Our Crypto AI Thesis (Part I): The Massive Lifetime Opportunity

I often tell friends that in the future, we’ll look back on 2022–2024 as the pivotal years before humanity's massive technological acceleration.

AI is likely to be the most transformative tech trend we’ll witness in our lifetimes—unless, of course, we stumble upon a miracle in life extension that allows us to live for hundreds of years.

This means AI is red-hot, and everyone is scrambling to get a slice of the action.

In just the first half of 2024, $35 billion has been funnelled into AI startups. And that’s only the private side—big tech is investing even more internally, as seen in their enormous GPU purchases from NVIDIA, which has driven NVDA’s market cap to an eye-popping $3 trillion.

But amid the hype, there’s a massive underdog opportunity that many are missing: Crypto AI (or decentralised AI)

Historically, every decade brings transformative investment opportunities that seem unpromising—even foolish—only to later prove visionary.

Social media was once dismissed as a frivolous distraction for teenagers with no real business model. Today, Meta (formerly Facebook) is one of the most powerful companies in the world, and early investors reaped returns exceeding 1,000 times their initial investment.

The narrative for Crypto AI is urgent and compelling. When I explain it, most people understand it right away.

By its nature, AI centralizes power. Left unchecked, it risks concentrating control in the hands of a few monopolistic organizations that will surely wield AI for profit and dominance. Decentralizing AI is not just crucial for our future—it’s essential for a brighter, more equitable society. I’ve explored this philosophical angle in greater depth before.

Skeptics, though, argue that integrating crypto with AI is merely capitalizing on buzzwords, pointing to past disappointments in entertainment, gaming, and social media, where crypto has failed to achieve lasting impact or widespread adoption (yet). I hear this even from smart investors—and it’s a fair concern.

But I believe this time is actually different.

There are several reasons why Crypto AI is poised to follow a very different trajectory. And I’m writing this to explain why.

There’s much more to cover than I initially planned, so I’ve decided to break this into multiple parts.

In this three-part thesis, I’ll examine the technological & investment landscape for Crypto AI in depth. I’ll highlight the most promising sectors and show how I position myself to capitalize on this emerging megatrend.

Part I: Why Crypto AI is the sector to pay attention to

Part II: How I’m thinking about AI agents, decentralised training, verifiable inference, data networks (and other subsectors in Crypto AI)

Part III: A multitude of different ways to capture value from this opportunity

Be Where Secular Technology Trends Intersect

As savvy investors and entrepreneurs, we always want to be riding the waves of change.

To catch the biggest waves, we need to look for places where multiple major currents converge and position ourselves there.

This means identifying secular technology trends—deep-rooted shifts in behaviour driven by technological advancements that redefine entire industries.

Crypto is a secular technology trend representing a shift in how we think about and use money. Other examples include cloud computing, mobile technology, and clean energy.

Simply riding a single secular tech trend isn’t enough, though. Thousands of others see the trend and are making similar bets. To truly stand out, we need to look beyond the obvious.

That’s why it’s remarkable to spot the early convergence of two large-scale secular technology trends.

This is where magic happens.

Convergence + Scale = Opportunity

(1) The Power of Convergence

When two secular trends converge, a sweet spot for innovation and value creation is created — often in places others might overlook.

1. Multiple Growth Drivers: Companies at the intersection of multiple secular trends benefit from the growth momentum of each trend.

2. Reduced Competition: Companies can carve out unique market positions with high barriers to entry because of the deep expertise required on both sides

3. Cross-pollination of ideas: Leads to innovative products and business models

(2) The Power of Scale

When it comes to investing, market size absolutely matters.

Consider Amazon in the early 2000s. Its dominance was not just due to e-commerce but also because it tapped into the emerging cloud computing trend, birthing Amazon Web Services (AWS). AWS now generates billions annually and is a leader in the growing cloud infrastructure market

The larger the secular technology trend—both in terms of the total addressable market and growth potential—and the earlier you enter, the greater the opportunities at its intersection.

Big trends not only provide a buffer against failure but also significantly amplify potential rewards.

