t imagined yet. World models and related tech will fundamentally change how we think about building and distributing games altogether.

Every studio should be running small experiments now. Prototype something. Build internal expertise. Figure out what works for your audience. The biggest opportunities are the ones we can’t quite see yet. You want to be ready when they come into focus.

Disclaimers:

This is not an offering. This is not financial advice. Always do your own research. This is not a recommendation to invest in any asset or security.

Past performance is not a guarantee of future performance. Investing in digital assets is risky and you have the potential to lose all of your investment.

Our discussion may include predictions, estimates or other information that might be considered forward-looking. While these forward-looking statements represent our current judgment on what the future holds, they are subject to risks and uncertainties that could cause actual results to differ materially. You are cautioned not to place undue reliance on these forward-looking statements, which reflect our opinions only as of the date of this presentation. Please keep in mind that we are not obligating ourselves to revise or publicly release the results of any revision to these forward-looking statements in light of new information or future events.

Every major leap in computing has been defined by its interface: mainframes had punch cards, PCs had keyboards, smartphones gave us touchscreens.

AI now demands its own.

Typing prompts into a chat box was a start, but it isn’t natural. The real breakthrough comes when AI can see what you see, hear what you hear, speak to you, and show you what you need. Instantly, hands-free, and all day long.

That’s what smart glasses unlock.

Why Smart Glasses, Why Now

For the last decade, smart glasses have been “the technology of the future.” Always close, but not quite ready. The bottlenecks were obvious: clunky displays, heavy frames, poor battery life, and no killer application.

Those hurdles are now breaking.

• Waveguides and microLEDs have made displays clear and compact.

• Weight has dropped below 40 grams.

• Battery life exceeds 16 hours.

• And most importantly, AI is the killer app.

Unlike VR headsets, which immerse us in virtual worlds, smart glasses augment the one we already live in. They are not for escape. They are for enhancement.

For AI to become a true everyday assistant, it needs an interface that is ambient, persistent, and seamlessly woven into human perception. Smart glasses are the first computing platform built for that reality.

The Problem Holding the Market Back

If the hardware is ready, why hasn’t the market taken off?

Because the software layer has been missing.

Every device ships with its own software, limitations, and distribution rules. Developers are shut out or forced to choose between fragmented platforms with shallow app libraries and no guarantee of longevity. The result is no real ecosystem and limited utility for users.

It is the same problem mobile phones faced before Android. In the early 2000s, developers had to pick between Symbian, BlackBerry, Windows Mobile, or Palm, each with its own SDK and distribution rules. No single platform gave them scale. Android solved this by providing an open source core, a unified SDK, and a central app store. That combination let developers build once and reach dozens of devices, which unlocked the mobile boom.

Smart glasses are at the same crossroads today. Rokid, Even Realities, Vuzix, Xreal and others each ship with their own software and fragmented ecosystems. Without a unifying layer, developers face the same dead end that mobile faced pre-Android. Mentra is building the Android moment for smart glasses - and fittingly, Rich Miner, the founder of Android, shares that conviction as one of Mentra’s early backers.

The Solution: MentraOS

MentraOS is the first open, developer-first operating system and app store for smart glasses. It allows developers to write once and deploy everywhere, from Mentra’s own glasses to devices like Even Realities and Vuzix, with more to come.

This removes the fragmentation that has held the category back and gives developers a stable target. MentraOS is open source, licensed under MIT, and ships with real, working apps already in daily use: live captions, translation, proactive AI, and more. Deaf and hard-of-hearing users are logging over eight hours a day.

Mentra is also accelerating adoption with its own reference hardware, Mentra Live. These are camera-first smart glasses that include a microphone, speaker, and full MentraOS SDK for developers. Priced at $249, they give creators and enterprises an open alternative to closed camera-glasses. Shipping is slated for fall 2025, and the current weight is 44 grams.

If Android unlocked the smartphone boom, MentraOS is positioned to do the same for smart glasses.

Why This Team

Cayden Pierce and Alexander Israelov bring a rare combination of grassroots innovation and hard operational experience. Cayden’s track record spans early collaboration with Steve Mann, recognized as the inventor of smart glasses, hands-on prototyping with open-source wearables, and research at MIT Media Lab on proactive, context-aware AI agents designed for real-time augmentation. Like Palmer Luckey, who launched Oculus out of a beat-up trailer, Cayden’s own early chapter was shaped by turning a 40-year-old RV into a roaming hacker lab, criss-crossing North America to build and showcase open-source wearables.

What makes Mentra distinct is not just technical vision but execution capacity. The company operates from San Francisco and Shenzhen, embedding itself directly in the two global capitals of software and hardware. Cayden & Nicolo taught themselves Chinese to work closely with component suppliers and factories, giving them visibility and access that most Western startups lack. This dual presence reduces dependency on third-party contractors, shortens iteration cycles, and improves supply-chain security at a time when other smart glasses teams struggle to move beyond prototypes.

Most competitors are hacker-led, strong on ideas but thin on manufacturing depth and operational reach. Mentra combines that hacker culture with direct hardware integration and global supply-chain positioning. This combination enables them to accelerate hardware distribution, secure scarce components, and build at scale rather than getting stuck in pilot mode. At the same time, the founders believe lasting scale will come from building an open ecosystem and empowering developers, rather than trying to lock down the platform and control every application themselves.

The Round

To further that mission Hartmann Capital invested in Mentra in March, joining its oversubscribed $8 million seed round alongside Paul Graham and Y Combinator, the founders of Android, YouTube, and Pebble, Amazon’s Alexa Fund, Toyota Ventures, and TIRTA Ventures.

Backing Mentra is a bet that open ecosystems will once again outcompete closed ones, just as Android did in the 2000s.

Our Belief

The history of computing is a history of new interfaces. Each generation has shifted how humans interact with technology - from the keyboard, to the touchscreen, to voice.

AI is too powerful to be confined to the moments when we are behind a computer or looking at a phone. Its true potential is unlocked when it can operate in real time, enhancing and interacting with the world as we move through it.

Smart glasses make that possible. They are the form factor that turns AI from a tool we consult into a constant companion embedded in daily life.

Mentra is building the platform to make that shift real - and in doing so, to define the interface of the AI era.

If you’d like to learn more about our work at the frontier of AI and human augmentation, reach out at invest@hartmanncapital.com

Disclaimers:

This is not an offering and does not constitute financial advice.

Our discussion may include predictions, estimates or other information that might be considered forward-looking. While these forward-looking statements represent our current judgment on what the future holds, they are subject to risks and uncertainties that could cause actual results to differ materially. You are cautioned not to place undue reliance on these forward-looking statements, which reflect our opinions only as of the date of this presentation. Please keep in mind that we are not obligating ourselves to revise or publicly release the results of any revision to these forward-looking statements in light of new information or future events.

This sixth installment continues a thesis I began two years ago, when the groundwork for the next major cycle was quietly being laid, yet invisible to most.

Since then, we’ve correctly anticipated the ETF approval, Trump’s pro-crypto presidency, the rate cut pivot, renewed crypto M&A, the stablecoin trade, and the Q1–Q2 narrative vacuum that shook out tourists before the real move began.

On April 9th, during our bi-weekly Frontier Forum, we flagged “Trump’s Liberation Day” as a critical inflection point—calling the market bottom at peak panic.

This isn’t about chest-thumping. It’s about context. This will likely be my only write-up for the next 3–6 months. And unlike the past few, this one carries urgency. I believe we’re entering the phase where positioning matters most—where the asymmetry is highest and attention is still misallocated.

Most will read this in hindsight, when the headlines have caught up. But if you’ve been waiting for a signal—this is it.

It’s time to be bullish.

Near-Term Catalysts - Signals in the Static

There are reasons I’m more bullish on blockchain technology today than perhaps at any point in 2022-24. But before diving into the longer-term, more philosophical drivers behind this cycle, we need to examine the source of immediate urgency behind this piece.

Digital asset investors have a tendency to misread inflection points. They’re cautious when the market has stabilized and is preparing to accelerate, yet euphoric when it’s already gone vertical. Today, we are in the former—consolidated, coiled, and underpriced.

Just look at the total crypto market cap excluding Bitcoin. After five years of chop, it's quietly grinding toward all-time highs. BTC, when priced in gold, is nearing a five-year peak. And while $120K may sound high to bystanders, that’s less than a 2x from 2021 levels. ETH is still down over 30% from its ATH. A recovery to the $3Ks is nothing to write home about. These are not “late-stage” prices—they’re forming the base for the next major leg up.

The winning Layer 1s will be trillion-dollar assets. Today, Ethereum sits just above $400B. Solana, the next in line, barely broke $100B. Meanwhile, Mira Murati just raised a $12B Seed Round—pre-product, pre-revenue. For once, AI is the bubble, not crypto.

So who and what will drive prices higher? Let’s dig in:

ICO 2.0 – Wall Street Style: The Rise of Treasury Companies

Michael Saylor’s “Bitcoin treasury company” playbook is gradually becoming a template.

Saylor’s pivot was born out of a dilemma: his business was stable, but stagnant, and tech giants were poaching his best talent. Sitting on $500M in idle cash as the Fed slashed rates to zero, he made a call—convert the entire balance sheet to Bitcoin. What followed was the most aggressive balance sheet expansion in corporate history: a $1B software company morphing into a $126B Bitcoin vault, now holding ~3% of BTC’s supply. (For context, his recent interview with Jordan Peterson is worth a listen if you want to understand Saylor’s life story and philosophy.)

In December, I wrote:

“MicroStrategy may one day trigger the next bear market, but first, it could become the largest company in the world… Saylor will be put on par with Musk, Zuckerberg, and Huang… I believe an unraveling is inevitable, but for now, he’s sitting on tens of billions in dry powder, likely enough to push Bitcoin past $100k and unlock the next $100 billion in firepower…at least a dozen publicly traded companies are likely to adopt Saylor’s model in the next 24 months. Momentum begets momentum.”

We’re living Act II of that thesis.

But I underestimated how quickly others would follow. We're not talking about a dozen copycats in two years—we're already there.

Since then, we’ve seen high-profile launches like:

• XXI, led by Jack Mallers

• ProCap, founded by Anthony Pompliano

• Bitcoin Standard Treasury, helmed by Adam Back

More importantly, the playbook is expanding beyond Bitcoin:

• BitMine, led by Tom Lee, is an ETH treasury vehicle

• SBET, run by ETH co-founder Joe Lubin, holds ~$1B in ETH—making it the largest ETH treasury company to date

And it's not stopping there. Expect vehicles for Solana, Hyperliquid, XRP, and others. Equity-wrapped token exposure is a regulatory arbitrage—giving Wall Street the volatility it craves while satisfying restrictive mandates.

Traditional tech firms are also continuing to adopt Bitcoin as a treasury asset. Figma, in its S-1, disclosed $70M in BTC holdings and plans to acquire $30M more. Tesla, too, holds over $1bn worth of BTC.

We’re in the fourth inning of the treasury company era. Momentum is building, but for now, it’s mostly been led by high-profile insiders and figureheads. Only a handful of forward-thinking tech firms have followed suit by adding BTC to their balance sheets.

That will change fast.

Within 12 months, I expect treasury wrappers to emerge for nearly every asset in the top 10–20, alongside thematic vehicles—DeFi, DePIN, you name it. Just as ETFs multiplied once the first Bitcoin fund broke through, this structure will be replicated across the board.

More importantly, the number of publicly traded companies holding Bitcoin will likely 10x—from ~100 today to over 1,000 within two years. In an environment where artificially low interest rates no longer offset dollar devaluation, Bitcoin will increasingly be seen as prudent diversification, not speculation.

And once these flows begin, they don't reverse easily.

These entities are built to buy and hold. Selling breaks the mandate. Which means flows will be one-directional—funded by convertible debt, ATM offerings, or internal capital—but rarely unwound.

Outflows only happen in three scenarios:

• The company defaults on its notes

• Shares trade below NAV, and management chooses to unwind (unlikely for the large players)

• Shareholder activism forces in-kind distributions or asset sales

Until then, these vehicles function as black holes—soaking up scarce assets, never letting them out. The longer it continues, the tighter the float becomes. At some point, that ends in a supply squeeze.

This might be one of the easiest times in history to be long crypto.

And yet, most investors are still dramatically underweight. Why? Because they think this is all just a reflexive Ponzi—treasury wrappers, MSTR mania, too much froth.

But the real story is much bigger than Bitcoin ETFs or treasury companies. It’s the quiet rewiring of the global financial system.

Tokenize Everything - The World Goes On-Chain

It’s easy to dismiss the treasury company meta as financial theater—Bitcoin-maxi cosplay with balance sheet leverage. But that cynicism misses the deeper truth: blockchain is finally being used for what it was built to do—upgrade the world’s financial infrastructure.

The real opportunity is tokenization. Not just of crypto-native assets—but of everything: cash, stocks, bonds, treasuries, real estate, art. All of it.

I’ve said this for years, but now the world’s most powerful asset managers are echoing the same thesis. BlackRock CEO Larry Fink put it plainly in his 2025 Chairman’s Letter:

Every stock, every bond, every fund—every asset—can be tokenized. If they are, it will revolutionize investing. Markets wouldn't need to close. Transactions that currently take days would clear in seconds. And billions of dollars currently immobilized by settlement delays could be reinvested immediately back into the economy, generating more growth.

Perhaps most importantly, tokenization makes investing much more democratic.

- Larry Fink, CEO of BlackRock in Annual Chairman Letter

And he’s not alone. SEC Commissioner Paul Atkins put it even more bluntly:

“If it can be tokenized, it will be tokenized.”

- SEC Chairman Paul Atkins

We’re not talking about a fringe use case. We’re talking about tens if not hundreds of trillions of dollars in value migrating on-chain. Every tokenized asset will generate on-chain activity—and that activity flows directly into protocol revenue:

• Each transaction = fees to the underlying blockchain

• Lending tokenized assets = revenue for protocols like Aave, Maple

• Trading tokenized assets = revenue for exchanges like Uniswap, Aero, Raydium

As tokenized assets flow through the system, value accrues at every stage—settlement, lending, trading.

Just step back. The stakes are bigger than you think.

• Bitcoin may 10x from here—reaching parity with gold

• The winning Layer 1s may 50–100x—reaching parity with the GDPs of nations.

• The top DeFi protocols? They may 250x—surpassing the likes of JPMorgan

Crypto’s addressable market just expanded by orders of magnitude. The narrative is no longer about whether blockchains have utility—it’s about how much value they can settle, store, and monetize. The shift becomes even more compelling as we move beyond wrapping legacy equities and begin seeing startups launch as tokenized entities from day one. Public digital asset markets would then quickly become the new venue for discovering emerging talent—with liquidity built in from inception.

A narrow set of assets powers this entire stack—and they now stand among the most asymmetric risk-reward opportunities in global public markets.

All Roads Lead to Devaluation – Rates, Debt, and Failed Austerity

Bitcoiners have long warned of dollar debasement and fiscal irresponsibility enabled by fiat-driven monetary policy. Now, the debt bubble they pointed to has finally entered mainstream discourse—yet there’s still no appetite to do much about it.

How do you solve a debt crisis? Step one: end the deficit. And how do you end the deficit?

A. Austerity: spend less
B. Raise revenues: tax more (through taxes or tariffs)
C. Default or restructure debt
D. Print money and monetize the debt (devalue the currency)

The Trump administration came in guns blazing, promising to confront the problem head-on. DOGE, under Elon Musk’s guidance, was meant to deliver $1 trillion in annual savings via austerity repackaged as “government efficiency.” Howard Lutnick was tasked with generating another $1 trillion annually via tariffs.

Fast forward six months: DOGE has saved just $190 billion (less than $90 billion itemized), and Elon has exited in a highly public, explosive fashion. If Elon can’t succeed with austerity—backed by a public mandate and political cover—then no one will. That takes the first lever off the table.

Meanwhile, the "Big Beautiful Bill" has added an estimated $2.4 trillion to the deficit over the next decade, while also extending tax cuts, further reducing revenue. Lever two looks equally untouchable under a Republican administration. (As a capitalist, I detest taxes as much as the next person—but when the nation is speeding toward a fiscal wall, this isn’t the time to cut revenue.) Tariffs have doubled under Trump, but still only brought in around $100 billion in seven months.

It's too little, too late.

We're not going to save our way out. We're not going to earn our way out—not unless the economy enters a sustained boom that radically expands the taxable base. And as the global reserve currency, default is unlikely. Trust in that status is worth more than suffering moderate inflation via debt monetization.

It doesn’t take a rocket scientist to see where this leads (no Elon pun intended): quantitative easing and dollar devaluation. Lowering interest rates would reduce the government’s annual interest expense—we’re already paying close to $1 trillion just to service the interest on the debt. Looser policy could also stimulate growth, broadening the taxable base without raising tax rates.

But in this environment, cutting rates won’t attract sufficient demand for Treasuries—especially with inflation expectations still elevated. To keep yields from spiking, the Fed would need to step in as a major buyer, effectively monetizing the debt. That dynamic puts downward pressure on real yields, stokes further inflation, and undermines confidence in long-duration U.S. debt. If sustained, it risks accelerating the erosion of dollar reserve status.

This is exactly the fault line between Fed Chair Jerome Powell and President Trump. Powell—who famously underestimated inflation as transitory in 2020—has since evolved into one of the few true statesmen in American public life. In a political landscape obsessed with optics and short-term wins, he’s governed with principle and long-term thinking, even in the face of likely dismissal. He remains committed to the Fed’s dual mandate, knowing full well that the consequences of monetary irresponsibility may not show up until after he’s gone. Yet he continues to do what is unpopular, but likely right.

Trump, by contrast, approaches the world like a founder: win today, dominate the moment, and deal with tomorrow once you’re holding better cards. He’s pushing for a full 300 basis point cut—down to 1%—a crisis-era move last seen during COVID, and one that has historically sent both markets and the economy into overdrive.

It might work. If so, it will go down as a defiant reset of American dominance. Or it might mark the beginning of the end, with the U.S. fast-tracking itself out of reserve currency status and exposing critical social and defense systems due to fiscal collapse.

Either way, the conclusion is the same: in 2–3 years, few rational actors will want to hold bonds or cash. The rotation into risk assets—especially the foundational infrastructure of the new financial era—is already underway.

In my view, 2025 is the year to be net short USD—particularly when so many assets outside the Magnificent 7 and Bitcoin are still trading below prior all-time highs.

The Machine Economy - Scarcity in the Age of Abundance

When I’m not wearing the hat of a hedge fund manager, I wear that of a venture capitalist. Over the past three years, I’ve backed more than 30 startups across AI, XR, and frontier tech. And every now and then, a belief begins as fringe—tinfoil even—but over time, it hardens into conviction. This is one of them.

Let’s run a thought experiment.

An entrepreneur starts with a team of five. With the right AI tools, that team gains the output of ten. Productivity doubles. The company begins to outcompete peers, expand margins, and grow. When it's time to build a marketing department, the founder instead deploys an AI agent with a clear mandate and KPI. The agent manages a human team, driving strategy and delegation. The humans execute, supported by additional AI tools.

As the structure scales, the founder introduces more agents—specialized ones for advertising, copywriting, outreach. Over time, the company operates with an increasingly autonomous hierarchy. Eventually, the founder asks the obvious question: why not install a CEO agent? Why not build a board of agents tasked with tuning its mandate?

Eventually, the founder steps away altogether—still the beneficiary, but no longer involved.

Meanwhile, the cost of goods and services globally has collapsed. Supply is abundant. Competition is brutal. Efficiency is the only moat—and it's being optimized into oblivion.

In that world of infinite, hyper-personalized, low-cost everything—what retains value?

You might say things. But we’re already engineering superior materials in labs. Lab-grown meat is real (for better or worse). Lab-grown diamonds are indistinguishable. You might say gold—but 2025 marks the first year asteroid mining became commercially viable. Scarcity is melting.

You might say energy. But fusion breakthroughs and microreactors suggest even that’s on its way to becoming abundant.

And so, paradoxically, in a world of boundless digital abundance, the most valuable asset may be the one with hard coded scarcity: Bitcoin. A finite denominator in a world where everything else is trending toward infinite. Most analysts get it backwards. They fixate on demand—who’s the next buyer—when the real story is supply. Bitcoin doesn’t need demand to grow in value. It needs time. Everything else inflates. Bitcoin doesn’t.

Absent a superintelligence that runs this experiment to its extreme—where all goods and services are delivered by a single entity, a dystopian outcome that could be checked by antitrust crackdowns or competition amongst ASI—we’re far more likely to see millions of autonomous, non-human agents participating in the global economy on our behalf. These agents will need to transact value, access leverage, exchange assets, and settle payments—instantly, securely, and without centralized gatekeepers.

The only infrastructure capable of supporting that at scale today is public blockchains.

As economic velocity increases, the largest beneficiaries won’t be the applications or interfaces—it will be the networks themselves. The base layers and rails that can handle agent-to-agent economic coordination in a globally secure, trustless way. In a world where businesses are run by agents, execution will be pushed to its limits. No traditional system—wrapped in bureaucracy and compliance—will be able to keep up. Banks defend themselves with paperwork. Blockchains defend themselves with code.

And code scales.

Blockchains are the only infrastructure built to meet the demands of an agent-driven economy.

Once that becomes clear, assets like Bitcoin—and leading Layer 1 blockchains like Ethereum, Solana, and Hyperliquid—alongside decade-tested DeFi protocols like Aave and Uniswap—start to look absurdly cheap.

You may not see it today. But I suspect you’ll come back to this write-up a few years from now and realize how early we were. None of this negates the volatility along the way. But I’ll repeat the same advice I have given myself over the years: stop chasing waves—study the tide. From what I can see, the flood is near.

In 2019, I thought I was late—that the real upside in crypto had already passed. In 2025, I realize everything until now was just the prelude to what’s coming.

Disclaimers:

This is not an offering. This is not financial advice. Always do your own research.