Crypto 🤝 Artificial Intelligence

CPUs and GPUs have long been the backbone of computing, but now they power AI, which is emerging as the world’s supercomputer—harnessing humanity’s collective knowledge and creativity.

Crypto, on the other hand, enables the creation of open, decentralized networks, laying the foundation for a new internet.

Combined, these super-computers and super-networks are a beautiful, powerful fit for each other.

• AI enhances the user experience of crypto. Users can interact with the blockchain using natural language instead of learning how to manage wallets, seed phrases, and sign transactions.

• Crypto provides the trustless, permissionless, and secure infrastructure AI needs to stay open and censorship-resistant. It also serves as a powerful coordination layer for building decentralized networks.

• This convergence sets the stage for entirely new business models to emerge

The intersection of Crypto and AI will grow exponentially, benefiting from advancements in both fields, which are on separate development trajectories.

The key lies in identifying what crypto enables AI to achieve that was previously impossible — that’s the secret sauce. Hint: distributed training, data networks and private data. More on these in Part II.

Learning from the Past: The Case of NFTs and DeFi

The rise and fall of NFTs offer a valuable lesson.

NFTs rode high on a wave of speculation within the crypto community but lacked the tailwinds of another secular technology trend to take it further. Entertainment and gaming—the core use cases for NFTs—are complex, mature markets with heavy incumbents, and their dynamics are not driven by technology alone.

As a result, NFTs were unable to sustain their early momentum. While the use cases are real, achieving their promise will take much longer.

On the other hand, DeFi stands as a successful example of the power of intersecting secular tech trends.

By sitting at the crossroads of fintech and crypto, DeFi has revolutionized financial services, offering alternatives to traditional banking, lending, and asset management that address real-world financial needs. The total stablecoin market cap is at all-time highs ($170B) and rising, while $82B is locked in DeFi protocols today.

Tokens: Open Source AI Needs Crypto

When people ask me why AI needs crypto, my answer is simple: tokens.

Traditional software scales at nearly zero marginal cost. Write code once, deploy it everywhere.

AI is an entirely different beast. It operates with high capital and marginal costs.

Training and deploying large-scale AI models demand vast computational resources, making efficiency and access to infrastructure critical success factors.

Right now, we live in a world where centralized giants—OpenAI, Anthropic, Google—hold all the cards. These companies have the talent, the hardware, and the capital. But let’s be real: corporate-owned AI seeks to maximise profits. Always.

Meta’s contributions to open-source AI have been invaluable, but who’s to say when they might stop releasing state-of-the-art models like Llama 3? Developing these systems costs hundreds of millions, and initiatives could be discontinued if Zuck wakes up on the wrong side of the bed one day.

Expecting the open-source movement to compete with these giants purely out of ideals or goodwill is, frankly, unrealistic. We need a new strategy.

The thing is, there’s a ton of untapped compute, research, and talent in the open-source world, outside the top AI research labs. This includes contributions from universities, research centres, collaborative platforms like Hugging Face, and individual AI researchers. But right now, it’s fragmented. It lacks the coordination necessary to make major breakthroughs at scale.

That’s where tokens come in.

As outlined by @long_solitude from Zee Prime, tokens embody crypto’s most powerful property — permissionless capital formation.

Tokens do what traditional fundraising models can’t:

1. They create bottom-up funding opportunities for experimental AI projects that would never see the light of day under the VC model.

2. They bootstrap decentralized networks, turning them into thriving, self-sustaining ecosystems by distributing ownership and value to contributors. Just look at Bittensor as an early proof-of-concept.

3. They enable entirely new business models that slash costs, using DePIN-style tokenomics. One example: shifting upfront infrastructure costs (a huge burden for AI startups) to the network itself. Another: GPU marketplaces that tap into the latent global supply GPUs sitting unused in people’s homes, reducing costs per unit of compute.

Tokens unlock a massive opportunity for retail investors to ride the AI wave. And this is wildly understated.

Crypto is great at finding new markets people want to trade in. Just look at NFTs & cultural assets, social tokens for the creator economy, memecoins for virality etc.

My bet? There’s a huge, untapped demand for retail participation in early-stage AI projects.

As AI seeps deeper into our daily lives and enables us to do things we couldn’t do before, people are starting to realize just how big and real this is. Sure, there will be heavy speculation, but it’s also about democratizing access to the biggest tech revolution of our lifetime—and giving everyone a shot at the next big thing.

Expect things to get crazy.

So Why Now?

New technologies tend to follow a well-defined innovation cycle.

The Gartner Hype Cycle is one of the best-known frameworks. It tracks how innovations blowoff through a hype phase, crash into the trough of disillusionment, and eventually climb up into real-world use cases.

As an investor, the ideal moments to invest are either when you spot a new innovation trigger early, well before the hype peak, or during the trough of disillusionment, when you can identify the startups poised to climb onto the slope of enlightenment.

This begs the million-dollar question: Where are we in this cycle for Crypto AI?

I like to use the chart above from Syncracy Capital (smart folks) to highlight the consensus view today.

It suggests that decentralized AI is nearing or at the peak of inflated expectations. Crypto AI has had a strong year, with several protocols now at multibillion-dollar valuations.

But I disagree with this view. We’re probably nowhere near the peak for Crypto AI yet.

The consensus view underestimates the sheer scale of the opportunity ahead. We haven’t even hit mania territory. Ask around—how many people really understand Bittensor, the bellwether for Crypto AI? In fact, I believe it could be another 1–2 years before we see a true peak.

Here’s why:

1. Crypto AI tokens ($30B) represent just 2.9% of the total altcoin market cap ($1.04T).

The largest Crypto AI protocols with live tokens—TAO, NEAR, FET, and ICP—sit in the $5B–$10B market cap range, and there are only four of them.

Excluding these four, and keeping in mind that ICP and NEAR aren’t exclusively Crypto AI tokens, the total market cap of Crypto AI stands at just $11.7B. That’s <25% of the market cap of memecoins.

Only four other Crypto AI projects (RENDER, GRT, AKT, AIOZ) are valued between $500M and $5B, with the majority having a market cap under $100 - 200M.

This is minuscule when you consider the massive potential of this trend convergence. Crypto AI encompasses both infrastructure and applications, including next-generation smart contract platforms designed for AI.

By comparison, smart contract platforms today have a combined market cap nearing $600B. There are 8 Layer 1 protocols valued over $10B, and another 12 in the $1B–$10B range.

What could the market size for Crypto AI be? It’s such early days and anyone’s guess.

Bloomberg Intelligence estimates the generative AI market will grow at a CAGR of 30%+, reaching $1.3T by 2032. If decentralized AI captures just 10% of that broader AI market, and we apply a speculative premium common in crypto valuations (say 3X), the result is a $390B market in 2032—13X growth from today’s $30B.

Intuitively, I think this is very conservative yet too far out into the future to be useful.

Another way is to consider Crypto AI growing to 10% of the total altcoin market cap in the next 3 years (2027) as AI apps and smart contract platforms launch and gain momentum. If the altcoin market cap hits $2.7T by then (50% growth from its 2021 peak of $1.8T), that would place Crypto AI at $270B—about 9X growth from current levels. That’s $240B in value waiting to be unlocked.

But these numbers are more illustrative than definitive. There are simply too many unknowns. What they do highlight, though, is the scale of the opportunity and provide us with a sanity check when thinking about valuations.

2. Top-tier Crypto AI teams are just getting started

Many high-calibre teams have been heads down building for 1–2 years and haven’t even launched their products to mainnet yet.

Several have raised 8-figures in VC funding, including Sentient, Sahara, Vana, Story Protocol, Gensyn, Space and Time, Ritual, Nillion...

Over the next 12 months, we’ll witness major mainnet rollouts and token launches—the AO Computer ecosystem, for example, is one to keep an eye on and which I wrote about earlier this year.

3. AI is advancing exponentially.

Look at OpenAI’s recent o1 model, which made significant leaps in reasoning capabilities. Scaling laws continue to hold. Crypto AI will closely track broader AI growth.