Our discussion may include predictions, estimates or other information that might be considered forward-looking. While these forward-looking statements represent our current judgment on what the future holds, they are subject to risks and uncertainties that could cause actual results to differ materially. You are cautioned not to place undue reliance on these forward-looking statements, which reflect our opinions only as of the date of this presentation. Please keep in mind that we are not obligating ourselves to revise or publicly release the results of any revision to these forward-looking statements in light of new information or future events.

Agents Are Changing the Rules of Business

The first wave of AI was about creation: writing blog posts, generating images, drafting emails. Now we're witnessing something fundamentally different: AI that doesn't just create, it acts. These aren't content generators anymore; they're autonomous agents completing real tasks and achieving actual objectives.

AI tasks exploded 760% in two years on Zapier's platform alone, the fastest growth they've ever recorded. Over 200 million AI tasks were completed each month across 8,000+ connected apps. But this isn't just impressive growth. It signals a fundamental shift in how work gets done.

Despite these powerful capabilities, most firms continue to operate AI in silos. Marketing teams prompt Claude, sales teams analyze calls with Gong, and support uses ChatGPT. While each tool provides unique value, this dispersed approach leaves workflows unconnected and potentially unmet. The real breakthrough happens when organizations stop treating AI as scattered tools and start building integrated agent ecosystems.

When DoorDash deployed its first AI agents, they scaled from zero to 35,000+ automated calls in six weeks with 94% accuracy. According to IBM, 99% of enterprise developers create AI agents rather than AI tools. What's the difference? These are not just simple content creators. Instead, they are autonomous task executors. Previously, AI tools required constant human oversight. Modern agents work independently.

The momentum for this transition is evident. AI agent companies raised $3.8 billion in 2024, nearly double the figure in 2023, as every major technology company began creating agent platforms and tooling. Workforce composition will shift toward hybrid human-AI teams, and operational efficiency will skyrocket as mundane jobs are entirely automated.

Here's why it's happening now, how the best companies are capturing value, and what it implies for the future of work.

The Technical Leap

stems from advances in reasoning models. AI that can think through problems step by step. Poe's data tells the story: reasoning model usage jumped from 2% to 10% of all messages in early 2025.

Reasoning models can plan, execute, and adjust in real time, a fundamental shift from previous AI that relied on pattern matching. And this capability leap required new infrastructure.

Three foundational protocols emerged: Anthropic's MCP connects agents to tools and data sources, Google's A2A enables agent-to-agent collaboration, and AG-UI facilitates user interface interactions. Rather than competing, these protocols complement each other, with Microsoft's Azure AI Foundry Agent Service now unifying these standards with built-in observability.

The Emerging Ecosystem

An entirely new category of business software was unlocked. Unlike rigid automation that breaks when interfaces change, AI agents use reasoning and context to navigate PDFs, spreadsheets, and legacy systems without pre-programmed pathways.

The result? Specialized agent platforms emerged across every business function, driving sector value from $5.4 billion to $7.6 billion in just one year.

However, the deeper trend is mass democratization; domain experts can now build complex agents through no-code platforms, and no developers are required. The enterprise numbers prove this isn't just hype. Salesforce reported over 200 Agentforce deals weekly, with thousands in the pipeline. Microsoft's Copilot platform enables low-code agent creation, removing technical barriers that previously limited AI deployment to engineering teams.

The Brains Behind the Operation

But here's what makes this transformation possible: a new generation of LLMs that don't just generate text. They reason, plan, and orchestrate entire workflows. These "agentic" models work like a conductor leading an orchestra, with the LLM as the brain coordinating a whole toolkit of specialized components.

The leaders in this space each bring something unique to the table:

Anthropic's Claude 4 (May 2025) dominates code with its Opus 4 variant hitting 72.5% on SWE-bench. It can work through multi-hour sessions using external tools via Claude Code and the Modular Context Protocol (MCP). At the same time, the Sonnet 4 variant offers strong reasoning and memory at lower cost, helping push Anthropic to $2B in annualized revenue by Q1 2025.

OpenAI's GPT-4o launched with an Agents SDK and Responses API, enabling multi-step task completion through its "Operator" agent that can navigate websites and complete transactions.

Google's Gemini powers Agentspace, an enterprise platform built around the Agent2Agent (A2A) protocol for multi-agent coordination across tools and workflows.

Manus AI, launched in early 2025, achieved 86% task success on the GAIA benchmark, surpassing many peers. It specializes in autonomous multi-agent execution for domains like travel planning, web design, and finance.

Mistral AI joined this elite group in May 2025 with their Agents API, taking a different approach focused on enterprise orchestration. Instead of emphasizing individual agent capabilities, Mistral built a framework for multi-agent coordination. Their system handles agent handoffs seamlessly, and a finance agent can delegate to a web search agent or calculator agent mid-conversation, creating collaborative workflows that tackle complex problems piece by piece.

The technical foundation is impressive: built-in connectors for code execution, web search, image generation, MCP tools, and persistent memory across conversations. On the SimpleQA benchmark, Mistral Large with web search jumps from 23% to 75% accuracy, proving that connected agents dramatically outperform isolated ones.

We're witnessing the emergence of an "agent economy" where AI systems don't just respond to prompts but independently execute complex, multi-step tasks that deliver real-world results.

Software Development

Nowhere is this agent revolution more visible than in software development. The coding industry is experiencing its most significant disruption since the Internet, and the numbers are staggering.

Take Anysphere, creator of Cursor (an AI-powered VSCode fork). They raised $900M at a $9B valuation, jumping from $2.5B just months earlier. The AI editor generates nearly a billion lines of code daily and ranks among the fastest-growing software startups with $500M ARR and over a million daily users.

Tech giants are racing to integrate AI agents into entire product suites. Microsoft 365 Copilot, Salesforce Einstein, Adobe's generative features. Users now expect applications to have intelligent assistants, making AI capabilities a competitive necessity rather than a nice-to-have.

This reflects broader momentum: AI coding tool investment doubled from $420M (2023) to $780M (2024), and speed records continue falling. Swedish startup Lovable hit $17M ARR in just three months with 15 employees. Behind this overnight success lies 18 months of iteration.

Founder Anton's GPT Engineer gained 50,000 GitHub stars but struggled to monetize. After multiple pivots, they rebranded to Lovable in November 2024 with a breakthrough: solving the AI problem of getting "stuck" on large codebases.

The enterprise angle is equally compelling. Factory.AI emerged from stealth after quietly serving enterprise clients for two years. Their specialized AI "droids" handle complete development workflows, letting human teams focus on architecture and product strategy instead of grinding through tickets.

CEO Matan Grinberg envisions a near future where AI agents meet developers everywhere. Imagine tagging droids in Slack to turn brainstorming sessions into working code or watching backlogged tickets automatically become production-ready pull requests. Development becomes less about coding and more about understanding, planning, and testing.

The competition is getting fierce. Anthropic recently cut Windsurf's Claude access amid rumors that OpenAI wants to acquire the coding assistant for $3B. "It would be odd for us to be selling Claude to OpenAI," Anthropic's Chief Science Officer noted dryly, highlighting just how strategic these tools have become.

The stakes justify the aggression. Anthropic CEO Dario Amodei predicts AI will write 90% of code within 3-6 months, practically all code within a year. He envisions individuals running billion-dollar businesses solo by 2026, using AI for work that previously required massive teams. His warning about 50% job losses in entry-level white-collar work and 20% U.S. unemployment by 2030 might sound extreme, but the current pace suggests his timeline isn't far-fetched.

The Infrastructure Powering It All

While coding assistants grab headlines, the foundation is built in the frameworks and toolkits layer.

LangChain dominates this space with its modular approach to building AI applications. As of February 2025, its developer platform, LangSmith, has surpassed 250,000 user sign-ups, processed over 1 billion trace logs, and supports around 25,000 monthly active teams. The company was valued at $200 million following Sequoia’s $25 million Series A.

Salesforce's AgentForce proves the enterprise appetite is massive. The platform automates 84% of support tickets, closes over 1,000 paid deals, and delivers up to 40% cost reductions. With 380,000 internal inquiries resolved, it shows the scalability enterprises demand.

n8n exemplifies the no-code revolution in AI workflows. After pivoting to AI-powered automation in 2022, the Berlin-based platform saw 5X revenue growth, doubling in the last two months. Their recent $60M raise at a $270M valuation reflects investor confidence in democratizing AI development. "It's a prompt to build workflow," explains founder Jan Oberhauser.

Users can write "get information from X and send it to Y" in natural language instead of coding 50 lines for email integration. With 75% of their 3,000+ enterprise customers now using AI tools, n8n proves the importance of the infrastructure layer.

E2B tackles the elephant in the room: safety. Their $11.5M seed round funds open-source sandboxes that let AI-generated code run securely. Processing millions of instances monthly, they're becoming the essential safety layer that makes everything else possible.

When AI Agents Come to Play

Here's where things get even more interesting: gaming has become the ultimate testing ground for AI agents. Virtual worlds provide the perfect sandbox for testing how AI agents behave, cooperate, and evolve when given complete freedom to act. The gaming industry is witnessing a fundamental shift as AI agents evolve from simple NPCs to sophisticated autonomous entities.

Altera made waves by securing $11.1M after demonstrating agents forming complex societies in Minecraft, complete with politics, trade relationships, and cultural evolution. Now they've revealed their flagship product, Fairies, a general-purpose agent capable of executing over 1,000 actions across popular applications, from code generation to task parallelization.

Meanwhile, Nunu.ai has raised $6M to support their vision of unembodied agents that seamlessly adapt across different gaming environments.

Their practical impact is already evident: Roboto Games deployed Nunu to automate QA testing for Stormforge, achieving a 50% cost reduction while reclaiming 160 hours monthly through 400+ automated tests generated via natural language commands. The five-hour integration now delivers daily testing coverage spanning performance, combat mechanics, and UI validation, complete with video playback and instant Slack notifications.

DeepMind's SIMA agent demonstrates visual-based gameplay in titles like GTA V, responding to natural language commands through controller emulation.

Similarly, Stanford's Generative Agents showcased 25 autonomous NPCs orchestrating coordinated events while pursuing individual goals within a sandbox environment.

These developments signal gaming's evolution from static, predetermined systems to dynamic, intelligent simulations that fundamentally reshape both game design and AI-human interaction.

The Sound of Intelligence

Voice is becoming the dominant way we interact with AI. After all, talking feels more natural than typing, especially when you're asking an AI to book your flight, debug your code, or manage your calendar.

So Voice has emerged as the primary interface for AI agents, with several companies driving innovation in this rapidly evolving space. OpenAI's Realtime API launch in 2024 catalyzed speech-to-speech application development across multiple use cases, and voice AI companies now comprise a larger share of Y Combinator's recent cohorts than ever before.

Deepgram’s new enterprise-grade speech-to-speech architecture marks a key milestone in voice AI, enabling real-time, emotionally rich interactions without converting speech to text. This end-to-end design reduces latency and preserves nuance, crucial for enterprise use. With built-in debugging tools and efficient training via transfer learning, the platform lowers infrastructure costs at scale. As 84% of organizations increase budgets for voice agents, Deepgram stands out as a foundational voice AI provider.

ElevenLabs on the other hand, continues to lead in text-to-speech synthesis. Its latest v3 model introduces rich emotional control via inline audio tags like [excited], [whispers], and [sighs], making it ideal for media production, audiobooks, and character-driven content.

However, latency constraints make it better suited for asynchronous rather than real-time use cases.

Platforms like VAPI offer a comprehensive developer stack for real-time voice agents, handling voice recognition, synthesis, and conversation orchestration. For instance, Fleetworks leveraged VAPI to enable over 10,000 AI-driven calls per day, streamlining operations and slashing engineering overhead in the transportation sector.

Together, these advancements mark a shift toward fully integrated, immersive, and emotionally intelligent voice agents, enabling companies to deploy naturalistic voice experiences without building complex infrastructure from scratch.

From Voice to Action: Agents Take the Wheel

But voice is just the beginning. The real magic happens when AI agents can act on what you tell them, not just in games or through APIs, but directly in the web browser where most of our digital work happens.

Google's Project Mariner operates in the cloud, juggling up to ten tasks simultaneously while learning from user behaviors to automate complex workflows.

OpenAI's Operator goes further, directly interacting with live websites, filling forms, clicking buttons, executing transactions with near-human precision.

Amazon's Nova Act gives developers comprehensive browser automation tools built on Playwright, consistently outperforming competitors in benchmarks.

Anthropic's Computer Use, powered by Claude, prioritizes safe browsing experiences while maintaining strong reasoning capabilities, a focus on safety that distinguishes it in an increasingly crowded field.

Scouts tackles a different angle: always-on web monitoring. It watches for specific triggers like news updates, price drops, or reservation availability, delivering timely email alerts based on your criteria.

Think of it as having a tireless assistant constantly refreshing dozens of tabs for you.

Manus represents the current pinnacle, seamlessly combining browsing, reasoning, and tool usage to complete complex, multi-step tasks without human intervention.

The Infrastructure Explosion

All these agents need somewhere to store their data, learn from their interactions, and scale their operations. The result? A massive infrastructure buildout that's reshaping the entire data platform landscape.

Supabase has transformed from a simple Firebase alternative into the backbone of AI development. Native vector search support enables semantic search, similarity queries, and embedding-based retrieval, essential components for RAG pipelines and agent backends. Supabase reached 2 million developers by June 2025, doubling in just six months as AI-native developers gravitate toward its open-source stack combining authentication, real-time capabilities, and Postgres extensibility.

Neon presents an even more dramatic transformation. Three weeks ago, following Databricks' ~$1 billion acquisition, Neon disclosed that 80% of its serverless Postgres databases are now created by AI agents, a stunning reversal from October 2024's 30%. Agents now provision databases at 4× the human rate. Neon's sub-500ms provisioning, auto-scaling, instant branching, and pay-per-use model perfectly align with machine-driven experimentation and ephemeral database lifecycles.

Each platform demonstrates how data infrastructure is being reimagined for agentic workflows, where machines, not humans, are the primary consumers of database services.

The Human Element in an Agent World

With all this tech firepower, Claude 4's coding, Operator's web navigation, it's tempting to focus purely on capabilities. But here's the thing: the real transformation isn't what AI can do, it's how we work with it.

Winners aren't just deploying the latest models. They're rethinking work entirely. Smart companies identify where AI amplifies humans versus where it creates expensive automation nobody wants. They figure out what tasks people want to offload and what work they prefer to keep.

This isn't replacement, it's partnership. The best implementations treat agents as amplifiers, not substitutes. The agent economy is already here, reshaping business. You're either driving this change or watching competitors pull ahead.

Disclaimers:

This is not an offering. This is not financial advice. Always do your own research. This is not a recommendation to invest in any asset or security.

Past performance is not a guarantee of future performance. Investing in digital assets is risky and you have the potential to lose all of your investment.

Our discussion may include predictions, estimates or other information that might be considered forward-looking. While these forward-looking statements represent our current judgment on what the future holds, they are subject to risks and uncertainties that could cause actual results to differ materially. You are cautioned not to place undue reliance on these forward-looking statements, which reflect our opinions only as of the date of this presentation. Please keep in mind that we are not obligating ourselves to revise or publicly release the results of any revision to these forward-looking statements in light of new information or future events.

As AI increasingly takes over our work, one critical question emerges: Where will humans spend their time—and their newfound wealth?

For a closer look at the evolution of AI, join our Frontier Forum series where we’ll be talking to founders who are leading the charge in this emerging space. » Register here «

Productivity and Leisure

The consensus trade is clear: AI will rival the productivity boom of the industrial revolution or the internet. It will make businesses more efficient, reduce friction, and replace rote tasks with machine learning. The global economy will benefit.

Projected AI productivity gains by country:

But the best opportunities don’t sit in the consensus trade — they emerge in the knock-on effects.

As productivity increases, human labor becomes less essential. As automation handles our needs, we’re left with more time, more money, and a deeper desire for entertainment, challenge, and self-expression.

That surplus doesn’t just disappear. It flows somewhere.

Historically, it flows into leisure. Since the 1920’s the average household has more than doubled their expenditures in this category, now making up nearly half the average household income.

Over the past 30 years, this has increasingly meant excess wealth flowing into gaming.

Since 1990, gaming industry revenue is up over 800%.
In 2024, it crossed $347 billion globally.
Over 3.3 billion people now play games — over 40% of the world’s population.

Gaming is no longer a side industry — it’s the digital commons.

And in the AI era, that commons is transforming again.

The Rise of the AI Creator Stack

Gaming’s evolution mirrors that of traditional media.

In the beginning, studios dictated the experience. Players took what they were given—passively consuming content built for everyone, yet tailored to no one.

Then came user-generated content (UGC)—starting with mods, maps, and skins, and eventually evolving into entire player-built worlds.

Today, platforms like Roblox and Fortnite Creative are enabling entire generations to build worlds from scratch — and thanks to AI, they don’t even need to code.

AI-native tooling is unlocking a new kind of creator:

• Text-to-3D generation (scenario: “build me a fantasy village”)
  

• No-code world builders (drag-and-drop game logic with generative props)
  

• LLM copilots in Unreal Engine
  

• Automated NPC dialogue powered by GPT-like models
  

• Instant level design, soundtracks, even economy balancing
  

In short: what YouTube did for video, AI is now doing for interactive experiences.

And when you combine frictionless creation with an audience of billions, you get the most powerful economic force in the modern web: network effects.

UGC is Eating the Gaming World

For Generation Alpha, UGC isn’t a feature — it is the game.

They’re not just players — they’re creators, curators, and influencers. Roblox isn’t a product, it’s a platform for self-expression. Fortnite isn’t just a shooter, it’s a social hub.

We’re watching the same transformation that social media underwent 15 years ago:

• Traditional Media → Social Media
  

• Traditional Gaming → Social Gaming

And just like social media, the dominant platforms won’t evolve out of the games that have the best graphics or the biggest studios — they’ll be the ones that empower creativity at scale.

Social Gaming Is the New Social Media

Social platforms in the 2010s were feed-based: photos, captions, likes.

In the 2020s and beyond, social platforms are spatial and participatory.

Gaming is becoming the new public square.

It’s where kids hang out after school.
It’s where identities are formed, challenges are tackled, and friendships are made.

Roblox is the third place for hundreds of millions of players.
VRChat is the new dive bar.
Gorilla Tag is the new playground.

We’re already living in the Metaverse — we just don’t call it that anymore.

Where We’re Placing Our Bets

At Hartmann Capital, we’ve broken the opportunity down into four buckets:

1. Platforms – Environments where people gather, play, and build. (i.e. Squido Studio)
   

2. Tooling – AI-native creation engines, asset generators, multiplayer infrastructure. (i.e. Ludus AI)
   

3. Social Infrastructure – Digital identity, creator tools, communications infrastructure. (i.e. LIV)
   

4. Talent – Studios, brands, and creators crossing worlds and platforms. (i.e. DressX)
   

The Future Will Be Spatial

We believe that in an increasingly AI-driven world, the cost of goods and services will trend relentlessly downward, shifting the core of human consumption toward the richness of our experiences. Entertainment will evolve from passive consumption into active creation—experiences that challenge us, deepen our connections with others, and push us to explore the boundaries of our imagination.

AI is the catalyst turning every player into a creator, the accelerant making immersive worlds easier, faster, and cheaper to build, and the intelligence layer transforming generic digital experiences into deeply personal realities. Gaming platforms, empowered by AI, will become the new hubs of social interaction, collaboration, and creativity. Players won't just consume—they'll shape entirely new realities.

The future will be spatial, and spatial will be social.

Disclaimers:

This is not an offering. This is not financial advice. Always do your own research.

Our discussion may include predictions, estimates or other information that might be considered forward-looking. While these forward-looking statements represent our current judgment on what the future holds, they are subject to risks and uncertainties that could cause actual results to differ materially. You are cautioned not to place undue reliance on these forward-looking statements, which reflect our opinions only as of the date of this presentation. Please keep in mind that we are not obligating ourselves to revise or publicly release the results of any revision to these forward-looking statements in light of new information or future events.

For a live deep dive on stablecoins, join our bi-weekly Frontier Forum on Wednesday, March 12th, at 9 AM EST. » Register here «

The digital asset space is undergoing a structural shift. The cycle is broken. I foreshadowed this phase in June of 2024 in my 2024 Outlook, where I explored the fundamental shifts that could occur in a post-Gensler era. In particular, I wrote that:

“The most meaningful graduation from being speculative assets to investable assets for the wider digital asset space beyond Bitcoin’s digital gold thesis occurs the moment tokens can accrue value without being at odds with regulators. While part of the debate is whether or not digital assets are commodities or securities, the other meaningful debate is how crypto assets can be compliant in a way that does not go against the underlying technology.”

Despite Trump’s victory and sweeping regulatory changes, the market is having one of its worst stretches and appears to be charting a new course, away from its traditional four year cycle. We are at a unique juncture for the digital asset market, where all of the following appear true to me over the next twelve months:

1. Bitcoin has become the consensus trade, and Bitcoin dominance as a portion of the crypto ecosystem is likely nearing its all-time high for the remainder of history over the next 3–6 months.

2. Bitcoin will nevertheless continue to compound growth year over year.

3. The wider digital asset space will continue to undergo a period of crisis over the next six months.

4. The wider digital asset space will outperform Bitcoin in the decades thereafter.

The reason for this prediction is twofold.

Long Term: From Financial Nihilism to Utility

We are approaching the zenith of financial nihilism, exemplified by phenomena such as Trump coin, Milei's Libra, and the billions lost in memecoin ventures. As momentum has now firmly broken and most fundamentally unsound digital assets have declined by 90–99.9%, their return seems unlikely. Retail investors, as a result, are burned out. The speculative games have proven to be zero-sum and rigged, leaving the market fleeced and unwilling to participate any longer.

The next marginal buyer is Wall Street. Digital assets with strong fundamentals and revenues will be bid up by institutional buyers before any other digital asset is likely to draw institutional interest. It follows that retail, eager to make money, will flock toward whatever is appreciating, compounding the fundamental soundness premium further (as we see with the Mag 7 in equities).

The big innovation that the economy, government, and Wall Street are ready for is stablecoins—for those not familiar, fiat currency that has been moved onto the blockchain (tokenized dollars, euros, etc.). Already today, stablecoins have onboarded more than 1% of the entire M2 money supply and are growing at a breakneck pace —even in bear markets.. However, it will not end with mere transactions in USDT or USDC.

There will be increased demand for financial primitives like lending, borrowing, swapping, etc., of on-chain assets, driving material revenue toward leading DeFi protocols. Already today, stablecoins are the largest revenue contributors on the Ethereum blockchain, where Tether and Circle combined lead the list of fee sources. As the world continues its reliance on and use of stablecoins to save fees and time, the tokenization of most other financial assets follows closely.

Stripe, the world’s largest privately-held fintech company, raved about stablecoins in their 2024 annual letter, likening them to room-temperature superconductors for the financial services industry. They wrote:

“Stablecoins have four important properties relative to the status quo. They make money movement cheaper, they make money movement faster, they are decentralized and open-access (and thus globally available from day one), and they are programmable. Everything interesting follows from these characteristics.”

Today, stablecoins are primarily used for international remittances and as one of the fastest ways for those living abroad to gain exposure to U.S. dollars. However, stablecoin activity remains largely limited due to regulatory constraints, which the current administration is re-writing as we speak.

The even bigger innovation the world is likely not quite ready for yet—but which will yield some of the most outsized returns—is the intersection of crypto and AI. As AI agents gain access to a fully digitally native financial suite, they will become truly autonomous participants in our economy. The first step to unlocking this milestone, however, is onboarding the majority of online commerce onto stablecoins.

Short-Term Challenges: The Regulatory Gap and Catalyst Vacuum

Despite all these promising forces, regulation takes time. Until laws are drafted, signed, and adopted, it is unlikely that we will see fintech hit the go-button and roll out stablecoins at every checkout, instantly bringing every American on-chain overnight.

We are, in a way, stuck in an awkward in-between spot, where the “Trump trade” is behind us, macro risk-off weighs heavily on the market, and yet the actual meaningful changes will take longer than a simple Executive Order signing. While the market is starved for near-term catalysts, flows remain lopsided toward sellers.

Michael Saylor—the man propping up the market with over $20bn worth of purchases since November—is slowly running out of ammo. MSTR’s premium to NAV is one of the lowest in the last 12 months. I highlighted this in my December write-up, when the premium was near 2.5x. Today, the premium is down to 1.5x, a material decline in just three months. (ELI5: Saylor has been issuing and selling MSTR shares to finance Bitcoin purchases. This generates “yield” as he puts it as long as he sells the shares at a premium—however, that premium is now nearly depleted.)

Without Saylor manning the gunships, the market relies on real fundamentals. Unfortunately, as I have pointed out for years, 95% of assets are massively overvalued or even vaporware. VCs and Founders alike know this. As a result, their tokens, which are vesting daily, are heading for the exit, adding material recurring sell pressure.

Unfortunately, the market is punishing the other 5%—the fundamentally sound digital assets with millions of users and earnings—just the same. This may seem counterintuitive:

How can a decentralized exchange or lending market in a high-growth industry decline to 2–4 PE ratios or even sometimes trade at book value?

The answer lies in the composition of market participants pre-Wall Street:

1. Retail traders—who trade purely based on sentiment. They will sell AAVE at a single-digit P/E ratio for the same reason they buy Fartcoin on the way up. They react emotionally to price movement, with no fundamental thesis.

2. Quant traders and funds—who trade based on historical correlations. Crypto follows cyclical, highly correlated patterns: few tokens appreciate in a bear market, and even overvalued projects continue rising in a bull market. If 95% of tokens experience net outflows and decline, statistical arbitrage strategies will short or sell the remaining 5%, expecting mean reversion due to historically high correlation.

3. Crypto VCs—who raised funds between 2017–2021 and are now 4–9 years into their vintage, pressured to deliver returns to LPs.

4. Crypto-native liquid funds—which are not seeing inflows. If anything, many are facing outflows, as Bitcoin ETFs have diverted much of the institutional demand.

What you are left with is an awkward vacuum. While BTC is near all-time highs, the average crypto asset is down 50–99%, including Ethereum and Solana.

The largest resignation in the industry’s history is unfolding.

People are giving up, announcing on X that they are selling everything and leaving. The market has been thankless and treacherous for anyone outside of Bitcoin.

Even I have questioned whether running a liquid fund makes sense in these conditions.

But to me, it is undeniable that stablecoins will be globally adopted in the next 2–3 years. As a result, the underlying infrastructure will experience exponential revenue growth.

This is an asymmetric opportunity for those patient over the next six months—one we have not seen in four years.

How am I positioning our liquid fund for this environment?

In mid-February, we materially de-risked our portfolio after the market broke long-term momentum structures.

From here forward, we will take a strategic liquid venture approach, searching for the most fundamentally sound projects that will benefit from:

1. Stablecoin proliferation

2. AI integration into finance

The entire DeFi industry has a combined market cap of just $75bn—half the size of Uber.

In fact, every single token outside the Top 10 has a cumulative market cap of just $250bn. Our belief is not only that this segment will evolve into a $10TN+ economy (40x growth), but even more so that the market has violently mispriced its assets. Currently, perhaps $200bn in value is tied up in vaporware, while only $50bn or so is allocated to projects that are truly defining the technology. As capital rotates, we expect this imbalance to correct, driving exponential upside for the assets with real fundamentals.

At this juncture, I believe the era of trading has passed, and the time to invest has come. The noise of sentiment has likely cost most active funds—ourselves included—more upside than it has helped prevent downside. Going forward, I will be spending less time looking at price and more time focusing on value. When those two meet or come close, I will be a buyer, regardless of daily volatility.

My belief is that the right portfolio construction—a blend of 20–40 assets, accumulated amidst the coming sell-off, with exposure to both public and private offerings—will materially outperform not only Bitcoin but most asset classes.

The Four Year Cycle is Broken

Why?

Because mature industries don’t rely on four-year hype cycles—growth is constant. No matter how grim the market outside of Bitcoin looks today, and though it may get worse first, we are nearing a point where stablecoins will be used daily by everyone, whether they realize it or not.

And then? Everything tokenizes.

The first mover to watch is Coinbase, which recently announced plans to launch a tokenized version of its stock. When this succeeds, it will set the precedent for a future where all financial assets can be traded 24/7, instantly, globally, and used as collateral.

The $100bn on-chain economy will become a $10TN+ industry.

Nobody said it would be easy. But getting the stablecoin and tokenization trade right will make all this volatility worth it.

To explore this coming era in more detail, check out my write-up from last June, where I analyzed what a post-Gensler landscape could look like.

Disclaimers:

This is not an offering. This is not financial advice. Always do your own research.

Our discussion may include predictions, estimates or other information that might be considered forward-looking. While these forward-looking statements represent our current judgment on what the future holds, they are subject to risks and uncertainties that could cause actual results to differ materially. You are cautioned not to place undue reliance on these forward-looking statements, which reflect our opinions only as of the date of this presentation. Please keep in mind that we are not obligating ourselves to revise or publicly release the results of any revision to these forward-looking statements in light of new information or future events.

I wrote about the dawn of photorealistic video creation a year ago, analyzing early developments like Google's Lumiere and OpenAI's initial Sora announcement. Back then, we were limited to 5-second clips at 512×512 pixels. Today, the landscape has transformed so dramatically that those early achievements feel like ancient history.

From Preview to Product: Sora's Evolution

Sora has evolved from research preview to commercial product in OpenAI's suite, now generating 1080p videos up to 20 seconds long with support for text, image, and video inputs across multiple aspect ratios. Its professional toolkit includes Blend (transitions), Loop (sequences), Remix (editing), and Recut (regeneration), managed through Storyboard timelines.

The $200/month Pro Plan provides tiered resolution access (480p-1080p), 500 priority videos, and five concurrent generations. While Sora's commercial evolution is impressive, it represents just one facet of a broader technological revolution underpinned by significant scientific breakthroughs transforming the industry's efficiency, quality, and accessibility.

Breakthroughs: The Science Behind the Magic

AI video generation transformed dramatically from 2024 to 2025, marked by several groundbreaking achievements. Meta's VideoJAM (February 2025) revolutionized motion coherence in AI videos, reducing temporal artifacts by 95% while using 97% less training data.

Simultaneously, Google's VideoPoet emerged as a versatile large language model, unifying diverse video generation tasks, including text, image, and audio inputs with accurate lip-syncing. Using autoregressive architecture, VideoPoet delivers coherent video and audio through zero-shot generation. Further advancing video quality, MAGVIT v2 improved video consistency through sophisticated temporal modeling, resulting in significant stability gains for extended video sequences.

Step-Video-T2V emerged as a 30-billion parameter model in a significant open-source breakthrough, generating high-quality 540p videos at 30fps with unprecedented 16x16 spatial and 8x temporal compression ratios.

Underpinning these platform-specific innovations are three fundamental approaches: Denoising Diffusion Models (DDPMs) work like digital restoration experts, gradually refining video quality frame by frame. Score-Based Models (SGMs) act as artistic directors, ensuring videos follow natural patterns of movement. Stochastic Differential Equations (SDEs) serve as motion choreographers, creating smooth transitions between frames.

The field also saw breakthroughs in cinematic control and 3D understanding:

NVIDIA's ReMatching Dynamic Reconstruction technology can now reconstruct dynamic scenes from multiple viewpoints with unprecedented accuracy, while ByteDance's SceneDreamer generates explorable 3D environments from standard photos.

CineMaster (February 2025) introduced director-level 3D control through depth maps and camera trajectories, while Hailuo's T2V-01-Director brought professional cinematography to AI videos with natural language camera control supporting complex movements like tracking shots, dolly zooms, and multi-action sequences.

The industry also saw strategic partnerships for high-quality training data. In September 2024, Runway partnered with Lionsgate to create a customized AI model trained on the studio's proprietary catalog to augment filmmakers' creative workflows.

In December 2024, Lightricks partnered with Shutterstock to leverage their premium HD and 4K video library for training the open-source LTXV video AI model.

These collaborations marked a shift toward using professional-grade content for AI training, significantly improving output quality.

In practical terms, modern consumer GPUs can now handle Full HD (1080p) video generation in real time, while cloud platforms offer increasingly cost-effective processing options. Topaz Labs launched Project Starlight (February 13, 2025) - the first diffusion AI model to achieve complete temporal consistency in video enhancement, automatically handling upscaling, denoising, and sharpening of degraded footage without manual adjustments.

Leading platforms have integrated these capabilities into their workflows. Adobe Premiere Pro now features AI-powered auto-reframe and multicam editing.

ByteDance's innovations led the field with multiple breakthroughs in February 2025: Phantom achieved precise identity preservation in video generation, while OmniHuman-1 introduced the first unified framework for human animation that works with any aspect ratio and body proportion, driven by audio, video, or both.

SkyReels made history by open-sourcing SkyReels-V1 (February 2025), the first human-centric video model for AI short dramas, generating cinematic-grade performances with 33 facial expressions, over 400 natural motion combinations, and 24 fps clarity, while running on optimized consumer GPUs like the NVIDIA RTX 4090.

Founded by former Stability AI engineers, Black Forest Labs announced plans in 2024 to create competitive video generation tools, aiming for high-definition output and consumer GPU compatibility.

While no public release has occurred, partnerships like the February 2025 deal with Deutsche Telekom and posts on X suggest ongoing progress, positioning their text-to-video model as a potential rival to OpenAI’s Sora and Runway’s Gen-3, though specifics remain limited.

The Competitive Landscape: A New Era of Choice

The AI video generation market has evolved into a diverse ecosystem with platforms developing distinct identities across strategic segments.

Premium providers like OpenAI's Sora deliver 1080p quality with advanced storyboarding at $200/month, while Google's Veo 2 leads in physics simulation with 4K capability at premium rates ($1,800/hour). Adobe Firefly offers IP-safe 1080p content at competitive prices, targeting commercial applications where copyright matters.

Motion specialists have carved out their own territory: Alibaba's Wan 2.1 excels in complex body movements despite 720p limitations.

Luma AI's Ray2 achieves natural motion quality at 1080p with premium pricing ($0.75-$2.40/5s).

Genmo's Mochi 1 implements high-fidelity physics through efficient AsyncDiT architecture at 720p.

Specialized capabilities continue to emerge: KLING 1.6's multi-image interaction system, MiniMax's optimization for minimal movement, and TransPixeler's RGBA video with alpha channels for transparency. Meanwhile, niche players like Haiper 2.0 focus on interactive features, SkyReels V1 specializes in human-centric content, and MoonValley.AI emphasizes ethically trained models.

As this market matures, we'll likely see both consolidation among general-purpose platforms and further specialization from niche providers, with success favoring those that integrate seamlessly into existing creative workflows.

Real-World Integration in Professional Workflows

As AI video platforms mature, their integration into professional workflows reveals both potential and limitations. Professional productions often require custom solutions that transcend individual platform limitations. This reality is driving a new trend: the development of hybrid systems that combine multiple AI capabilities with traditional production tools.

Roland Emmerich's 'Space Nation' television series exemplifies this hybrid approach by combining Cybever's 3D-EnGen technology with Unreal Engine 5. As Emmerich notes, 'This is a technology that nobody can stop. It is too good, too realistic. I just hope that I can work with that and create incredible backgrounds.'"

Cybever's solution runs on conventional NVIDIA RTX workstations and interfaces with Maya and Houdini, decreasing pre-visualization time by 80%. Unlike traditional text-to-video approaches, which require full regeneration for changes, their technology allows for real-time modification of interactive sequences. "Nexus Rising" is another production example in which teams cut costs by 70% compared to traditional ways.

Meanwhile, social media platforms are adapting existing tools to their specific needs: YouTube has optimized Veo 2 for Shorts' vertical format, Instagram leverages Runway's rapid generation for Reels, and professional suites like Adobe Creative Cloud have integrated high-resolution capabilities. Beyond traditional media production, AI video changes digital marketing through the rise of virtual influencers.

Virtual Influencers and Industry Applications

The virtual influencer market exemplifies AI video's commercial potential, projected to grow from $12.3B (2023) to $95.6B (2032) with established revenue models already proving their worth. Top-tier virtual personalities like Lu do Magalu command $21,000 per post with 7.7M followers, while mid-tier creators such as Aitana Lopez generate $10,000 monthly.

These digital personalities consistently outperform human counterparts in engagement (2.84% vs 1.72%), though results vary dramatically by campaign - Lil Miquela's Samsung content reached 126M views while her BMW content achieved just 0.6% engagement against a 3.6% human benchmark.

Cross-industry adoption accelerates with healthcare implementing AI video in 79% of training programs, while education and manufacturing report significant ROI through personalized content and enhanced quality control. The ecosystem has specialized accordingly: Synthesia and HeyGen focus on content creation, D-ID and Hour One develop avatar technology, and Elai.io specializes in enterprise integration.

Investment now targets fundamental infrastructure - CGI, machine learning, and natural language processing platforms - as the market evolves beyond novelty into standardized practices with sustainable business models, signaling the transition from experimental technology to essential marketing tool.

Looking Forward

A year ago, we speculated about AI video generation's potential. Today, these tools have become essential components of professional content creation workflows. The question has shifted from technical feasibility to creative possibility, with several key trends emerging:

• Seamless Workflow Integration: AI video tools now complement rather than replace traditional creative software, enabling hybrid approaches that leverage the strengths of both.

• Advanced Physics Simulation: The latest models demonstrate significantly improved understanding of how objects interact naturally in physical space, eliminating many of the uncanny artifacts that plagued earlier systems.

• Granular Creative Control: Creators now enjoy unprecedented control over specific elements within generated videos, from lighting and camera movement to character expressions and environmental details.

• Post-Generation Refinement: The development of sophisticated editing tools specifically designed for AI-generated content allows for precise adjustments without complete regeneration.

The democratization of video creation continues at an unprecedented pace, transforming not just how we create content but how we conceptualize visual storytelling itself. What seemed like science fiction in early 2024 has become our creative reality in 2025, with the boundary between imagination and realization growing thinner by the day.

Disclaimers:

This is not an offering. This is not financial advice. Always do your own research. This is not a recommendation to invest in any asset or security.

Past performance is not a guarantee of future performance. Investing in digital assets is risky and you have the potential to lose all of your investment.

Our discussion may include predictions, estimates or other information that might be considered forward-looking. While these forward-looking statements represent our current judgment on what the future holds, they are subject to risks and uncertainties that could cause actual results to differ materially. You are cautioned not to place undue reliance on these forward-looking statements, which reflect our opinions only as of the date of this presentation. Please keep in mind that we are not obligating ourselves to revise or publicly release the results of any revision to these forward-looking statements in light of new information or future events.

Welcome to the Q4 Edition of the AI in Gaming Industry Report. Building on our Q3 coverage, which examined AI co-pilots, substantial investments, and the rapidly changing landscape of AI-driven technologies in the gaming industry, this quarter's briefing focuses on how these advances transform development processes and gaming experiences.

This report will focus on:

• The latest AI Developments

• High-Profile Funding and Acquisitions

• Cutting-Edge Research Highlights

Whether you're a developer, publisher, investor, or simply interested in the fast-evolving opportunities, this Q4 report will be your compact guide to everything going on at the intersection of AI and gaming.

In The News

• Millions of player-captured augmented reality scans from Pokémon Go and other Niantic games were used to create a new "Large Geospatial Model" (LGM), which the company unveiled. Niantic's Visual Positioning System gathers these scans, which map actual locations from a pedestrian viewpoint - often in places off-limits to cars. Integrating AI and 3D geometry, Niantic hopes to provide cutting-edge augmented reality and location-based game experiences. Players in 2016 probably didn't expect their gameplay data to power a commercial AI platform, which has raised privacy issues.

• Google DeepMind showcased Genie 2, a “foundation world model” that turns a single prompt image into a fully interactive 3D scene. Demonstrations highlighted advanced physics, character animation, and continuous world consistency lasting up to a minute. The system allows rapid prototyping, allowing humans or AI agents to interact with newly generated environments in real-time. Though hardware demands remain steep, Genie 2 offers a glimpse of how AI-driven tools could soon create entire game worlds on the fly.

• Microsoft and Activision unveiled plans to integrate advanced AI into future Call of Duty titles. AI features such as more intelligent NPC behaviors, procedurally generated missions, and real-time player feedback systems will enhance gameplay dynamics.

Investments

Over Q4 2024, these startups collectively raised about $315 million, with Luma AI ($90M), Moonvalley AI ($70M), Genmo ($28.4M), and Decart’s two rounds ($21M and $32M) representing the biggest fundraises. This quarter’s financing shows investor appetite in generative storytelling, AI-driven 3D creation tools, game distribution, and social gaming platforms.

The Latest Research

1. DeepMind's Genie 2 represents a significant leap in interactive game worlds. It features an 11B-parameter model that transforms single images into dynamic 3D environments.

   The system showcases sophisticated physics simulation and character animation, maintaining coherent worlds for up to a minute. However, computational demands remain a key challenge.

2. Dablander's research introduces a cognitive framework that places LLMs at the core of game AI. The system's modular design - incorporating perception, memory, thinking, action, and learning components - proved highly effective in tests with 25 AI agents who demonstrated sophisticated social behaviors and adaptive interactions.

   

   While Google DeepMind's Genie 2 successfully extended these concepts to generate interactive 3D game worlds with impressive physics and character animations, it still faces key limitations, including low frame rates (around 1 FPS) and occasional stability issues in maintaining world consistency beyond one minute.

3. The study "Integrating artificial intelligence into game development" highlights The Witcher 3's NPCs, which use behavior trees and emotional modeling to make realistic decisions based on simulated feelings, and Left 4 Dead's AI Director, which analyzes player performance to dynamically adjust difficulty.

   

   The research found these AI systems significantly improved player retention and satisfaction by creating personalized experiences that maintain an optimal challenge level throughout gameplay.

4. DreamGarden, developed by researchers at NYU, UT Austin, and Microsoft, bridges the gap between concept and creation with its hierarchical LLM system. Its innovative node-based interface gives developers precise control throughout development, proving particularly valuable in early prototyping.

   

   The system's strength lies in its ability to break down high-level prompts into actionable tasks through a "garden" of interconnected nodes that can be pruned or expanded based on developer needs. While studies showed users found value in the planning and visualization aspects, the system currently faces limitations in generating complex simulations and maintaining consistent performance in real-time environments. This balance between autonomous generation and user control positions DreamGarden as a promising tool for rapid prototyping and early-stage game development.

5. Microsoft's GamePlot revolutionizes narrative design with real-time story refinement and dynamic NPC control. However, a key study of 14 designers revealed a significant divide: developers favor automation, while writers prioritize creative control, highlighting the need for specialized AI tools.

6. Tencent's PlayGen achieves a breakthrough in real-time game generation, delivering 20 FPS on consumer hardware through advanced data sampling techniques. The system maintains exceptional visual quality and precise mechanics for extended periods, effectively solving the performance-accuracy trade-off.

   

7. UPenn's REGENT introduces a novel approach to agent adaptation, combining retrieval-based learning with transformer architecture. The system achieves superior results while using significantly fewer resources—3x fewer parameters and 10x less training data—challenging conventional scaling wisdom.

8. UCSC's ConvMesh offers a fresh perspective on 3D mesh optimization, employing convex optimization instead of neural methods. It achieves remarkable quality improvements using just 500 reference points, though the current two-hour processing time suggests room for efficiency gains.

9. UC Berkeley and Starward Games merge AI with DeFi in their GameFi ecosystem. They use GPT-4 and Claude to create context-aware agents that actively shape in-game economies. The system demonstrates enhanced player engagement while preserving blockchain principles despite computational challenges.