That said, there’s a lot of noise today. Perhaps more than any other crypto sector. Many startups and protocols will inevitably fail, even if they see some short-term success.

A selective, discretionary approach to identifying winners will likely outperform a broad “spray-and-pray” strategy.

Positive Catalysts Ahead for 2025

I expect several major tailwinds for Crypto AI that will propel the sector and narrative in the coming year.

• Traditional Tech VC Funds Entering Crypto AI: While a16z is already in, major players like Sequoia, Lightspeed, and Accel have only dipped their toes lightly into this sector. As they move further along the investment curve towards decentralized AI, they’ll bring both capital and legitimacy to the space.

• OpenAI IPO: With OpenAI’s private valuation already at $150B and growing revenue, a potential IPO in 2025 could spark retail mania for AI investments. There are very few opportunities for retail investors to directly capture value from the AI trend outside of NVIDIA and hardware. This pent-up demand could spill over into Crypto AI tokens, which will be easily accessible on platforms like Coinbase.

• A More Token-Friendly US Government: If the next U.S. administration post-November 2024 elections takes a friendlier stance towards crypto tokens, it could provide massive tailwinds for the Crypto AI sector, opening up regulatory clarity and encouraging broader adoption

• Breakthroughs in AI: The exponential pace of AI development shows no signs of stopping anytime soon. In a recent blog post, Sam Altman outlines how he believes the future for AI is extremely bright. Progress in AI agents and distributed AI training will unlock new use cases and drive further momentum.

The Bear Case for Crypto AI

While I’m bullish on the massive potential of Crypto AI, I also recognize that nothing is guaranteed. My thesis would shift if these bear case scenarios play out:

• Unfriendly Regulatory Regime: A hostile crypto regulatory environment in major markets like the U.S. could stifle innovation, limit funding, and push crypto projects into grey areas. Tokens are central to decentralized networks and heavy restrictions would undercut one of crypto’s key value drivers. This is less likely in a Trump win but could happen if Kamala becomes the US president.

• AI Fails to Live Up to the Hype: Despite the current frenzy and billions in investment, there’s a chance AI won’t deliver on its promise. If progress slows or hits roadblocks, we could see the AI bubble deflate. I think this is unlikely, though.

• Failure to Find Large Markets and Product-Market Fit (PMF): No matter how innovative, Crypto AI projects must find real business models and solid product-market fit. The sector could stagnate if decentralized AI fails to integrate meaningful commercial use cases. In this scenario, Crypto AI risks becoming more of a niche.

• Mediocre Talent Pool: The pool of top machine learning scientists and engineers is limited. If Crypto AI can’t attract enough top talent who believe in the vision of an open, democratised AI future, innovation will slow, and the sector will struggle to keep up.

Part I: Closing Thoughts

Unironically, AI might be the best thing that’s ever happened to crypto.

It opens the door to true mainstream adoption and practical use cases—something that’s eluded gaming, NFTs, and social apps.

We’re heading into a decentralized AI future fueled by open, public networks. The savviest founders and forward-thinking investors are already taking notice.

The toughest part about Crypto AI is that you can't just watch the crypto space. Otherwise, you'll have a very narrow, superficial view of what’s coming. You need to stay plugged into machine learning developments, dig into the latest Arxiv papers, and talk to founders who truly believe they’re building the next big thing in AI.

But boy, I've never been so excited.

In Part II, I’ll dive into and map out the key subsectors within Crypto AI. Some of the most critical areas I’ve identified include:

1. Decentralised Compute: Training, Inference & GPU marketplaces

2. AI Agents living on-chain

3. Data networks

4. Crypto AI consumer apps

5. Verifiable AI, and more.

I’ve been working on this for several weeks, so it won’t be long before it’s ready. Stay tuned!

Cheers,

Teng Yan

If you have any feedback, feel free to DM me on X or email me at [email protected].

This article is intended solely for educational purposes and does not constitute financial advice. It is not an endorsement to buy or sell assets or make financial decisions. Always conduct your own research and exercise caution when making investments.