10. Modl.ai and Riot Games use an innovative sensor-based approach to tackle the challenge of human-like gaming AI. Their 14.9M-parameter model achieves impressive 9.59ms inference times on consumer CPUs and is remarkable in its authenticity—fooling 30% of players in human evaluation tests.

How AI Shaped Gaming in 2024

In 2024, gaming split into two camps over AI. While the PlayStation Boss says AI can "Revolutionize" gaming, Nintendo took a decisive stand against using generative AI in its titles.

Major publishers moved quickly to implement AI solutions. Electronic Arts led with over 100 AI projects, demonstrating their commitment through College Football 25's AI-generated stadiums and player models. Their innovation extended to Battlefield's development with AgentMerge for issue tracking. Meanwhile, Modl.ai achieved a technical breakthrough with its tactical shooter AI, using innovative sensor technology to achieve unprecedented processing speeds.

Research teams pushed technological boundaries even further. Google DeepMind's Genie 2 maintained consistent environment generation for up to 60 seconds, while Ubisoft's Project NEO demonstrated sophisticated NPC interactions. More importantly, these advances began reaching consumers. Retail Mage launched on Steam as one of the first AI-native games, while 1001 Nights moved into final development.

The industry underwent structural changes to embrace AI development. NCsoft established three dedicated AI studios, and Microsoft began integrating AI features into Call of Duty. Gaming AI startups secured $400 million in a single quarter. However, this AI boom coincided with industry-wide struggles, as over 30,000 developers lost their jobs since 2022.

Implementation challenges emerged as AI met reality. Palworld faced criticism over AI-generated content, while players quickly identified flaws in Call of Duty's AI-created items. The temporary shutdown of Itch.io over AI-related copyright concerns prompted Adobe to develop protective tools for artists, highlighting the need for better safeguards.

As 2024 drew to a close, the industry gained valuable perspective on AI's role. The most successful implementations focused on enhancing specific production processes rather than replacing human developers entirely.

Predictions for 2025: Breaking New Ground

1. Local Processing Takes Center Stage

In 2025, the gaming industry is witnessing a fundamental shift in AI implementation. NVIDIA's Jetson Orin Nano Super is at the forefront of this transformation, a breakthrough in local AI processing that delivers 1.7x better performance at an accessible price of $249.

NVIDIA's ACE technology has laid the groundwork for lifelike digital humans in games, providing sophisticated AI models for speech, animation, and behavior generation. This foundation is now enabling even more ambitious projects in local AI processing.

One of the most promising demonstrations of local AI's potential is GARP (Generative Agents Real-time Playground), a new technology that runs complex NPC simulations entirely on local GPU hardware. Running on a single NVIDIA 3090, GARP can simulate up to 20 unique characters, each with persistent memories, learning capabilities, and dynamic decision-making systems. These AI-driven characters engage in natural language interactions, develop distinct personalities, maintain daily routines, and even form relationships while coordinating activities with other NPCs.

2. Game Engines Embrace AI Integration

The major game engines are rapidly evolving to meet this AI-driven future. Unity and Unreal Engine are expanding their native AI capabilities, while specialized tools like Coplay, Ludus AI, and Cursor are revolutionizing development workflows.

This ecosystem of AI-powered tools is creating a more efficient and accessible development environment for studios of all sizes.

3. The Rise of AI-First Game Studios

A new breed of game studios is emerging, built from the ground up around AI-centric workflows. These studios are reimagining the entire development process, from initial concept to final playtesting. Advanced features previously requiring significant resources - such as persistent NPC memory systems, dynamic narrative generation, and real-time content adaptation - are becoming standard offerings thanks to improved consumer hardware capabilities.

4. AI Revolutionizes Esports

The competitive gaming landscape is set for transformation as AI tools become integral to professional player development. Advanced AI systems will fundamentally change how teams and players approach training, strategy development, and performance analysis.

Platforms like iTero are leading this movement, demonstrating the power of AI-driven coaching in League of Legends. These systems provide an unprecedented depth of personal performance analysis by analyzing over a billion data points through sophisticated machine-learning algorithms. Unlike traditional statistics that rely on general averages, AI coaches can process hundreds of unique stats per match, offering personalized insights and specific improvement recommendations that can significantly boost win rates.

Teams across the industry are increasingly leveraging AI for automated testing, asset optimization, and localization. Early adopters report incredible improvements in efficiency, with compressed development timelines and streamlined team structures. While adoption rates will vary between major publishers and independent studios, integrating AI tools into core development processes appears inevitable.

Success in 2025 will hinge on finding the right balance between innovation and practical implementation across all these fronts - from local AI processing and game engine integration to AI-first development and competitive gaming analysis. The winners will be those who can harness AI's potential to enhance development efficiency and game quality while maintaining the creative vision that drives memorable gaming experiences. As these technologies mature, they promise to transform not just how games are made but also how they're played and experienced by everyone, from casual players to professional competitors.

Andrew Wilson, EA CEO

Disclaimers:

This is not an offering. This is not financial advice. Always do your own research. This is not a recommendation to invest in any asset or security.

Past performance is not a guarantee of future performance. Investing in digital assets is risky and you have the potential to lose all of your investment.

Our discussion may include predictions, estimates or other information that might be considered forward-looking. While these forward-looking statements represent our current judgment on what the future holds, they are subject to risks and uncertainties that could cause actual results to differ materially. You are cautioned not to place undue reliance on these forward-looking statements, which reflect our opinions only as of the date of this presentation. Please keep in mind that we are not obligating ourselves to revise or publicly release the results of any revision to these forward-looking statements in light of new information or future events.

Bitcoin, in particular, has performed well during this period, reinforcing its role as a hedge against government debt, fiat currency risks, and Modern Monetary Theory's potential misuse. Its appeal as an alternative to traditional financial systems grows amid mounting government debt and fiscal irresponsibility. As governments overspend and central banks step in to support economies, the resulting interventions could heavily impact financial markets and the broader economy.

While the Fed often dominates financial discussions, the ECB's equally sizable balance sheet underscores its global market impact. The Eurozone's 20 member states face unique challenges, including diverse regional debt-to-GDP ratios. Despite recent quantitative tightening, the ECB's dollar reduction surpasses the Fed's. This article focuses on pressing developments in Europe, with a separate report analyzing U.S. policies under the Trump presidency and their implications for the Fed.

Project Eurozone

The Eurozone is a unique initiative uniting multiple European countries under the euro. Since its inception, it has grappled with economic disparities among member states.

The global financial crisis highlighted these challenges, particularly in Greece, Ireland, Spain, Italy, and Portugal, where debt crises exposed existing vulnerabilities. The Troika—comprising the European Commission, ECB, and IMF—played a central role by providing bailouts and implementing reforms. The ECB maintained low interest rates for nearly a decade, stabilizing the Eurozone, fostering recovery, and enabling affordable financing for debt-laden countries.

From 2009 to 2022, near-zero interest rates supported debt financing and economic growth. However, rising inflation—fueled by global supply chain disruptions, stimulus measures, and the Russia-Ukraine conflict—forced a policy shift. Higher energy prices further strained economies.

The key question is how monetary policy will adapt. With many Eurozone countries still burdened by high debt, the ECB may need to revisit low interest rate policies or purchase government bonds to manage yields and ensure economic stability.

Germany, France, and Italy, the Eurozone's largest economies, contribute over 50% of its GDP, underscoring their pivotal role. The ECB’s future monetary policy will closely track their performance due to their significant influence.

We begin with France's political situation, where recent events have widened the spread between its 10-year bond yields and Germany's.

Political Turmoil in France

France's current political instability began with Emmanuel Macron’s decision to call snap elections following his centrist alliance's poor performance in the European parliamentary elections. Marine Le Pen’s far-right Rassemblement National (RN) claimed 31.4% of the vote, dwarfing Macron’s alliance at 14.6%. Seeking to restore legitimacy and political direction, Macron dissolved parliament, but the snap elections only deepened divisions. The resulting hung parliament is now split into three major blocs: Macron’s centrist Renaissance party, the far-right RN, and the left-wing coalition led by France Insoumise.

France’s fragmented parliament quickly created governance challenges. Michel Barnier, appointed prime minister to lead a minority government, struggled to secure parliamentary support. His administration collapsed in December after just three months, following a no-confidence vote over a deficit-cutting budget, highlighting the government’s inability to navigate the fractured political landscape.

François Bayrou succeeded Barnier as prime minister, bringing political experience but inheriting the same obstacles: a divided parliament, economic strain, and the urgent need to pass a stop-gap budget to prevent a government shutdown. The deep divisions within the assembly and the rise of far-right and far-left forces have made coalition-building nearly impossible, leaving the government vulnerable to repeated disruptions.

France’s fiscal outlook is increasingly dire, with a deficit of 5.5% in 2023, projected to rise to 6.3% by 2025, and a debt-to-GDP ratio expected to climb from 113.3% in 2024 to 120% by 2027. On December 14, 2024, Moody’s downgraded France’s credit rating, citing political instability and limited prospects for deficit reduction.

To understand these challenges more fully, it is essential to analyze France’s economic and fiscal data over the past 24 years. Comparing GDP and government debt trends in France, Italy, and Germany provides critical context for the credit rating downgrade and its underlying causes.

The Never-ending Race?: GDP and Debt

When examining cumulative nominal GDP growth, Germany and France have shown similar development over the past 24 years, with 101% growth for Germany and 99% for France. In contrast, Italy has significantly lagged, particularly since 2008, achieving only 76% nominal GDP growth since 2000.

Turning to the government debt-to-GDP ratio, we see that France was similar to Germany until 2009, after which it diverged and began to rise significantly, spiking with the impact of the global financial crisis and the COVID-19 pandemic. Italy, on the other hand, has shown a steady increase in this ratio since 2008.

This raises a critical question: Why has France's debt-to-GDP ratio increased dramatically when GDP growth has been relatively similar to Germany's? The answer is that France's debt has grown much faster than Germany's, as the graph below clearly shows:

The reason for the faster-growing debt has been the high government deficit over the last 24 years:

After the global financial crisis, France ran a government deficit of up to 7% of GDP and needed 10 years to reduce it to 2.25% in 2019. This is remarkable, as government financing costs were suppressed by interest rates that were below 1% from 2009 to 2022 and negative for 8 years during that timeframe:

France’s high deficit is not the sole issue; the core problem lies in its inability to drive higher GDP growth. If the government had effectively used low-interest debt to boost economic development, the debt-to-GDP ratio would be far healthier. France's GDP growth per 1% debt increase is just 0.36%, compared to Italy's 0.65% and Germany's 0.93%, highlighting its inefficiency as an economic investor.

This underperformance has consequences: rising bond yield spreads compared to countries with more stable debt-to-GDP ratios, like Germany.

To identify potential solutions, it is worth examining Italy's strategies during the Eurocrisis for lessons that could guide France forward.

Italy’s Austerity

After the global financial crisis, Italy grappled with significant economic challenges, exacerbated by its already high public debt. Bond yields surged from 4% in 2010 to over 6.5% in 2012, making debt financing increasingly difficult. The turning point came in 2012 when Mario Draghi, then ECB President, delivered his pivotal "Whatever it takes" speech, restoring confidence in the Eurozone and stabilizing markets. Concurrently, the Italian government implemented austerity measures to curb public spending and rebuild trust in its finances. This blend of decisive monetary and fiscal actions restored market confidence, reducing bond yields to around 2% within two years.

Over the past 24 years, Italy’s GDP growth reached just 76%, significantly lagging behind Germany’s 101%, despite both countries taking on similar debt increases (Italy 108%, Germany 117%). This sluggish growth contributed to Italy's rising debt-to-GDP ratio. According to Baccaro and D’Antoni (2020), if Italy had matched Belgium’s growth rate, its public debt in 2007 would have been 90% of GDP instead of 106%.

Italy’s austerity measures from 2012 to 2015 further exacerbated its challenges. These measures caused a 5% drop in GDP, worsening debt rather than reducing it (Storm, 2019; Fazi, 2018). Cutting public spending to balance budgets shrank GDP, reducing revenues and deepening deficits, underscoring the intricate relationship between GDP and debt. Additionally, Italy faced disproportionately high interest payments, allocating 7–15% of annual expenditures to debt servicing, compared to Germany’s 1–7%, which severely limited fiscal flexibility and public investment capacity.

France now faces similar risks, with rising debt and financing costs as the ECB maintains higher interest rates. To avoid Italy’s pitfalls, France must control its debt-to-GDP ratio to reduce interest burdens, improve the efficiency of public spending to drive GDP growth, and stabilize its political landscape to address governance challenges and populist pressures. However, France’s high taxes, including the steepest capital gains rates in Europe (30–34%) and the highest tax-to-GDP ratio in the OECD, restrict its ability to raise revenues further. Combined with rising private and public debt, as noted by the BIS, these factors demand a comprehensive strategy to ensure stability and growth.

Debt Beyond the Government

Data from the BIS Credit of the Non-Financial Sector series shows growth in both public and private-sector debt in France.

With a total credit-to-GDP ratio of 319%, France leads the Eurozone's three largest economies and ranks second among developed countries globally.

Private debt levels play a critical role in assessing a country's vulnerability to economic crises. When considering austerity measures, the dependency of private debt on government spending must be factored in, as reduced government expenditure could exacerbate private-sector vulnerabilities.

Additionally, hidden potential government debt in the form of guaranteed loans warrants attention. Following the Ukraine war, European governments increased loan guarantees for companies affected by rising energy costs. In these arrangements, if a company defaults, the government assumes the debt, increasing public liabilities. Recent Eurostat data shows these guarantees represent significant GDP proportions.

In recent years, Germany and Italy have primarily used this tool to provide liquidity to companies without immediately increasing government debt.

Germany Under pressure

Germany is navigating a challenging economic landscape, primarily driven by the Ukraine war and rising energy costs, which weigh heavily on key industries like chemicals and steel. To provide relief without breaching the constitutional "debt brake" (Schuldenbremse)—capping structural deficits at 0.35% of GDP—the government has relied on guaranteed loans. While this approach offers short-term liquidity without directly increasing public debt, it has not spurred significant investments in an uncertain economic environment, limiting GDP growth.

The shutdown of Germany's last three nuclear power plants in 2023 has heightened reliance on fossil fuels and renewables. Despite ambitious renewable energy targets, the transition has been challenging, with high energy costs and supply concerns prompting companies, particularly in energy-intensive sectors, to delay investments or shift production abroad.

The debt brake itself poses limitations. While it underscores fiscal discipline, it restricts the government’s ability to implement direct economic stimulus or invest in long-term growth. External pressures could also test these fiscal rules. For instance, during Donald Trump’s presidency, calls for increased NATO defense spending and trade protectionism, including tariffs on German exports, strained Germany's economy. A resurgence of such pressures could lead to debates about relaxing the debt brake to enable targeted fiscal measures.

Germany’s political instability further compounds economic challenges. The departure of a coalition party has triggered new elections, scheduled for February, amid rising support for right-leaning nationalist parties like the AfD. Endorsements from high-profile figures such as Elon Musk have bolstered their standing, raising concerns about coalition-building. If the AfD gains significant influence, it could upend Germany’s political landscape, complicating governance as centrist parties refuse to cooperate with them.

These financial, energy, and political challenges are deeply interconnected and mutually reinforcing, leaving Germany vulnerable to economic stagnation and heightened uncertainty. With these dynamics at play, attention turns to the ECB’s role in navigating this complex environment.

The Role of the European Central Bank

The ECB plays a pivotal role in mitigating the compounding debt effect caused by high interest payments, especially as GDP growth lags behind debt expansion. To address this, the ECB can maintain low borrowing costs by either reducing interest rates across the Eurozone or purchasing government bonds in the secondary market to lower yields through increased demand. These tools are critical to stabilizing economies under fiscal strain.

France faces mounting fiscal challenges with its precarious debt-to-GDP ratio, high tax burden, and political instability, leaving limited options for deficit reduction or revenue generation. Reducing government spending could harm GDP growth, as seen in Italy, making France increasingly dependent on ECB support to avoid rising debt servicing costs.

Anticipating ECB Monetary Policy

Post-COVID inflation prompted the ECB to raise interest rates to 4%, later reduced to 3%. However, countries like France and Italy remain vulnerable due to higher debt growth and widening bond spreads. France's interest payments as a share of government expenditure risk reaching Italy’s 10% level, which would severely constrain fiscal flexibility. With current rates above France’s interest payment on debt level of 1.7%, the ECB may need to consider broad rate cuts or targeted interventions to lower French bond yields.

If inflation limits such measures, the ECB could deploy the Transmission Protection Instrument (TPI) to purchase government bonds from France and Italy, preventing unsustainable debt-to-GDP ratios and elevated interest payments.

The Transmission Protection Instrument

Introduced in 2022 to curb rising Italian bond yields, the TPI has acted as a reliable backstop. Its flexible framework allows the ECB to assist countries like France and Italy while influencing their fiscal policies. By leveraging this tool, the ECB can maintain stability in the Eurozone without immediately reducing overall interest rates.

The Future of the Eurozone

As populist movements gain traction, tensions over fairness and fiscal distribution within the EU are intensifying. The TPI’s targeted bond purchases may provoke perceptions of favoritism, creating divisions among member states. Simultaneously, increasing reliance on state-backed loan guarantees to support economies raises questions about long-term debt sustainability and coordination across the bloc.

The aftermath of Brexit underscored the economic risks of leaving the EU, yet dissatisfaction in some member states could reignite debates about exits. The ability of the EU and ECB to address these challenges will determine whether they can sustain the Eurozone’s unity and ensure its economic resilience.

Bitcoin: An antidote for debt?

In conclusion, Bitcoin emerges not only as a hedge for individuals but as a strategic lifeline for European institutions grappling with the Eurozone’s deepening economic and political crises. Its decentralized, debt-free structure and fixed supply make it immune to the inflationary pressures and fiscal mismanagement that plague fiat currencies. Bitcoin’s geopolitical neutrality and global accessibility position it as an unparalleled asset, free from the influence of major powers like the U.S. or China. For European institutions seeking to diversify away from euro risk and reduce dependency on dollar-dominated financial systems, Bitcoin offers a compelling alternative. Its unseizable nature and resistance to wealth confiscation measures, taxation, or capital controls further reinforce its value, especially in times of political upheaval. As wealth preservation becomes increasingly challenging in the face of rising debt-to-GDP ratios and fragmented governance, Bitcoin provides a sovereign, incorruptible means to store and transport wealth. In a rapidly evolving financial landscape, Bitcoin is not just a beneficiary of the Eurozone’s challenges—it is the antidote to them.

Disclaimers:

This is not an offering. This is not financial advice. Always do your own research.

Our discussion may include predictions, estimates or other information that might be considered forward-looking. While these forward-looking statements represent our current judgment on what the future holds, they are subject to risks and uncertainties that could cause actual results to differ materially. You are cautioned not to place undue reliance on these forward-looking statements, which reflect our opinions only as of the date of this presentation. Please keep in mind that we are not obligating ourselves to revise or publicly release the results of any revision to these forward-looking statements in light of new information or future events.

Past performance is not necessarily indicative of future results.

As virtual spaces increasingly mirror the real world, the demand for tools that enable self-expression and storytelling in these digital environments has skyrocketed. This shift has given rise to innovations like the LIV Creator Kit (LCK), which empowers creators to document and share their experiences directly from virtual reality.

A decade later, we have entered the era of the virtual selfie, where people capture digital representations of themselves as in game avatars.

LIV has introduced the Unity SDK—the LIV Creator Kit (LCK)—for VR creators to document their digital lives through various camera perspectives, including selfies, drone shots, and first-person views, all managed via a digital, tablet-like interface.

This tool supports the growing trend of active gameplay over passive consumption. According to a recent report by Midia Research, gamers spend 7.4 hours weekly playing games and 8.5 hours watching gaming videos.

The early demand for LCK is already validated. So far:

• 1M videos recorded in the first 6 hours of launch, >6M within the first 48 hours

• Avg recording length is 20mins

• 25 years of video was recorded in the first 48 hours (>250k hours)

LIV’s camera system benefits both developers and creators, amplifying engagement and content opportunities. Here’s how:

Benefits for Game Developers

Developers no longer need to focus solely on in-game retention and usage. Monitoring how games are shared off-device is equally critical:

Discoverability:

Player-shared content on platforms like TikTok, Instagram, or YouTube are mission critical for game discovery, especially indie/VR titles.

Since Meta’s switch from a curated store to an open one, 3rd party reach has become even more crucial. Developers rely on their own shorts channels, and having creators distribute their message for them.

Retention and active usage doesn’t need to stop when a player logs off - watching gameplay can still drive active engagement, and keep games at the top of mind, breeding long term engagement.

Importantly, LCK also gives game developers the ability to make promotional material for their content more accessible, without any additional tooling or permissions required.

Content Creators

One of the most fascinating insights into smaller content creators is their drive for social capital as a priority over monetary gain. Developers who can offer tooling to make gaining social capital easier or with an edge are positioned for success. Creators also benefit from tools that simplify and elevate their creative process. LIV’s features empower creators to expand their creativity by capturing gameplay for unique tutorials, highlights, or entertainment pieces tailored to their audience. High-quality content enhances their chances of achieving greater reach, gaining traction, and broadening their influence. Additionally, creators can diversify their formats by producing a range of content types, from live streams to bite-sized clips, tailored to different platforms and viewer preferences.

LIV Creator Kit: Transforming Content Creation

The LIV Creator Kit (LCK) merges gaming and content creation seamlessly. As an SDK, it brings virtual camera functionality to standalone Meta Quest (as an official partnership) and PCVR applications, enabling users to capture and share their in-game moments directly from VR. A low friction integration for developers and even lower for creators is expanding the relationship between play and consumption

So far, LIV LCK is officially enabled in Gorilla Tag and Cactus Jam, and soon 90+ other VR titles. If you’re a Unity developer interested in integrating LCK into your game, check it out here. LCK for Unreal is coming in Q1 2025.

Disclaimers:

This is not an offering. This is not financial advice. Always do your own research. This is not a recommendation to invest in any asset or security.

Past performance is not a guarantee of future performance. Investing in digital assets is risky and you have the potential to lose all of your investment.

Our discussion may include predictions, estimates or other information that might be considered forward-looking. While these forward-looking statements represent our current judgment on what the future holds, they are subject to risks and uncertainties that could cause actual results to differ materially. You are cautioned not to place undue reliance on these forward-looking statements, which reflect our opinions only as of the date of this presentation. Please keep in mind that we are not obligating ourselves to revise or publicly release the results of any revision to these forward-looking statements in light of new information or future events.

Google is making a fresh push into the world of headsets and extended reality (XR) with the launch of Android XR, a new operating system designed specifically for XR devices like headsets and smart glasses. Working alongside partners like Samsung, Google is gearing up for device launches as early as next year, with the platform now available for developers to begin building experiences, with a Hartmann portfolio company for XR wellness, Tripp, starting development.

The XR Market Explosion

The global XR market is experiencing unprecedented growth, with projections suggesting it could reach $450 billion by 2030. This surge is driven by technological advances in chip computing power, display technology, and the growing integration of AI. Enterprise adoption has been particularly strong, with companies using XR for training, remote collaboration, and design visualization. Additionally, consumer VR has grown meaningfully throughout the 2020s thanks to Meta’s consumer Quest line headsets.

Meta's Dominant Position

Meta has emerged as a clear leader in the consumer XR space, with its Quest series becoming the de facto standard for VR headsets. The consumer standalone Quest line of headsets , has seen significant success, shipping over 30 million units and establishing a thriving ecosystem of games and applications. The company's partnership with Ray-Ban for smart glasses has also proven more successful than expected, showing there's appetite for lighter, more casual XR experiences in addition to full tracking.

A New Era for Apps and AI

Android XR aims to enhance how users interact with immersive forms of existing technology by offering XR-specific versions of popular Google apps like Maps, Photos, and YouTube. A revamped version of Chrome will support multitasking, while the system will also accommodate existing Play Store apps, mirroring Apple's approach with iPad apps on Vision Pro. Central to the experience is Google's Gemini AI, which promises to make XR more intuitive and seamless.

Apple Vision Pro's Impact

Apple's entry into the market with Vision Pro has significantly raised the stakes. Despite its premium pricing, the device has garnered attention for its sophisticated eye and hand tracking, high-resolution displays, and seamless integration with Apple's ecosystem. Its focus on mixed reality and productivity applications has expanded the conversation beyond gaming and entertainment, potentially opening new markets for XR technology.

Bridging Virtual and Real Worlds The term "XR" reflects Google's ambition to cover the entire spectrum of virtual reality (VR), augmented reality (AR), and mixed reality (MR). Android XR-powered devices are expected to transition effortlessly between virtual and real environments. This includes standalone headsets, phone-connected accessories, and even audio-only devices like Ray-Ban Meta Smart Glasses.

Lessons from the Past, Eyes on the Future

Google has tried its hand at headsets before with projects like Glass, Cardboard, and Daydream, but none gained widespread adoption. Android XR represents a culmination of Google's investments in AI, the Android ecosystem, and wearable technology. A flagship device developed in partnership with Samsung, codenamed Moohan, is slated for release next year.

This also comes at a time when the XR market has matured, with an active developer ecosystem, and available standalone headsets with significantly reduced price tags.

The Competition Heats Up

With Apple, Meta, and now Google/Samsung all invested in XR, competition is driving rapid innovation. Each company brings unique strengths:

• Apple's hardware excellence and ecosystem integration

• Meta's social networking capabilities and gaming focus and market lead

• Google's AI prowess and Android ecosystem

This three-way race of the biggest tech firms will accelerate development and adoption of XR technology.

While Android XR is still in its infancy, Google is betting that advances in AI and immersive technology will make these devices indispensable. The real challenge? Taking market share from Meta.Google's latest venture marks an ambitious leap forward.

Whether it succeeds will depend on how well Android XR delivers on its promise of merging AI and XR competitors.

Disclaimers:

This is not an offering. This is not financial advice. Always do your own research. This is not a recommendation to invest in any asset or security.

Past performance is not a guarantee of future performance. Investing in digital assets is risky and you have the potential to lose all of your investment.

Our discussion may include predictions, estimates or other information that might be considered forward-looking. While these forward-looking statements represent our current judgment on what the future holds, they are subject to risks and uncertainties that could cause actual results to differ materially. You are cautioned not to place undue reliance on these forward-looking statements, which reflect our opinions only as of the date of this presentation. Please keep in mind that we are not obligating ourselves to revise or publicly release the results of any revision to these forward-looking statements in light of new information or future events.

Subscribe now

Never in my wildest dreams did I imagine the political landscape for digital assets would be as bullish as it is today. To be blunt: Bitcoin was expensive at $60k under Biden/Gensler. Today, Bitcoin is cheap at $95k.

It’s hard to overstate how positive the current set-up is, and I am no perma-bull. I’ve been cynical and bearish for much of the last two years, serving on the front lines of the crypto wars—experiencing everything from debanking to regulatory crackdowns, to the fraud at FTX. We’ve been through it all. Many will glance past digital assets today, only to buy in at the top in a year or two. Unless you’re already unfathomably wealthy, take 30 minutes to listen and pay attention to what happens next. Price asymmetries rarely exist in efficient markets. But in markets with complex regulations and sudden shifts, inefficiencies thrive. It’s critical to pay attention now, not in two years. If you haven't read my prior pieces, I highly recommend Part I and Part III of this series. Both serve as cold showers, unpacking the realities of the last four years. Their contrast will make today’s write-up resonate even more.

In this piece, I’ll uncover two core drivers for the price action over the next three years:

1. Mapping out the new political landscape.

2. Mapping out the new flows of buyers and sellers.

In my next piece, we’ll return to technology, exploring the latest breakthroughs in blockchain—from the first AI x Crypto experiments to early successes in Decentralized Science. But for now, this is the single most important thing you’ll read. If you miss the big picture, the nuance is a waste of time. If you get this right, it might just be the greatest 2-year trade of a lifetime.

It’s time to lock in.

A Game of Thrones: Toppling the Old Guard

The past four years in digital assets were defined by survival: banking crackdowns, lawsuits, and outright persecution by federal agencies. For many crypto entrepreneurs, this was the darkest period in the industry’s history.

But the war is won, and the tone has shifted drastically.

The previous administration sought to kill crypto because of its inability to control it, aligning itself with traditional banking interests. The new administration values crypto for its freedom and stands to profit from its success—whether for better or worse.

Here are the key political shifts:

Out: SEC Chair Gary Gensler

As I wrote six months ago, "If Trump wins, expect Gensler to be resigned at worst or in jail at best." One down, one to go. Gensler was one of the most corrupt SEC chairs in recent years, reprimanded by both the courts and Congress for his abuse of power. Removing him from the chessboard is the singular most significant move, likely unlocking meaningful guidance for crypto companies, clearing up what’s legal and what’s not, instead of continuing with regulation by enforcement.

In: Howard Lutnick - US Secretary of Commerce

Howard Lutnick is the CEO and Chairman of Cantor Fitzgerald, the financial services firm that holds ~$100 billion worth of treasury securities backing the world’s largest stablecoin, Tether. Shortly after his appointment, Cantor Fitzgerald invested in Tether, securing a 5% stake in the company. This is extremely bullish for several reasons.

1. Senators like Sherrod Brown and Elizabeth Warren have tried for years to kill stablecoins. Now, stablecoins are in the President’s ear.

2. Tether has been a source of FUD for nearly a decade, with rumors about it not being fully backed. I believe these concerns will finally be put to rest.

3. Howard is also extremely pro-Bitcoin, openly admitting in a recent interview that he personally owns hundreds of millions of dollars worth of BTC.

Read that again, and let it sink in.

Out: Senator Sherrod Brown

Senator Sherrod Brown, a three-time incumbent and former chair of the Senate Banking Committee, has long been one of crypto’s fiercest adversaries. He’s worked relentlessly to de-bank the crypto industry, cut off its access to traditional banking rails, and criminalize digital asset activity. Many of his statements can be found here.

During my seven years managing a digital asset fund, I had to switch banks four times, and saw three pro-crypto banks forced out of business under Brown’s rule. The crypto lobby mobilized to take him down, installing a pro-crypto Senator in his place. This was a pivotal race to end banking discrimination, often referred to as Operation Chokepoint 2.0. It also demonstrated the growing strength of the crypto lobby—step on our toes, and you might lose your job.

Speed Round - Crypto Advocates in the President’s Ear

• United States Secretary of Defense – Pete Hegseth

• United States Secretary of Health and Human Services – RFK Jr.
  RFK Jr. has publicly stated that he holds most of his net worth in Bitcoin.

• Department of Governmental Efficiency – Elon Musk
  The world’s greatest entrepreneur is perhaps as much a right-hand man to President Trump as JD Vance. Both Tesla and SpaceX hold approximately $1 billion in Bitcoin each.

• Secretary of the Treasury - Scott Bessent

  “Crypto is about freedom and the crypto economy is here to stay. Crypto is bringing in young people, people who have not participated in markets.”1

The rebels struck back and toppled the empire. The primary anti-crypto foes remaining in DC are Elizabeth Warren and Brad Sherman. Both are in their 70s and are unlikely to sustain their antagonistic rhetoric after this election cycle. Elizabeth Warren, in particular, who once ran her campaign on "building an anti-crypto army," has effectively ostracized herself from a significant portion of Democrats. Following the crushing defeat in November, many of them are likely to shift further toward the center rather than continue their drift to the far-left.

There will come a day when digital assets are likely used as a money-making scheme by some of those in power. It would be naive not to recognize that many of the names involved are driven by self-interest. Just as Pelosi insider-trades her way to billionaire status, I fully expect things to get dirty in DC, with government deals being handed to certain projects. It's inevitable, and will likely lead to new forms of inefficiency and friction—such as a technologically inferior protocol lobbying politicians, who then vote to put state activity on it, instead of adopting the better tech.

Politically, for now, crypto essentially operates under a regulatory Iron Dome. There are numerous bullish catalysts ahead, with few material bearish threats able to breach the ranks of pro-crypto leaders.

The music is playing. So we dance.

Follow the Money: Flows are Coming

The floodgates are open, so where is the flood going to come from? Here are a few key buyers to keep an eye on:

MicroStrategy - The King of Bitcoin

As of today, MicroStrategy has accumulated a staggering 402,100 Bitcoin—nearly 2% of all Bitcoin ever mined. The company, which was the first publicly traded firm to adopt the Bitcoin standard, has recently ramped up its mission to own more Bitcoin than anyone else. A remarkable 37.5% of its Bitcoin holdings were purchased since November. Founder and CEO Michael Saylor has devised a straightforward strategy: issue convertible debt and buy Bitcoin with it. Additionally, the company sells its own shares on the market whenever their premium is deemed too high and uses the proceeds to buy Bitcoin. MicroStrategy’s ongoing capital raises could be a major market-moving factor, as they’ve already raised $21 billion since November and plan to double that in the next three years. At present, MSTR trades at a 2.4x premium to its net asset value (NAV), leaving ample room for further share sales to acquire Bitcoin.

MicroStrategy may one day trigger the next bear market, but first, it could become the largest company in the world. If MSTR were a Shakespearean tragedy, Act I would be the rise (2020-2024), enduring countless storms. Act II will be the Saylor era, where he is put on par with Musk, Zuckerberg, and Huang. Currently worth $70 billion, MicroStrategy is among the largest publicly traded companies. During this act, we’re likely to see MSTR added to the S&P 500, which will further fuel Bitcoin's rise. When Saylor becomes one of the richest people alive and MicroStrategy the biggest company in the world, it will be worth tracking the MSTR premium to NAV to see when he runs out of ammunition, and how high we will have driven his cost basis and debt load by then. I believe an unraveling is inevitable, but for now, he’s sitting on tens of billions in dry powder, likely enough to push Bitcoin past $100k and unlock the next $100 billion in firepower. It’s like the Terra Luna trade in a way—except we’re still phenomenally early, and every short-seller will be punished before the trade unwinds (for context, Terra Luna was the biggest P/L trade of my career).

Part of the reason we’re so early is that at least a dozen publicly traded companies are likely to adopt Saylor’s model in the next 24 months. Momentum begets momentum, and talk of supply shortages will come sooner than expected.

Supply Shortages – Not Enough Bitcoin to Go Around

Back in 2019, I toured various colleges to give seminars on blockchain and Bitcoin. One of the simplest mental models to illustrate Bitcoin’s scarcity was the math of Bitcoin per millionaire. There are approximately 58 million millionaires globally, but only 21 million Bitcoin—likely fewer than 16 million when accounting for lost coins. This means there’s not even half a Bitcoin for each millionaire in the world. When adjusted for companies like MicroStrategy gobbling up 2%, the scarcity becomes even more pronounced.

There are two efficient ways to track supply shortages:

1. Issuance
   Since the halving in May, approximately 400-500 new Bitcoin are mined daily. At $100k per BTC, that’s only $40-50 million in new supply per day. In contrast, the average daily ETF inflow is $137 million according to Farside2. So for every 1 Bitcoin mined, roughly 3 Bitcoins are being snatched up by institutional investors.

   Where does the extra supply come from? OTC desks and directly from exchanges...

   

2. Exchange Balances
   Long-term Bitcoin holders tend not to keep their coins on exchanges but rather in custody (self or otherwise). The amount of Bitcoin on exchanges peaked in 2022 and has been on a 5-year decline since. Today, there are fewer Bitcoin on exchanges (despite issuance!) than at any point since 2019. As the tradable supply shrinks, supply squeezes will occur, leading to significant price movements, followed by phases of redistribution by profit-takers and new buyers.

Sovereigns Come Knocking

Donald Trump, RFK Jr., and Senator Cynthia Lummis have all promised something that remains widely under-reported: the birth of a national strategic Bitcoin reserve.

Yes, the United States plans to buy Bitcoin, and according to Lummis, quite a lot of it. In her BITCOIN Act, introduced in the US Senate, she proposes that the US purchase 5% of the Bitcoin supply—roughly 1 million BTC—over a span of several years.

The idea itself is a powerful narrative.

The implementation would trigger flows of historic proportions.

The ripple effects could lead to at least half of the world’s sovereign nations acquiring Bitcoin within a few days to five years of the US formally approving the strategy. I explored the game theory behind sovereigns contemplating Bitcoin adoption four years ago, and it remains a compelling thought experiment.

While sovereigns are lining up to buy Bitcoin, another massive injection of capital is coming from the aftermath of FTX’s collapse.

FTX - A $14BN+ Cash Injection

One of the industry's greatest collapses will soon provide it with renewed life. After years of forced selling by the bankruptcy courts, it’s expected that the FTX estate will return anywhere from $14-16 billion in cash to former users who lost assets due to the fraud committed by Sam Bankman-Fried. While the selling had pushed markets lower, the inverse will soon occur. Billions in cash flowing into the hands of crypto natives amidst a raging bull market will lead to one thing: more buying.

Institutional Buying Beyond Bitcoin

I expect 2025 to be the year when the first wave of Wall Street traders secure compliance approvals, regulatory green-lights, and infrastructure to expand beyond Bitcoin. Currently, traditional funds are predominantly focused on Bitcoin and ETH ETFs, but soon, they will likely begin buying and trading tokens across various themes—ranging from DeFi to DePin, and RWA. While $130 million per day in buying doesn’t move Bitcoin significantly, these buying patterns could materially reprice the broader crypto market. The charts suggest that this shift has already begun.

In bear markets, Bitcoin regains a significant share of market dominance. Typically in bull markets, it loses some of that dominance. As attention shifts again from Bitcoin to the broader ecosystem, we’re likely to see dominance drop to 25-40%, which suggests multi-trillion-dollar appreciation for alternative digital assets. This is the trade we are most focused on.

A Perfect Storm

The current set-up for digital assets could not be better. We’re as close as possible to a regulatory blank check, with interest rates being cut, the M2 money supply expanding after a two-year contraction, and historic amounts of capital flowing into digital assets through ETFs, MicroStrategy, sovereign nations, and more.

For now, the set-up is so simple that I expect the majority of tokens to rise—regardless of how little utility some tokens have (e.g., XRP is now worth a staggering $250 billion). This phase could last anywhere from a few more days to several months. In the latter half of the bull market, when alt season truly kicks in, we’ll likely see greater return dispersion. “Dino coins,” as we call them—tokens from previous cycles with little to no use case—will be more suppressed as teams and investors take profits (see the return profile of EOS, NEO and IOTA for reference). Meanwhile, new token launches will create fresh opportunities and blockchain use cases. Emerging projects, like Virtuals and Hyperliquid, are already showing promise—expect many more breakthroughs as this cycle matures.

Crypto can run on fumes, hope, and excitement for 3-6 months, but to complete a full cycle, the industry will need to rest on major technological breakthroughs that capture the attention of millions. The intersection of crypto and AI appears to be the first such opportunity—injecting new life into digital assets the way DeFi did in 2020.

The next 2020-2021 is on the horizon, but this time, we have reason to believe it will be materially bigger. The industry is likely to hit its first batch of products reaching mass adoption this cycle. There will be sell-offs along the way, but in our view, as long as the underlying assets remain sound, future sell-offs will be frequent but short-lived, representing attractive buying opportunities. In previous cycles, Bitcoin often experienced drawdowns of 30%, sometimes 7-12 times per cycle. If you don’t yet have a digital asset strategy, now is the time to formulate it.

Disclaimers:

This is not an offering. This is not financial advice. Always do your own research.

Our discussion may include predictions, estimates or other information that might be considered forward-looking. While these forward-looking statements represent our current judgment on what the future holds, they are subject to risks and uncertainties that could cause actual results to differ materially. You are cautioned not to place undue reliance on these forward-looking statements, which reflect our opinions only as of the date of this presentation. Please keep in mind that we are not obligating ourselves to revise or publicly release the results of any revision to these forward-looking statements in light of new information or future events.

Past performance is not necessarily indicative of future results.

Did you know that some of gaming's biggest successes started as fan modifications? Let's dive into the surprising numbers and facts behind the modding phenomenon that shapes the $200+ billion gaming industry.

The Birth of Modding

While modding traces its roots to 1983's Castle Smurfenstein, the release of Doom in 1993 marked a true landmark moment. It introduced the WAD (Where’s All the Data) system, a digital container holding game content like graphics, sounds, and levels. The system was structured with IWADs (Internal WAD) containing the original game content, while PWADs (Patch WAD) allowed players to create and share custom modifications. This modular approach made it simple for players to modify or expand the game by adding their content or replacing existing elements, establishing modding as a standard feature in gaming.

The rise of the internet in the late 1990s dramatically amplified this impact by providing an efficient platform for distributing mods globally. While console gaming remained a closed ecosystem, the PC platform's open nature made it the perfect environment for modding communities to flourish. These communities emerged wherever players identified limitations or unmet needs, pushing games to adapt and evolve through practical solutions - from bug fixes to enhanced features. But mods didn't just extend games. Justin Fisher's 'Aliens Total Conversion' for Doom introduced stealth mechanics years before they became mainstream, demonstrating how modders often pioneered new gaming concepts.

A prime example is Team Fortress, a 1996 Quake mod that revolutionized multiplayer gaming with its team-based, class-driven gameplay. Valve recognized its potential and hired the mod's developers to recreate Team Fortress using their GoldSrc engine (also used in Half-Life) to promote Half-Life's software development kit.

This led to Team Fortress Classic, published by Sierra Studios in 1999. The mod's popularity soared, becoming second only to Counter-Strike in online player count. In 2003, it transitioned to Valve's Steam platform and inspired Team Fortress 2 in 2007.

As modding became more popular, dedicated platforms emerged to organize and distribute these modifications. CurseForge, founded in 2007, transformed from a World of Warcraft mod site into gaming's premier modification platform, notably expanding to Minecraft in 2011.

After acquisitions by Twitch (2016) and Overwolf (2020), it now hosts mods for titles like Minecraft (66.1B downloads), StarCraft II (16.2M downloads), and Terraria (16.6M downloads), among others. With platforms like CurseForge providing infrastructure, many mods achieved remarkable commercial success.

Success stories

Today's Steam most-played list reveals its lasting impact - three of the top games are "mod-born": Counter-Strike 2 (1.2M daily players), Dota 2 (577K players), and PUBG (645K players).

Counter-Strike evolved from a Half-Life mod to Steam's most-played game. CS:GO alone has generated over $1 billion in revenue and maintains 1 million+ daily active players, DotA transformed from a Warcraft III custom map into an esports phenomenon generating over $200 million in prizes, and PUBG emerged from an Arma 3 mod to sell 75 million copies and launch the billion-dollar battle royale genre.

Looking again at the Steam charts, many top games, such as Grand Theft Auto V, Left 4 Dead 2, Stardew Valley, and War Thunder, maintain popularity through robust modding support. While these success stories demonstrate modding's potential, the true scale of modern modding becomes evident in today's infrastructure.

Platforms and Infrastructure

Today's modding ecosystem operates on an unprecedented scale. Nexus Mods hosts over 300,000 mods across 1,000+ games, while Steam Workshop supports 20,000+ titles. The landscape is evolving further with cross-platform solutions like mod.io, which has achieved remarkable reach with over 1 billion downloads, 25 million users, and 4 million pieces of user-generated content across 180+ games. Through its professional-grade SDKs, game plugins, and REST APIs, mod.io enables developers to implement modding support in any game.

One prominent example using mod.io's platform is Baldur's Gate 3, which achieved 22 million total downloads, 1.7 million within 24 hours of console launch, and over 2,000 mods created by 768 authors. With 40% of players using mods and an additional 231 million downloads on Nexus Mods, it demonstrates the cross-platform reach and massive appetite for community content.

Professional developer tools are also evolving. Bethesda's Creation Kit, advancing from Gamebryo to Creation Engine 2 (2024), features Havok Behavior animation, Radiant AI for NPCs, and dynamic quest generation. This toolkit powers massive modding communities like Skyrim, with 60,000+ mods and its most popular mod, SkyUI, reaching 20 million downloads. The modding community's potential is evident in Starfield, where modders recently developed sophisticated NPC routines, adding 40 named characters with realistic daily schedules and 12 robot shopkeepers.

Even more ambitious is Skyblivion, which evolved from a simple asset port in 2012 to a complete remake of The Elder Scrolls IV: Oblivion, now backed by a 40+ person team including professional developers.

Their 2023 announcement trailer drew millions of views, and after winning ModDB's "Best Upcoming Mod" award in 2017, the project is scheduled for release in 2025.

While some games provide built-in modding support, others rely on community-developed solutions. Minecraft exemplifies this with its modding ecosystem built around Minecraft Forge, a third-party mod loader that enables Java-based modifications.

Despite lacking official tools, Minecraft's modding scene thrives through community innovation, demonstrating how passionate players can overcome technical barriers.

The Business Case for Modding

This thriving ecosystem of platforms, professional tools, and community solutions has created compelling business opportunities. GameDiscoverCo's analysis of Steam's top 1,000 games reveals games with user-generated content achieve 64% better retention after two years, climbing to 90% after five years.

Genre-specific revenue boosts are equally impressive: action games see a 40% increase, RPGs 38%, and casual games 23%. Companies embracing mods experience 50-70% longer product lifecycles and 50% higher player retention.

This success has transformed modding from hobby to career launcher - 28% of game developers started as modders. Notable success stories include Robin Walker evolving from Team Fortress mod creator to Valve's lead developer, and Brendan "PlayerUnknown" Greene parlaying his DayZ mod into PUBG and a $50 million net worth. Similarly, DayZ creator Dean Hall founded RocketWerkz studio, with DayZ reaching 4 million copies sold by 2018.

Even Rockstar Games, historically resistant to mods, recently acquired Cfx.re, the creators of FiveM and RedM multiplayer mods for GTA 5 and Red Dead Redemption 2 - the same team they had banned eight years earlier.

Key Success Factors and Future Outlook

The most successful modding ecosystems combine robust tools, suitable game design, and passionate communities. While monetization challenges persist, as seen with Starfield's controversial paid mods, new approaches are emerging. ARK: Survival Ascended partnered with CurseForge to enable cross-platform modding with premium options, offering mod authors 50% revenue share and hosting a $350K modding competition. This shift toward professional modding has attracted major industry investment, from Microsoft's $7.5 billion Bethesda acquisition to Epic Games' built-in tools.

With 25% annual community growth and AI-assisted creation emerging, modding's impact appears boundless. Already extending game lifecycles by 50-70% and generating substantial revenue (Minecraft: $350M annually), modding continues to blur the line between community creation and professional development.

Disclaimers:

This is not an offering. This is not financial advice. Always do your own research. This is not a recommendation to invest in any asset or security.

Past performance is not a guarantee of future performance. Investing in digital assets is risky and you have the potential to lose all of your investment.

Our discussion may include predictions, estimates or other information that might be considered forward-looking. While these forward-looking statements represent our current judgment on what the future holds, they are subject to risks and uncertainties that could cause actual results to differ materially. You are cautioned not to place undue reliance on these forward-looking statements, which reflect our opinions only as of the date of this presentation. Please keep in mind that we are not obligating ourselves to revise or publicly release the results of any revision to these forward-looking statements in light of new information or future events.

Welcome to the first edition of the AI in Gaming Industry Report. Following our last article about "The AI Revolution in Gaming," we've decided to provide quarterly updates to help you stay informed on the rapidly evolving landscape of AI in games.

In this report, we'll dive into the latest developments and news, including:

• The most recent announcements and breakthroughs.

• Notable fundraising rounds and investments in AI gaming startups.

• Summaries of cutting-edge academic papers and industry studies.

For this quarter’s feature story, we look at the emerging trend of AI co-pilot for developers. From GitHub Copilot to Anysphere's Cursor, these AI-powered assistants are changing how games are coded.

Whatever your role—a game developer, an industry executive, an investor, or a casual observer eager to get a handle on what the future of gaming might look like in the context of AI—this report is designed to serve as your quick and up-to-date briefing.

In The News

• Electronic Arts, a widely criticized game studio and publisher, fully supported generative AI during its Investor Day 2024 presentation. CEO Andrew Wilson emphasized that AI is more than a buzzword; the company has over 100 active AI projects. The company also introduced its new Imagination to Creation tool, which uses AI prompts to create a video game in real-time.

  

• During an investor Q&A session, Nintendo's president, Furukawa, highlighted intellectual property concerns and the company's emphasis on unique experiences. While Nintendo is open to technological advances, it remains cautious, prioritizing dependable solutions over rushing to embrace new trends.

• After CEO Christoph Hartmann's comments about AI potentially streamlining game development, Amazon Games faced backlash. The company clarified that voice actors are still "essential contributors" and emphasized responsible AI use.

While big companies made headlines, the sector saw significant investment activity, indicating a growing belief in AI's ability to change how games are created and experienced. Let's take a look at the most notable funding rounds.

Investments

All of these startups raised $418.35 million this quarter, with World Labs ($230 million), Volley ($55 million), and Qodo ($40 million) leading the way. Investments support companies developing AI-driven game creation tools, social gaming platforms, narrative generation systems, and AI-enhanced game distribution channels.

The Latest Research

1. The GameNGen Doom project, a groundbreaking AI engine that allows neural networks to simulate playable video games, was probably the most talked-about story this quarter.

The AI imitates games like Doom with such accuracy that it is nearly impossible to distinguish between authentic gameplay and simulation, thanks to diffusion-based models similar to those used in image generation. However, the project still faces several challenges, including high computational requirements and the need for large amounts of training data. Despite these challenges, it demonstrates AI's growing ability to understand and reproduce complex game mechanics and visuals.

2. Building on the theme of AI comprehension in gaming, researchers presented VIDEOGAMEBUNNY, a specialized AI model for understanding game content.

   

   Trained on a diverse dataset of over 185,000 images from 413 games, this model outperforms larger, general-purpose AIs in game-specific tasks such as identifying game elements, understanding mechanics, and spotting glitches.

3. To bridge the gap between AI understanding and player interaction, Rao et al. studied the dynamics of human players collaborating with AI-powered non-player characters (NPCs) in a custom Minecraft game.

   

   Using GPT-4 for natural interactions, the study found promising potential for engaging gameplay and mutual assistance, which is especially beneficial to new players. However, it also revealed significant limitations in AI's understanding of 3D environments and real-time game states, resulting in inconsistent NPC behavior and spatial awareness.

4. Moving from understanding to creation, Zhang et al. presented a text-to-game engine capable of creating playable role-playing games (RPGs) from simple text descriptions. The "Zagii" engine uses advanced AI models to create game stories, characters, and visuals in real time.

   

   With 803 games generated and over 60,000 gameplay sessions logged, this technology demonstrates its ability to democratize game production, making it more accessible and efficient.

5. Complementing the text-to-game approach, the Sketch2Scene system represents another leap in AI-assisted game development.

   

   This unique application turns simple doodles into realistic, interactive 3D game worlds. Sketch2Scene's ability to automate the production of terrain, objects, and textures has the potential to streamline game world creation.

6. Researchers developed an AI system to evaluate and improve video game tutorials, addressing the crucial aspect of player onboarding.

   

   The technology uses Vision-Language Models (VLMs) to evaluate tutorial screenshots and identify possibly confusing or unclear instructions. Tested on a tower defense game, this method promises to improve tutorial quality while eliminating the need for intensive human playtesting.

While this research demonstrates AI's potential for game comprehension and production, another key trend is emerging: AI-powered tools for game makers. Our feature story explores this trend, focusing on the rise of AI co-pilots in game creation.

The Rise of AI Co-Pilots

When we think of AI in games, we tend to think of improving the gameplay – more intelligent enemies, more realistic environments, or personalized content. However, another ‘wave’ is gaining momentum and is focused on solving problems present in software development itself.

Stack Overflow’s data from the 2022 developer survey reveals that 63% of the respondents spend more than 30 minutes daily searching for answers or solutions to such problems. This indicates that one would have spent 333 to 651 hours per week searching for answers on behalf of a team of 50. This inefficiency reduces the output and small-scales the overall project performance to the extent of compromising developers’ morale and productivity.

AI co-pilots are built for these inefficiencies and are designed to alter how developers code. A significant milestone was reached in June 2021 when GitHub launched a preview of its AI-powered programming aid, GitHub Copilot. This tool, which uses AI to suggest code when a developer starts typing, quickly became popular. Microsoft's latest quarterly earnings report reveals GitHub Copilot's remarkable growth: 1.3 million paid accounts (up 30% quarter-over-quarter), adoption by 77,000 companies, and over 1.5 billion lines of code generated daily. Currently, Copilot produces more revenue than GitHub had in all of 2018, adding $2bn in annual revenue run rate to GitHub’s revenue and 40% of the service’s growth.

Research shows that AI services that enhance developers’ efficiency could grow the global economy by $1.5 trillion in 2030.

The Stanford AI Index Report emphasizes the influence on developer productivity and happiness:

• 88% of engineers report enhanced productivity

• 74% find themselves engaged in more rewarding work

• 88% noticed a marked increase in task completion speed

• 8% higher overall success rate in task completion

• Tasks are completed 56% faster (reduction from 161 minutes to 71 minutes)

AI co-pilots increase efficiency and improve job happiness, allowing engineers to focus on more challenging and gratifying areas of their work.

Investors and founders have noted all of these benefits. As depicted in the graphic below, the market for AI-powered development tools is quickly developing, with considerable investment going into both startups and established organizations.

With an overall funding of $2.73 billion across multiple ventures, it's evident that the race to produce the most effective AI co-pilot is on. Poolside, a prime example of this trend, recently announced a staggering $500 million fundraising to accelerate its progress toward artificial general intelligence (AGI). Poolside believes that software development will be the first field where AI outperforms humans, and they've created a new approach called Reinforcement Learning from Code Execution Feedback (RLCEF) to help AI improve its coding abilities. They aim to democratize software development, enabling anyone to create complex software without coding knowledge. Let's look at the basic capabilities of these co-pilots and how they might be deployed.

How does it work

Co-pilots for coding leverage machine learning models trained on large code repositories to generate real-time contextual suggestions. As programmers type, the AI analyzes the codebase and provides code snippets to help enhance and extend the written code.

Let's take Cursor, for example. It is a fork of Visual Studio Code that adds advanced AI capabilities to a familiar interface for most developers.

As you can see, Cursor can produce code on demand and even build an application’s entire framework. Developers can communicate with the AI in plain English about their code. Another fascinating VSCode fork is the Haystack Editor, a 2D digital whiteboard-based IDE that dynamically generates code connections as you navigate and edit files.

Game Engine Integrations

Integrating AI-powered editors into game engines is gaining momentum. For Unity, Prithvi Bisht's Medium article provides a step-by-step guide on how to use Cursor AI with Unity. Unity Code Assist Lite, a free Visual Studio plugin with over 50,000 installations, provides improved code completion, in-line visuals, and AI-powered support.

Moreover, there's also an in-engine plugin available on the Unity Asset Store that allows developers to control the Unity Editor using AI commands directly within the engine. This tool enables you to set up scenes, configure GameObjects, adjust project settings, generate scripts, and create textures simply by typing tasks. With features like full undo operations and detailed execution history, it significantly enhances workflow efficiency and flexibility for Unity projects.

Meanwhile, Unreal Engine developers are looking into tools like Bluepy, a plugin that uses OpenAI's API to produce Blueprint nodes from natural language inputs. Bluepy allows developers to describe the required functionality directly within the Blueprint graph, and the AI constructs the necessary nodes and connections.

But this is only the tip of the iceberg. Game creation involves unique challenges that differ from regular software development. These include complex physics, custom pathfinding, day-night cycles, and multiplayer networking. Large codebases can be intimidating for new team members, and tight deadlines necessitate quick onboarding and delivery while remaining consistent with the existing design.

Startups like Code Maestro are developing solutions beyond traditional co-pilots. These tools offer customized assistance for game-specific tasks, from AI behavior design to multiplayer debugging, while creating interactive knowledge bases to streamline onboarding.

Code Maestro's capability includes legacy code and custom engines, and local hosting provides data security, which is critical for protecting intellectual property.

However, the impact of these AI assistants on game production remains to be seen. We'll be keeping a careful eye on how these technologies fit into existing workflows and whether they can live up to the promise of more efficient development processes and creative gaming experiences.

Current Limitations and Future Potential

• GitHub Copilot's output often requires extensive testing and verification. A study found that 40% of AI-generated code contained vulnerabilities in high-risk scenarios.

• Developers like Darren Horrocks worry that relying too much on AI can lead to a misunderstanding of core ideas and encourage lazy development methods.

• Intellectual property and ethical considerations include the possibility of disclosing proprietary code and creating content that duplicates protected work.

• Tools like GitHub Copilot and ChatGPT lack the deep integration required for complicated coding activities, especially in game development sectors where platform-specific optimizations and interdependencies between code components are crucial.

• Research indicates that these techniques have variable success rates. ChatGPT generated accurate code for 65.2% of problems, followed by GitHub Copilot (46.3%) and Amazon CodeWhisperer (31.1%). Even the best-performing tool fails to produce accurate code in over one-third of circumstances.

• Furthermore, these tools' performance is strongly reliant on input quality. Without clear descriptions, their usefulness suffers dramatically.

On the other hand, OpenAI's new 'o1' series demonstrates considerable advancements in reasoning, maths, and coding abilities. These models outperform GPT-4 on complicated tasks, scoring 83% on International Mathematics Olympiad qualifying exams and achieving the 89th percentile in Codeforces coding competitions.

Conclusion: Balancing AI and Human Expertise

While AI models are improving and co-pilots help write some of the code, they do not replace the need for experienced coders. Human oversight is still required to ensure code correctness, maintain security, and manage large-scale projects. Developers that adopt and use these AI tools will have a significant advantage in an increasingly competitive sector.

As NVIDIA CEO Jensen Huang said:

Disclaimers:

This is not an offering. This is not financial advice. Always do your own research. This is not a recommendation to invest in any asset or security.

Past performance is not a guarantee of future performance. Investing in digital assets is risky and you have the potential to lose all of your investment.

Our discussion may include predictions, estimates or other information that might be considered forward-looking. While these forward-looking statements represent our current judgment on what the future holds, they are subject to risks and uncertainties that could cause actual results to differ materially. You are cautioned not to place undue reliance on these forward-looking statements, which reflect our opinions only as of the date of this presentation. Please keep in mind that we are not obligating ourselves to revise or publicly release the results of any revision to these forward-looking statements in light of new information or future events.

1 in 3 Quest owners have downloaded Gorilla Tag - making it the most successful game ever on standalone VR. Gorilla Tag, created by studio Another Axiom, has emerged as a standout success in the virtual reality gaming space, not just as a game but as a cultural phenomenon. Launched in 2021, it has quickly amassed over 10 million users and generated more than $100 million in revenue. While other games have seen success in VR, Gorilla Tag is unique in that it combines social interaction with a movement system that only works in VR. Much like other revolutionary free-to-play games, such as Fortnite and League of Legends, Gorilla Tag has carved out its own space by fostering a community-driven ecosystem that keeps over a million players coming back daily. Its rise also signals the emergence of a new sector of games—those built exclusively for VR—driven by the immersive experience only this platform can offer. At the heart of this success is the game’s social nature, compelling gameplay loop, and creative use of free-to-play monetization. However, being primarily played by young teenagers, it sits outside the cultural zeitgeist as a popular game and isn’t covered to nearly the same scope as other successful titles. So why is Gorilla Tag so popular?

Why Gorilla Tag is Successful

Whilst VR falls short against other gaming platforms in compute power, the medium has an over powered feature for engagement: Player embodiment.

Unlike flatscreen games where users control a character with a keyboard or controller while looking at a screen, VR players embody characters, making them feel as if they truly are the main character.

This means that spatially native social interactions are more akin to hanging out in real life, rather than zoom. The technical-social disconnect is heavily diminished.

Gorilla Tag is at its core a social space with tag as the engagement loop - similar to a poker night with friends or a pickup basketball game.

Gorilla Tag is designed with this in mind and is built to reflect the social nature of the game. Players spawn in a small hub inside of a tree. As soon as they step outside, they’re placed in a random social lobby and thrusted into the playground with a dozen others already there.

Voice chat is turned on by default and doesn’t require a press-to-talk action. Players speak into the world, and everyone in the server can hear them. The audio uses spatial proximity chat, where voices vary in position and volume based on players' physical locations relative to one another.

These circumstances foster natural camaraderie, as players form friendships with those they spend time with. They often connect on the Horizon platform (Meta Quest’s Operating System) and create private lobbies to continue playing together at different times.

One of the most intriguing observations in user interactions is the natural adherence to in real-life (IRL) social norms. While kids may spam words, groups often physically gather, form social circles, make eye contact, and focus on the speaker. However, if a conversation becomes dull, it’s not seen as rude to leave and start climbing branches. There's little dead social time or awkwardness. With enough people in a server, multiple subgroups often form, allowing players to hop between them or join a different server, given the ~50,000 concurrent users online.

A common critique among older Gen Z is the "death of the third place"—a space to gather outside of school/work and home. Social media has poorly filled this void, as authentic interaction feels fragmented and lacks the depth of real-life connections.

It’s clear that Gorilla Tag, predominantly played by boys aged 10-18, has become their third place. With 30% of MAUs being DAUs, one-third of monthly players are logging in daily, with server activity peaking after school hours.

Kids are drawn not just by the game itself, but by the chance to meet friends and play on a virtual playground.

To foster a social environment, users need to spend time in a dedicated space for interaction. While Gorilla Tag's average session time isn't publicly available, the Quest average is around 30 minutes—significantly higher than traditional social media. I estimate Gorilla Tag's session time to be even longer, similar to other successful VR multiplayer titles, around 60 minutes.

This would align with the average PC gaming session length of approximately 58 minutes (with the top 12% of players reaching 5 hours, making this a conservative estimate).

It’s important to note that these statistics reflect single sessions, and users often have multiple sessions in a day. However, Gorilla Tag excels in retaining users within a single social space, unlike social media, where users frequently jump between different threads or interactions.

Status and In-Game Purchases (IAPs)

The key question remains: how has a free-to-play game like Gorilla Tag managed to generate $100 million in revenue?

While the game is free on the Meta Quest Store, it is priced at $20 on Steam, though the vast majority of installs—over 90%—occur on the Quest platform.

Prior to the mobile gaming revolution, most games monetized through traditional models such as paid titles, downloadable content (DLC), or subscription services like World of Warcraft. However, mobile gaming was a pivotal moment for the free-to-play model, which has since evolved to encompass more complex monetization strategies, including microtransactions in AA and AAA titles.

With 10 million users and $100 million in revenue, Gorilla Tag boasts an impressive average lifetime value (LTV) of around $10 per player—an exceptional figure for a free-to-play game. But how exactly does Gorilla Tag generate this revenue? The answer lies in in-game cosmetics.

Given Gorilla Tag's emphasis on social interaction, players naturally seek ways to express their individuality and status within the game. Cosmetics offer a powerful tool for this, allowing players to personalize their characters in ways that reflect their personalities and achievements. This creates a sense of identity and social prestige within the community, especially when certain items become status symbols.

Within the game, there is a dedicated store area that players physically travel to, designed like a traditional clothing store with skeuomorphic elements. Hot and seasonal items are available on a rotating basis, often for a limited time, adding a sense of urgency to purchases. This setup mimics real-world shopping experiences and taps into scarcity-based consumer psychology, where players want to "flex" their unique or rare cosmetics to others in the game.

The in-game purchasing process is frictionless. Players can quickly browse and select items within the Gorilla Tag environment, with only one additional Meta OS confirmation pop-up to finalize transactions. By minimizing the steps needed to make a purchase, Gorilla Tag ensures that players can seamlessly acquire new items, making the buying experience smooth and unobtrusive.

As a result, players regularly purchase new items to keep their avatars fresh and distinctive. This constant cycle of cosmetic updates, combined with a user-friendly transaction flow, has proven to be a highly effective monetization strategy, leading to the game’s significant financial success, unseen before for VR titles.

Another VR title to surpass $100MM in revenue, Beat Saber, offers an initial content purchase along with additional level DLCs.

Here’s a video I filmed while visiting the store, where the kids in Gorilla Tag helped me choose, buy, and show off my bowtie.

In-game currency, Shiny Rocks, can be purchased with real money or earned incrementally by logging in daily.

Less friction —> More play time

One of the biggest hurdles for widespread VR adoption is the inherent hardware and user experience (UX) friction associated with accessing VR content. On the hardware side, users need a charged headset, a suitable play area, and the motivation to physically put on the headset. These challenges are universal across all VR applications, including Gorilla Tag.

However, Gorilla Tag addresses a different type of friction, specifically tied to UX design. A common complaint in VR gaming is the tedious process players go through just to start an experience, often involving 2D menus, loading screens, and multiple steps before entering the game world. This can feel clunky and counterintuitive for a 3D environment, and if players encounter too many barriers, they may lose interest or abandon the session altogether.

Gorilla Tag mitigates this friction by streamlining the player's journey from game launch to social engagement. When players start the game, they bypass traditional main menus and complex interfaces. Instead, they are immediately placed in the homeworld "tree," where they can seamlessly walk into a social lobby with no extra steps. This simple design significantly reduces the time between launching the game and participating in its core experience.

The entire process—from pressing start on the Quest home screen to being inside a social lobby—takes about 10 seconds. This near-instant transition eliminates much of the friction typically associated with VR gaming, allowing players to jump right into the action. By making the game loop accessible in seconds, Gorilla Tag removes one of the key frustrations of VR, contributing to its high retention and daily active user base.

Demonstration of going from putting on my Quest to being in a Gorilla Tag social lobby: ~12 seconds.

Unique to VR: Gorilla Tag's Groundbreaking Locomotion

When we think of the most talked-about and successful VR games, what do they all have in common? They share a design philosophy that leverages the core strengths of VR—specifically, the immersive tech stack that includes motion controls, six degrees of freedom (6DoF), and spatial awareness. These elements create experiences that are uniquely tailored to VR, making them impossible to replicate on traditional gaming platforms.

Gorilla Tag exemplifies this approach, offering an experience that is not only compelling but also fundamentally tied to the VR medium. At the heart of this is its innovative locomotion system, colloquially known as "Gorilla Locomotion." Traditionally, player movement in games has been controlled by joysticks or other digital inputs, where characters move in the direction indicated by the controller. Gorilla Tag, however, breaks this mold by requiring players to physically swing their arms to propel themselves through the game world.

Instead of relying on button presses, players use their real, physical arms (which are mirrored by their virtual arms) to push themselves in the desired direction. The only other control input is a grab function that allows players to cling to surfaces. This system is intuitive yet challenging, like learning to ride a bike. It feels awkward at first, but over time, players become more skilled and fluid in their movements. This design invites a more immersive form of play that’s impossible to replicate with traditional game controllers.

There is only a basic, self-serve tutorial to teach the fundamentals of movement. However, new players are often encouraged to ask for tips from more experienced players, fostering a sense of community and shared learning. Veteran players frequently assist newcomers, showing them how to master the game's core mechanics, such as swinging from branches or scaling trees.

This dynamic movement system sits in the "easy to learn, difficult to master" range, making it highly rewarding for players as they improve. It’s similar to genre-defining games like Counter-Strike or Rocket League, where mastery comes with practice and dedication. This rewarding feedback loop is a major reason why players keep coming back. Additionally, Gorilla Locomotion has spawned its own category of VR locomotion, incentivizing other developers to create movement systems that can only exist in virtual reality.

GT has designed a system that feels natural yet challenging and ensures that its gameplay remains deeply tied to the physicality and immersion of VR. This is not a game that could be ported to a traditional screen with a joystick—it’s fundamentally a VR experience, and it helps to keep player retention inside the game.

The Memeability and Social Media Presence of Gorilla Tag

One of the key drivers of Gorilla Tag's virality is its popularity across social media platforms. Its memeable nature and easy-to-share gameplay have made it a hit among content creators, particularly on platforms like TikTok, YouTube, and Twitch. The game’s distinctive art style and chaotic, physical movement make for highly entertaining clips, leading to millions of views and shares. The rise of user-generated content, from memes to full-length gameplay videos, has contributed to the game's growing community, with players often showcasing their skills, humorous interactions, or unexpected moments in lobbies.

In particular, Gorilla Tag lends itself to the type of organic, spontaneous content that social media algorithms favor. Videos of players flailing their arms to move, dramatic chase sequences, or hilarious voice chat moments quickly go viral, enticing new players to join in on the fun. Creators across different platforms have built entire followings based on Gorilla Tag content, further embedding the game into the social media fabric. As a result, the game benefits from constant word-of-mouth marketing, with each viral clip acting as free advertising. In an era where games often live or die by their presence on platforms like TikTok, Gorilla Tag has mastered the art of staying relevant by being inherently shareable.

Gorilla Tag’s Impact on the Future of VR Gaming

Gorilla Tag has not only proven to be a financial success but has also set a new standard for what VR games can achieve. Its innovative approach to player movement, its seamless integration of social interaction, and its free-to-play model are shaping the future of VR gaming. In an industry where hardware limitations and user friction are common barriers, Gorilla Tag has shown that a well-designed game can break through those challenges and create a lasting impact - divorcing away from the idea that VR games are not commercially scalable.

As VR technology continues to evolve, the lessons from Gorilla Tag's success influences the next generation of VR titles. From its unique locomotion system to its organic social spaces, GT has set the precedent on how VR can offer experiences that are impossible on other platforms. In doing so, it has helped usher in a new era of social VR gaming, recentering the definition of social.

With its ongoing updates, expanding player base, and cultural resonance, Gorilla Tag is more than just a game—it’s a movement. Its success shows that virtual reality has the potential to transform gaming by offering players a truly immersive, social, and unforgettable experience.

The VR social renaissance is soon in full swing - and Gorilla Tag is only the first of many more success stories to come. We are here to find the next.

Disclaimers:

This is not an offering. This is not financial advice. Always do your own research. This is not a recommendation to invest in any asset or security.

Past performance is not a guarantee of future performance. Investing in digital assets is risky and you have the potential to lose all of your investment.

Our discussion may include predictions, estimates or other information that might be considered forward-looking. While these forward-looking statements represent our current judgment on what the future holds, they are subject to risks and uncertainties that could cause actual results to differ materially. You are cautioned not to place undue reliance on these forward-looking statements, which reflect our opinions only as of the date of this presentation. Please keep in mind that we are not obligating ourselves to revise or publicly release the results of any revision to these forward-looking statements in light of new information or future events.

Imagine a world in which video games are not only played, but dynamically created and shaped by AI in real-time. As a gamer and investor, I've spent months researching and discussing the possible impact of Generative AI on the gaming industry. The market map below highlights the leaders who are driving these developments, each with their own set of tools that are changing the way games are produced and enjoyed. In this article, I'll showcase some of the best demos from those startups, highlighting the latest advances in this field.

To fully understand the significance of this AI revolution in Gaming, we need to first understand the issues that the industry is suffering from. Open-world games, for example, sell unparalleled freedom and immersion yet often fall short because of repetitive dialogue, predictable behavior, and limited interactions.

This is evident in scenes like players queuing to complete quests in World of Warcraft Classic, highlighting how traditional design struggles with large-scale player interactions.

Creating convincing NPCs and detailed environments is extremely time-consuming. Game writers create massive dialogue trees, sometimes with thousands of lines for a single character. Red Dead Redemption 2 features around 300,000 animations and 500,000 conversation lines, with even small NPCs having 80-page scripts. Even with substantial investment — Rockstar Games spent over 100,000 hours on dialogue recording — NPCs remain constrained by pre-written scripts, unable to truly react to unexpected player actions or engage in dynamic conversations. This limitation can break immersion and reveal the game world's static nature.

To address these challenges, developers have been exploring various solutions — with procedural generation as a key technique.

Procedural Content Generation uses algorithms to automatically create game content, combining human designed assets with computer randomness. This enables games like No Man's Sky or Minecraft to create nearly infinite worlds.

Another example is Tiny Glade, which started as a procedural wall generator built with the Bevy engine in Rust. It has surpassed 1 million wishlists and attracted 357,000 players for its demo, demonstrating how games offering creative tools and personalized experiences can resonate strongly with audiences, even without the scale of larger open-world titles.

Similarly, platforms like Roblox have seen remarkable growth, with daily active users surging to 79.5 million and engagement hours reaching 17.4 billion in Q2 2024.

However, procedural generation has its limits, often resulting in repetitive or predictable patterns and lack of meaningful narrative structure. Additionally, the diversity of procedurally generated content is constrained by the assets and rules the developers provide, limiting the potential for truly unique creations.

This is where generative AI emerges as a potential solution to bring this level of content richness and dynamism to traditional game formats.

The Unity Gaming Report 2024 reveals that 62% of game developers are already leveraging AI tools, with an additional 29% planning to incorporate AI into their development processes.

In May, we explored how developers are using these technologies to expand game content, with early signs of real-time creation tools starting to emerge.

NVIDIA's recent GenAI demo at SIGGRAPH 2024 showcased real-time world-building capabilities. Their NIM microservices for OpenUSD enable quick creation and assembly of 3D assets and scenes using AI and prompt engineering.

This technology points towards a future where game environments and assets are generated and adapted in real-time, with AI-driven, culturally tailored content that can be easily localized for different markets.

In March 2024, NVIDIA CEO Jensen Huang predicted at the GPU Technology Conference that fully AI-generated games could emerge within 5-10 years, potentially replacing traditional graphics rendering with AI-driven content generation.

While this vision may seem distant, Unity is already laying the groundwork with its Sentis neural inference engine. It allows developers to run AI models directly on users' devices, bypassing the need for cloud servers and ensuring faster response times and reduced latency.

This capability opens up a new realm of possibilities for creating dynamic, responsive, and personalized gaming experiences. With Sentis, game developers can implement real-time AI-driven features like adaptive gameplay, intelligent NPCs, and on-the-fly content generation. As Unity leads the way with practical tools, academic research is pushing the boundaries even further.

Runtime Behavior Generation in Games

In early August 2024, a master's research project from the University of California, Berkeley introduced GROMIT, a novel system for Runtime Behavior Generation (RBG) in video games. The project explores the use of Large Language Models (LLMs) to generate game behaviors in real-time based on player input. GROMIT demonstrates this capability through three distinct demos: a Sandbox where players can directly manipulate the environment (e.g., making an apple float or shrinking buildings), an Escape Room where object interactions are generated on-the-fly (such as burning a bookshelf with a torch), and an Adventure Game where players can combine spells to create new ones.

The research showed that GROMIT could successfully generate behaviors in 70% of cases in a controlled environment, with some generated behaviors having significant impacts on gameplay, like creating alternative solutions to puzzles. However, it also highlighted challenges, including a lower success rate (37%) in more complex, pre-existing game environments not designed for such manipulations. The study emphasizes the need for additional guardrail systems to ensure appropriate and controlled behavior generation, pointing towards a future where games could become more dynamic and personalized.

One promising example of this future direction is Determin AI's TREK platform. It is an AI Reality Engine that aims to revolutionize interactive experiences by generating millions of real-time interactive video bots.

Users will be able to watch, interact, and play with their favorite game characters and fictional worlds, as well as request new bots.

While GROMIT and TREK showcase the potential for dynamic, player-driven content generation, another crucial aspect of creating immersive, responsive game worlds lies in the behavior of non-player characters (NPCs). As games evolve to incorporate more real-time generated content, the role of NPCs in creating believable and interactive environments becomes increasingly important.

NPCs

Several companies are advancing significantly in their use of runtime generative AI to produce more interactive and dynamic characters.

Convai, for example, has shown off the possibilities of runtime produced actions by demonstrating how NPCs may react more quickly and flexibly to player interactions.

Another key player in this space is Inworld AI. Their approach centers on creating autonomous AI agents that game developers can direct and control.

Key features include the Goals and Actions system, which allows AI agents to act on developer-defined motivations triggered by player or environment interactions. These advancements make it possible to create NPCs with more depth and believable decision-making processes.

Taking NPC interaction to the next level, Ubisoft's ZeroEGGS generates stylized character gestures in real-time based on speech input. The system generalizes to new styles, voices, and languages not seen during training, creating diverse and responsive characters on-the-fly.

In user studies, it outperformed baseline models in naturalness, appropriateness, and style portrayal. This tool paves the way for more dynamic, personalized game worlds where NPC behaviors and animations adapt in real-time to player actions and preferences. Complementing these advancements in NPC behavior, innovative tools are emerging to address the challenge of dynamic asset creation.

Object Generation

Sloyd allows for on-the-fly creation of unique models based on player input or descriptions, bridging the gap between player imagination and in-game reality.

Imagine a player describing a sword with specific features to an in-game weaponsmith. Using Sloyd, the game could generate that custom weapon in real-time, bringing the player's vision to life. This technology leverages AI trained on a vast component library to interpret prompts and assemble models accordingly. This empowers developers to create more interactive and personalized game worlds, where players can shape their environment and items beyond pre-made assets.

World Building

As game developers seek to create ever more expansive and detailed environments, tools that can assist in crafting entire worlds become increasingly valuable.

Moonlander is one of the first AI-powered copilots for virtual world building, helping developers construct vast, detailed game environments with ease and speed.

Taking this concept even further, Playo.ai aims to generate fully playable, hyper-personalized 3D games at an unprecedented scale and cost-efficiency. The company is building a foundation AI model to build games on demand for people. If players want a unique game to experience (“Barbie racing in the street of Ramallah”), it will create a game in a few seconds that can be played anywhere.

The mission is to simplify game development, making it possible for anyone to transform their game ideas into reality effortlessly.

FRVR's AI-powered game creation tool, currently in public beta, enables both experienced programmers and non-coders to bring their game ideas to life. The platform streamlines the entire process from ideation to publication, allowing users to create games through simple prompts and customizable templates.

With just one click, creators can release their games on various platforms, including mobile app stores, Facebook Instant, and Microsoft Windows. Published games have been played by over 1.5 billion players globally, with some of those titles drawing in over 100 million monthly players. It demonstrates the potential for quick, AI-driven content creation. Future versions could potentially integrate into games, allowing for runtime generation of new levels, challenges, or game modes based on player preferences or actions.

Animation and Rigging

One of the most labor-intensive, costly, and skilled aspects of making a game is making great animation. Anything World has recently launched their "Animate Anything" plugin for Roblox Studio. The plugin allows users to upload their own 3D models and have them automatically rigged and animated within minutes, dramatically reducing the technical barriers to creating dynamic in-game assets.

In a similar vein, Replikant has been showcased in a presentation at SIGGRAPH 2024 as part of the Nvidia GenAI talks. The platform enables the rapid creation of high-quality 3D animations and interactive, no-code chatbots.

As we move from bringing characters and objects to life through animation, another crucial element in creating immersive game experiences is the auditory landscape.

Sounds & Music

Unlike current pre-recorded samples, AI models could produce diverse, contextually appropriate sounds on the fly. For instance, footsteps could dynamically adapt to surface types, character weight, and movement speed, creating a more realistic and varied soundscape without repetition.

ElevenLabs has recently introduced Text to Sound, which can generate a wide range of audio content, including sound effects, short instrumental tracks, soundscapes, and various character voices, all from simple text prompts.

Splash's AI music platform has gained traction, particularly through its integration with Roblox. The "Splash" game on Roblox has attracted over 21 million players, allowing them to create and perform music in-game. Splash's AI can compose, sing, rap, and play instruments, making music creation accessible to a wide audience.

This technology opens up possibilities for dynamic, personalized soundtracks in games. For runtime generated games, such technology could enable adaptive music that responds to player actions, environmental changes, or story progression, enhancing immersion without the need for extensive pre-composed soundtracks.

Ethical and Technical Considerations

One of the primary challenges facing generative AI applications across industries is the acquisition of high-quality, diverse, and ethically sourced training data. Many AI models have faced criticism for using copyrighted or personal data without proper consent, leading to legal and ethical concerns. For instance, image generation models like DALL-E and Midjourney have been scrutinized for potentially training on copyrighted artworks, while large language models have raised questions about the use of personal information scraped from the internet. Recently, NVIDIA has come under fire for allegedly scraping videos from YouTube and Netflix without permission to train its AI models, according to a report by Engadget.

In contrast to these contentious practices, Activision Blizzard has taken a notably different approach. Their recent release of the Warzone map Caldera as an open-source dataset represents a significant and proactive step in addressing AI data challenges within the gaming sector.

By openly sharing this high-quality data, Activision Blizzard is not only supporting advancements in game-related AI applications but also setting an example for ethical data sharing which also leads to more sophisticated content creation tools.

It remains to be seen whether other major gaming companies will follow suit. Nonetheless, this move demonstrates a possible path forward in balancing the need for robust AI training data with ethical considerations.

Summary

The future of gaming will likely undergo an important shift, with predictions indicating that fully AI-generated games will be available within the next five to ten years, thanks to improved GPUs that enable real-time creation of content. This approach has the potential to revolutionize various genres, from role-playing games (RPGs) to massively multiplayer online games (MMOs). One example: AI-powered non-player characters (NPCs) will engage in dynamic, unscripted interactions, greatly enhancing player immersion.

Furthermore, we believe that removing entry barriers to game production will encourage a new creator economy, enabling individuals to create high-quality games with minimal resources. This trend might give rise to 'one-person unicorns,' in which creative artists capitalize on the latest AI tools to bring their ideas to life. As a result, we also expect the average age of developers to fall dramatically, allowing in a new generation of innovation.

Disclaimers:

This is not an offering. This is not financial advice. Always do your own research. This is not a recommendation to invest in any asset or security.

Past performance is not a guarantee of future performance. Investing in digital assets is risky and you have the potential to lose all of your investment.

Our discussion may include predictions, estimates or other information that might be considered forward-looking. While these forward-looking statements represent our current judgment on what the future holds, they are subject to risks and uncertainties that could cause actual results to differ materially. You are cautioned not to place undue reliance on these forward-looking statements, which reflect our opinions only as of the date of this presentation. Please keep in mind that we are not obligating ourselves to revise or publicly release the results of any revision to these forward-looking statements in light of new information or future events.

At the core of VR's appeal is its ability to create a sense of presence – the feeling of truly being inside a virtual world. Physical activity plays a pivotal role in enhancing this presence. When players use their bodies to interact with the virtual environment, whether it's swinging a sword, ducking behind cover, or reaching out to grab objects, the immersion delineation between reality and the virtual world is depreciated more than any other gaming experience.

For the last 9 generations of game consoles, improvements have been dictated by moore’s law - more general computing capacity for a similar price/form factor for each successive generation, usually launched about every 5 years.

Improvements include rendering speed, graphics capacity, online connectivity and more detailed environments.

The initial launch of consumer viable VR headsets in the mid-late 2010s saw the first divergence of this principle - where a new medium was more computational expensive than flat screen equivalents, making VR games sacrifice the evergreen principles of graphical improvement.

This trend has become more exacerbated at the transition from PC powered VR to standalone - meaning all of the devices computing power is built into the headset, rather than delegated to a dedicated computer.

The standalone Quest 3’s computing power is roughly equivalent to Nvidia GTX 960, a mid level GPU from 2015.

Whilst not winning a compute war, VR benefits from a bespoke and exceptional feature - 6 degrees of freedom, and the immersion of player embodiment.

The hallmark tenant for building a VR game is to make it best & only possible in VR.

If the core loop and design doesn’t utilize the unique features, then it must fight and compete with attention on high compute, lower friction devices (PC & console)

The best of these games really utilize mechanics that make it where it would be near impossible to replicate without a VR setup.

Games that achieve embodiment and physical activity

Note, while these titles brought about unique-to-vr movement components, the success of these titles are multifaceted.

Beat Saber

Beat Saber was one of the first viral successful games for VR - a rhythm game where players hit moving music notes propelling towards them.

Rhythm isn’t a new category - games like Rockband and Guitar Hero dominated the 2000s with a similar concept of scoring notes with timing precision.

Beat saber brought this concept to VR. Not only does the game take place in 3D space, but utilize physical player movement to play. Users have to strike blocks moving their physical hands, and have to physically move sideways or downwards to avoid obstacles going towards them.

This has made the title incredibly popular, and a surprisingly good workout. Most notably, it validated the idea that top experiences in VR are unique to VR.

Superhot

Superhot VR exemplifies the "best & only possible in VR" principle, while still being derived from an original flat screen release.

The game's core mechanic - time progresses only with player movement - is brilliantly adapted for VR. This creates a unique mix of strategy and action intrinsically linked to the player's physical movements. Every slight gesture advances time, generating an intuitive connection between player action and game progression that traditional controllers can't replicate.

Superhot VR leverages VR's capacity for 360-degree awareness, requiring players to constantly survey their surroundings. This spatial interaction, combined with natural, gesture-based controls for actions like punching, throwing, and shooting, enhances immersion in ways button inputs literally can't match.

The game's core loop, paired with its physical interactions, generates a powerful sense of presence. Players find themselves fully immersed in a surreal, time-bending world, manipulating time and space through their own movement - making the 2D version pale in comparison.

Gorilla Tag

The most popular VR game to date (with an article coming out soon digesting the reasons for it’s success) Gorilla Tag stands out as the prime example of a game that fully embodies the "best & only possible in VR" principle.

While Gorilla tag’s core driver is its social playground, its brilliance lies in how it leverages VR movement to create a unique locomotion system that would be impossible to replicate on traditional gaming platforms.

Players move by using their arms to propel themselves, mimicking the movement of gorillas. This innovative approach to locomotion creates a physically engaging experience that's intrinsically tied to VR's motion controls. The game transforms the player's entire body into a controller, requiring them to swing their arms to move and climb, creating a sense of embodiment that's uniquely possible in VR.

The game's multiplayer aspect further capitalizes on VR's strengths. Players must be constantly aware of their surroundings, looking around to spot other players and using the full range of motion afforded by VR to dodge and chase. This spatial awareness and physical engagement create a social experience that's both intuitive and deeply immersive.

As aforementioned, the success of these titles are greater than just their unique-to-vr movement mechanics, but it is a defining pillar to their success.

In essence, the future of VR gaming lies not in competing with traditional platforms on graphical prowess, but in pioneering experiences that are uniquely suited to the medium. By embracing VR's strengths - immersion, physical interaction, and spatial awareness - developers are creating games that justify the technology and offer players something genuinely novel.

As the industry continues to evolve, the most successful VR titles will likely be those that fully embody the "best & only possible in VR" principle, pushing the boundaries of what gaming can be and redefining consumer expectations of interactive entertainment.

Disclaimers:

This is not an offering. This is not financial advice. Always do your own research. This is not a recommendation to invest in any asset or security.

Past performance is not a guarantee of future performance. Investing in digital assets is risky and you have the potential to lose all of your investment.

Our discussion may include predictions, estimates or other information that might be considered forward-looking. While these forward-looking statements represent our current judgment on what the future holds, they are subject to risks and uncertainties that could cause actual results to differ materially. You are cautioned not to place undue reliance on these forward-looking statements, which reflect our opinions only as of the date of this presentation. Please keep in mind that we are not obligating ourselves to revise or publicly release the results of any revision to these forward-looking statements in light of new information or future events.

Google and Facebook's success was largely due to web analytics. As Web3 emerges, there is a much greater demand for apps that can not only make sense of massive amounts of data but also customize user experiences and improve advertising for companies and developers alike. 

Despite the benefits of social graphs and targeted advertising in Web2, Web3 continues to work on the technology required to successfully filter and capitalize on massive volumes of on-chain user data to improve user engagement.

One such company that we believe is at the forefront of this monumental shift is Cookie3. Put simply, they develop tools that optimize the interaction between decentralized applications (dApps) and their users. Benefitting both businesses and consumers.

For businesses: Cookie3 Analytics is an advanced tool for dApp developers, providing transparent insights into user interactions and wallet activities. This helps developers deeply understand their audience and improve the user experience while maintaining a seamless integration that respects user privacy.

For affiliates: Cookie3's Affiliate platform allows Affiliates to join campaigns and receive automated rewards, simplifying the process and ensuring fair compensation.

For consumers: Cookie3 Score unveils a MarketingFi Score that evaluates a user's engagement with marketing initiatives and dApps. Similar to a credit score, it anticipates the likelihood of future rewards, thereby motivating continued engagement and offering users a way to capitalize on their online presence.

Enhancing Security with Cookie3 Airdrop Shield

Airdrops are popular in Web3 marketing, although they pose issues such as vulnerability to bot attacks and Sybil schemes. Cookie3's Airdrop Shield addresses these problems in its entirety. It uses strong off- and on-chain analytics to screen out troublesome players, ensuring that airdrops are sent to actual, interested customers. This preserves the integrity of marketing campaigns and protects key project resources.

On Track to reach 500 business users and $1M ARR by 2025

In just one week, ending March 2024, Cookie3 rapidly ascended as the largest Key Opinion Leaders aggregator in Web3, enlisting over 17,000 KOLs with a collective following exceeding 295 million, featuring notable figures like IvanOnTech, Altcoin Daily, BoredElonMusk, and Jihoz, all without incurring marketing costs. The platform has attracted more than 170 business clients. It supports 16 blockchains, tracked over 80 million visits through its Analytics Plugin, processed over 9 billion transactions with its Tracking Engine, and developed advanced AI models analyzing off-chain and on-chain behavior.

A Community-Powered Ecosystem

Cookie3's innovation, in partnership with the Cookie DAOcenters on the DAO’s $COOKIE token, the foundation for the MarketingFi value and reward system. This token enables equal distribution and incentives throughout the community and encourages participation. Holders of $COOKIE are more than simply investors; they actively participate in the Cookie DAO, influencing the platform's development and making critical choices.

Cookie3's collaboration with Polygon Labs enhances the $COOKIE Ecosystem through Cookie3 Analytics, providing AI-powered data insights for smarter decisions and handling over 1 billion wallets and 14 billion transactions. Additionally, the project is supported by a think tank including Tal Cohen, a Web3 expert with experience at companies like McKinsey, EY, and Google, ensuring all initiatives are backed by deep industry knowledge.

Successful Token Launch

With a $52M trading volume on its listing day and over 10K holders, the 60-day $COOKIE first movers campaign promotes early participation with perks like 120% APY staking, multi-token farming, and exclusive referral programs. The token has successfully launched on exchanges like Bybit, MEXC, Kucoin, Bitget, BingX, and Gate.io, with KOL rounds oversubscribed by over 4500%.

Our investment in $COOKIE, which has already raised over $5.5 million in partnership with industry leaders like Spartan Group, CMT Digital, and the accelerator ChainGPT, positions Hartmann Capital to leverage the projected growth of the digital marketing sector to over $650 billion by 2028. We believe in $COOKIE’s potential to lead this expansion with Cookie DAO and Cookie3, transforming web3 marketing with transparency, efficiency, and community engagement.

Share Hartmann Capital Research

Disclaimers:

This is not an offering. This is not financial advice. Always do your own research.

Our discussion may include predictions, estimates or other information that might be considered forward-looking. While these forward-looking statements represent our current judgment on what the future holds, they are subject to risks and uncertainties that could cause actual results to differ materially. You are cautioned not to place undue reliance on these forward-looking statements, which reflect our opinions only as of the date of this presentation. Please keep in mind that we are not obligating ourselves to revise or publicly release the results of any revision to these forward-looking statements in light of new information or future events.

Crypto’s Six Year Rise out of Obscurity

6 years ago, Hartmann Digital Assets Fund had its initial close. We (or more properly ‘I’ at the time, as a twenty-three-year-old one-man show) came onto a very novel scene with perhaps no more than 100 crypto funds ever having been in existence. It was easy to know everybody and everything. There were only a handful of OTC desks like Circle and Cumberland, one or two custodians like Kingdom Trust, a singular bank willing to back the space (Silvergate), and virtually, and I mean this earnestly, zero actual products or users outside sending and receiving BTC. The industry was almost exclusively built on vision and ideals. This was an era before DeFi, DePin, Web3 Gaming, and everything that has been built in recent years. In fact, most of the ‘basic’ financial primitives that are taken for granted daily, like on-chain swaps, lending, and borrowing, were virtually non-existent—so much so that in that year all of global DeFi was banking approximately $300k. This number, for context, has since risen to north of $100bn.

2018 was also the year when BlackRock’s Larry Fink said that he can’t think of a single client on the globe that has asked for or wanted crypto exposure. Today, BlackRock’s Bitcoin ETF recorded the most successful ETF launch in history, hauling in $20 billion in under 5 months.

If we were to shrink down the entire trading history of Bitcoin and ETH into just 10 days, I’ve traded and witnessed around 5 for Bitcoin and 8 for ETH. Every candle and wick tells a story I fondly remember. Equally, being on the younger side of CIOs/GPs with a 6-year digital asset track record, this crypto journey already makes up 25% of my life in years or 64% of my adult life. Being in a nascent industry for this long, and as I like to say, growing up both in it and alongside it, has shaped my perspectives profoundly.

The industry has gone through many phases on this shared path to maturity. A fairly homogeneous ideal-driven community of libertarians has splintered into many subcultures—from the suit-wearing digital asset crowd pushing for institutional adoption, to the meme-coin casino millennial who lives on Twitter, and finally the original missionary crowd pushing ideals of privacy and decentralization while regularly being targeted by governmental agencies. The first has become my job, the second makes me question my job, and the third is why I get over the stench of the second and remember what brought me here in the first place.

The Regulatory Bottleneck is Heating Up… and is About to Burst

While the industry continues to build, particularly on the scaling front, innovation and adoption have been hamstrung amidst the regulatory battlefield in the US. It’s been nearly a full year since I published my letter on the regulatory bottleneck. While Ripple and Grayscale won their respective cases, and the Bitcoin ETF got approved as a result, as predicted, the regulatory war has not been won yet. The SEC has continued its crackdown on the industry’s most illustrious participants as recently as April with their Wells notice to Uniswap Labs, a US-based unicorn company that is used by millions and has not stolen a penny from anyone. Instead of going after the countless scams and frauds prevalent both in finance and crypto, they chose to go to war with pillars of the industry like Coinbase, Kraken, and Uniswap. It’s not a war for ‘investor protection’; it's a war against technology.

This war against technology has had real repercussions. A few challenges come to mind when thinking of how this current toxic regulatory environment has led to an erosion in fundamentals in the industry:

1. Unless a builder is willing to rot in jail for their mission, many are steering clear of crypto. Running a startup is hard enough; the last thing you need is the Feds threatening your liberties.

2. Regulatory pressure is unevenly skewed towards legitimate products. In other words, you are likely safe from the SEC if you straight up scam someone with a pump-and-dump meme coin, but God forbid you build a new technology that may actually change the financial industry. As a result, we are seeing more worthless crap being launched than truly valuable tokens that try to capture and return value to their holders.

3. DeFi has not really grown in 3 years. While we hit a mind-blowing $100bn in assets being banked on decentralized rails (down from $180bn in 2021), growing materially past this mark will require institutional capital, which is not going on-chain until there’s regulatory clarity.

The Digital Asset trade did not require regulatory support to climb from the innovators to the early adopters. However, the next leg of the Digital Asset trade, to go from Early Adopters to Early Majority, very much needs either regulatory support or a collapse of the old system.

Remove the time horizon, and there is only one correct trade to take: being long on decentralization and shorting the fiat system, institutional inertia, and any nation suffocating at the hands of bureaucracy.

Running a liquid fund, however, time horizons matter. The AI industry, for example, while today evidently shattering records in adoption speed, suffered through two multi-year winters in the '70s and '90s1. The last few months, as Bitcoin struggled to defend its all-time high, I’ve spent a tremendous amount of time thinking about whether or not the industry would win this regulatory war in the West or be shelved until a more friendly (or as the media likes to call it—radical) administration takes over.

But then, things started changing in the last three weeks… As former (and currently polling future) President Donald Trump publicly threw his support behind the digital asset industry2, we started seeing a sudden tone change in DC:

1. The Ethereum ETF, against all expectations, got streamlined for a rushed approval by the SEC.3

2. Congressional Disapproval for the SEC's harshly restrictive crypto custody rules under SAB 121 passed with votes from countless formerly crypto-hostile Democrats.456

3. The Financial Innovation and Technology Act, vastly welcomed by the digital asset community for providing more regulatory clarity and limiting the SEC's jurisdiction over the asset class, passed the House of Representatives and made its way to the Senate. 78

4. Noticing the sudden influx of support for the Trump campaign from the digital asset community, the Biden campaign suddenly began their own outreach to the community according to The Block. Biden lost all favors with the community days later, though, when he vetoed Congress’s repeal of SAB 121.

Ending the Gensler Crusades

To get a better look at the situation, two weeks ago, I found myself in the same room as most of the industry’s policy and/or leadership teams at the annual CoinCenter Gala in Austin, Texas. The energy was rather elated, with speakers acknowledging that “we are finally starting to win” in light of the recent victories in DC. House Majority Whip Tom Emmer greeted the audience, too, with a warning to SEC Chairman Gary Gensler: “Hear my voice, Gary - Your days are numbered.” As I was sitting there, I asked myself: “Is it really the time?” 

The presidential election may just be the hammer that shatters the regulatory bottleneck. If Trump wins, expect Gensler to be resigned at worst or in jail at best. Beyond the autocratic abuse of power in his role as SEC chairman (which is now more frequently pointed out by courts), Gensler also used to serve as Hillary Clinton’s campaign CFO, which brands him as a foe to the Trump administration.

If Biden wants to win, canning Gensler may be his most straightforward path to winning the crypto vote and distancing himself from the radical left, particularly Elizabeth Warren, who has made it her mission to "build an Anti-Crypto army”. This dissent from Warren, who has served as the de-facto voice of the Democratic party on all things crypto, was visible in the vote on FIT 21, where even former Speaker of the House Nancy Pelosi voted in favor alongside Republicans.

I now place the odds of an end to the Gensler Era in the next 6 months at 70-80%. The day that happens, I expect the global digital asset market cap to appreciate by 30%+… or in other words, around a trillion.

Should the dice fall on the remaining 20-30% (or my estimate just be completely off), and we get another four years of Biden and another 2 years of Gensler before his term elapses, I would place my bets on a continual erosion of what Capo Gensler likes to call ‘crypto asset securities.’ The good guys will lose, and the only winners will be law firms collecting checks from crypto firms and US taxpayers as they go head to head for another four years.

Two camps of assets, however, remain mostly sheltered from this crackdown: officially deemed crypto commodities like Bitcoin and meme-coins that do not even attempt to generate value and are born as literal parodies yet evolve into being proxy trades on the cultural Zeitgeist. These two asset camps will continue to grow due to a mix of fiscal irresponsibility of central banks and financial nihilism of Gen Z and millennials. A treatise on exploring these markets is in the works too but will have to wait for another time.

A New Era in Digital Assets

The most meaningful graduation from being speculative assets to investable assets for the wider digital asset space beyond Bitcoin’s digital gold thesis occurs the moment tokens can accrue value without being at odds with regulators. While part of the debate is whether or not digital assets are commodities or securities, the other meaningful debate is how crypto assets can be compliant in a way that does not go against the underlying technology. For example, requiring KYC of each and every holder is a non-starter when the underlying principle is privacy and permissionlessness.

In the event of regulatory clarity, the digital asset market will likely transform in ways that unlock the greatest bull market yet. A few predictions stand out:

1. A shift from narratives to product market fit

   Since there is currently no path for crypto assets to compliantly accrue value, most crypto asset issuers do not even bother creating products that capture value. Ironically, the ability to capture value is a great litmus test in determining whether or not a product itself is genuinely demanded enough for consumers to part with their hard-earned money. Instead, crypto founders have often built things that consumers care so little for, they had to pay their users in tokens to use them. So a few things happen. The quality of building goes up, and...

2. Projects will have more clear metrics to measure success by

   Currently, many digital asset valuations appear to be free-floating numbers set purely based on sentiment and comps. While most markets certainly aren’t efficient, as even equities often trade far detached from their earnings, the equity market does do a fair job in raising the cream to the top. As such, the tokens with the most material product market fit and earnings are likely to start dominating both conversations and portfolios more frequently. This in turn leads to...

3. Easier funding environment for digital assets

   Funding in digital assets is most heavily skewed towards the private markets, while the ability to raise post token launch tends to turn into a roll of the dice for founders based on what market regime they find themselves in. This has caused a cyclical rise and fall nature of ‘alts,’ where each new cycle brings with it a cohort of new projects that raise fantastic rounds while private, and often run out of money or fail to properly capitalize the next bear market, sometimes even if they actually built a fine product. Then the private markets rotate to the next cohort. There is a fair amount of duplicate costs and value being discarded through this rotation. As such, stronger fundamentals will enable protocols to raise funding more easily while at the same time enabling...

4. A thriving M&A market

   Throughout 2022-23, we witnessed numerous DeFi projects fall by the wayside that could have been prime acquisition targets for better-funded DeFi projects. For example, a well-funded Uniswap or fairly well-capitalized AAVE may expand its offering and become a DeFi super app by acquiring some of the many well-functioning but under-capitalized players in the on-chain perps and options markets, or break more materially into real-world assets by facilitating a token swap with one of the leading real-world-asset (RWA) protocols that are trading at perhaps 1% the market cap of Uniswap. The maturation of crypto assets individually and of the market as a whole is likely to open the doors for truly savvy dealmakers and operators to build value in ways we have not properly seen executed before in Digital Assets and materially accelerate product development and innovation, which in turn leads us closer to adoption. For example, some Layer 1 blockchains are likely to utilize M&A to acquire highly desired products and turn them into public goods. This would reduce costs for users, while at the same time increasing usage and gas spend on the chain itself, driving value for the network token (the fat protocol thesis sends its regards). 9

Having traded catalysts and fundamentals in digital assets since the day they essentially arrived in the industry, this is perhaps the most promising era of this market that is yet to come. Such structural changes would also likely see the largest inflows in capital into the market as institutional allocators can apply similar models in finding true value as they have for a century in other asset classes.

At Hartmann Capital, our core digital asset thesis is that ‘Most Assets will be Tokenized’ in the coming decade. And while this was a contrarian take, Larry Fink and even Jamie Dimon (who can’t stand Bitcoin), are fond proponents of tokenization. While most of these leaders are focusing on the asset that could be wrapped in a token and the opportunities and savings it would unlock, we want to focus on and own a stake in the infrastructure that powers this whole process and could capture the fees of what will likely, in the endgame, be the market of all markets—from financial, to physical assets, to IP, and personal data.

Lighting the Fuse

The timing of the US presidential election, and thereby the regulatory fate of the digital asset space, happens to coincide with the looming central bank pivot. While the ECB took the first move in cutting interest rates, the Fed has yet to make its first cut, now projecting that it will cut rates once in 2024, with rates dropping to ~4% in 2025 per their June Economic Projections.

While the Fed has successfully reduced its balance sheet by approximately $1.7 trillion since the tightening started, M2 money supply remains near its all-time high.

By now, $6.44 trillion of this money supply are sitting in money market funds, earning a passive ~5% annualized. As rates get cut, not only are risk assets more attractive as access to capital becomes easier, but equally will the risk-free alternative become less attractive. When 2-3 trillion of sidelined cash re-enter markets, we are likely to see digital asset prices rise materially, both acutely during the early innings of easing, and throughout the entirety of the lower interest rate environment.

Conclusion

Rate cuts are on the horizon regardless of the political outcome. As such, money will flow. Where that money flows to is largely dependent on whether there will be changes to SEC leadership or mandate, or alternatively a loss of jurisdiction over digital assets through courts and Congress. The outcome remains binary, though, where under the current regime Bitcoin and memes thrive, while under a more constructive regulatory framework, the true innovations in digital assets can begin to serve trillions of dollars in financial value and capture hundreds of billions in value in exchange.

The trade remains two-fold—hold tight till the pivot & pick the right regime. Regardless of the outcome, either bucket can serve as a tradeable universe for the next four years with tremendous upside.

At the end, I’ll leave you with an excerpt from digital asset pioneer Erik Voorhees’ keynote speech at the annual CoinCenter Gala on the promise of crypto and why the battle for digital sovereignty must be won in America—a country founded upon the principles this industry most closely reflects.

9

---

Disclaimers:

This is not an offering. This is not financial advice. Always do your own research.

Our discussion may include predictions, estimates or other information that might be considered forward-looking. While these forward-looking statements represent our current judgment on what the future holds, they are subject to risks and uncertainties that could cause actual results to differ materially. You are cautioned not to place undue reliance on these forward-looking statements, which reflect our opinions only as of the date of this presentation. Please keep in mind that we are not obligating ourselves to revise or publicly release the results of any revision to these forward-looking statements in light of new information or future events.

Apple held its annual Worldwide Developers Conference on Monday and announced its latest AI developments - one day later, it spurred the stock to hit an all-time high of $200. Here are our key takeaways from the event:

1. ‘Apple Intelligence’ is a big leap into personalization for its users.
   
   The integration of AI, particularly in collaboration with OpenAI, enables Apple to provide more personalized and efficient user experiences, ranging from enhanced Siri interactions to more intelligent app features.

2. Apple is making privacy front and center.
   
   Apple Intelligence will leverage cloud-based models on special servers using Apple Silicon to ensure private and secure user data. If a request needs to go to a cloud server, Apple says it will only send a limited selection of data in a “cryptographically” secure way.

3. Siri is getting smarter.
   
   iOS18 will employ ChatGPT to improve the responsiveness and usefulness of its virtual assistant Siri — enabling users to seamlessly switch between text and voice.

4. New GenAI tools unlock creative potential.
   
   Apple's AI-powered picture and emoji-generating tools, such as Genmoji and Image Playground, allow users to customize and animate digital expressions more vividly.

5. The announcements weren’t without its detractors.
   

   Elon Musk warned that he’d ban devices over security fears in a Tweet shortly after the OpenAI + Apple announcements.

   

6. Slow and steady may win the race.
   

   Apple's AI vision is not about a single large model but rather a number of smaller models that do not require the same amount of processing power and memory and that work on Apple's devices and CPUs. This, along with Apple’s incremental approach to AI integration, is different than trying to jam new tech into old products. Many of the AI features Apple highlighted on Monday are comparable to tools launched over the last year. Apple's AI can summarize and rewrite papers, create tiny graphics, and interpret conversations in real time.

   

7. The market for third-party AI apps is getting smaller.

   Apple's ecosystem now directly includes features comparable to those seen in third-party programs such as Grammarly, 1Password, AllTrails, Soulver, Bitmoji, and Bezel. This integration reduces the need for separate apps by including these services inside Apple's software stack.
   

8. Apple's choice to keep AI services free might offer them a significant advantage over competitors.
   
   Samsung in particular, as the 2025 deadline for Samsung's free services approaches. Observing how these tech companies adapt and change the smartphone experience will be intriguing as the AI integration competition heats up.

Hartmann Capital Weekly: Choose which topics you want delivered to your inbox with this one question survey.

TAKE THE TEN SECOND SURVEY

9. Meet the new VisionOS 2.
   
   It is enhanced with gesture-driven controls and the ability to turn regular photos into spatial ones for Vision Pro, which now supports ultra-wide displays. The results are stunning, with no ghosting or distorted edges.
   

10. Safari 18 for visionOS 2 beta now supports WebXR
    
    This enables immersive VR experiences via the web.

11. Lag be gone.
    
    Hand tracking has been upgraded from 30Hz to 90Hz, eliminating the lag issues caused by the previous disparity with its 90Hz display, which was problematic for gaming and less responsive compared to the Quest's 60Hz.
    

12. APIs and Devs are a big driver for adoption.
    
    Other noteworthy updates include TabletopKit for building interactive, table-centric games and comprehensive Enterprise APIs that expand control over device sensors, including access to the main camera for enhanced application functionalities.
    

    And through the new VisionOS sessions, developers can access hands-on labs, sample code, documentation, and expert guidance to master developing this new spatial computing platform.

13. No longer an American exclusive.
    
    Apple's Vision Pro will be available for preorder in China, Hong Kong, Japan, and Singapore on June 13th at 6 PM PT and released on June 28th. Preorders for Australia, Canada, France, Germany, and the United Kingdom start on June 28th at 5 a.m. PT and will be available beginning July 12th.

Moving ahead, Apple Intelligence and visionOS 2 will encourage industry innovation and competition, focusing on privacy, better productivity, and immersive content. The emphasis on AI and VR integration, upgraded hardware, and new development tools will raise the standard, inspiring other companies to better their technologies and user experiences.

Disclaimers:

This is not an offering. This is not financial advice. Always do your own research.

Our discussion may include predictions, estimates or other information that might be considered forward-looking. While these forward-looking statements represent our current judgment on what the future holds, they are subject to risks and uncertainties that could cause actual results to differ materially. You are cautioned not to place undue reliance on these forward-looking statements, which reflect our opinions only as of the date of this presentation. Please keep in mind that we are not obligating ourselves to revise or publicly release the results of any revision to these forward-looking statements in light of new information or future events.