Primitive Ventures Founder: AI Creates Four Types of Opportunities — What Can Young Entrepreneurs Do?

Author: Dovey Wan, Founder of Primitive Ventures; Translation: @金色财经xz

Founders are now experiencing a real identity crisis in the AI space. We see the same script in San Francisco and Shenzhen: Founders are returning funds, trying to squeeze into top U.S. large language model labs (LLM Labs) or Chinese robotics giants before their windows close. More than one founder has told me that, compared to the enterprise value (EV) generated by continuing to operate their startups, the opportunity cost of not jumping on board feels much higher right now.

Through in-depth conversations with those who have exited, a rough consensus is forming: Over the past twenty years, “technology” has almost equated to the internet. Now, the pendulum has swung back completely, returning to atom-level (Atoms) super-industrial work rather than bits (Bits). The next-generation entrepreneurial opportunities—where AI truly accelerates outputs in aerospace, energy, life sciences, materials, robotics, and other fields—are no longer the stage for dropout scholars or founders without deep industry roots. AI makes building easier, but it also makes winning much harder.

For a certain type of young founder, this is psychologically brutal. When everyone has roughly the same product iteration speed and analytical brainpower, what is left of the moat for young founders?

If I were to give sincere advice to young founders, it would be: Do not confuse “the most visible game” with “the game with the highest win rate.”

I increasingly believe that AI is simultaneously spawning several distinct opportunity sets:

The first obvious category: super-industrialized AI in biotech, robotics, aerospace, defense, energy, and related fields. The greatest returns will come from here. But these are no longer lightweight software narratives. They are capital-intensive, operation-heavy, heavily regulated tracks, often built by those with deep technical or institutional leverage. Moreover, given their strategic importance to nation-states, these fields will be under high sovereignty protection.

The second category echoes fintech in the 2010s: the biggest upside often doesn’t come from selling tools to traditional giants but from becoming the service itself (like SoFi, Oscar Health, LendingHome, etc.). AI is pushing many industries in the same direction: acquiring traditional businesses, retaining cash flow, and reconstructing operations and data with AI. Recent deals like Long Lake with Amex GBT are excellent examples. This opportunity is real, but it’s a different game. It naturally belongs to those who are proficient in finance, M&A, workflow restructuring, and real-world service networks—not the typical young tech founder path.

The third category of opportunities isn’t as glamorous, but for those troubled by “permanent underclass fears” today, it’s more accessible: a large number of businesses operated by solo founders or small teams of three, with annual revenues between $5 million and $50 million. They aren’t necessarily suitable for large-scale venture capital giants, but they are solid, good businesses. When everything becomes highly personalized, you only need to serve a niche market with genuine purchasing power. In D2C (direct-to-consumer), combined with increasingly mature 3D printing technology, what you need are founders who can control flexible supply chains and precisely grasp community collective needs.

In content, I have a friend quietly building a female-oriented fan fiction platform with surprisingly impressive income figures. The origin was his girlfriend using generative AI to create spicy stories and anime for herself. While others rush toward infinite deflation, taste, unmet desires, and community intimacy can generate compound effects.

If the third category is about monetizing desires through taste and community closeness, then young founders may have more advantages—or at least fewer disadvantages—in the fourth category: building new market structures that turn information asymmetry into fee flows.

In constructing new market structures, winners are not necessarily those with the best models, but those who create new trading targets/payment methods or new trading pathways. They can create new venues for capital flow before traditional giants enter, organize liquidity, bundle risks into more tradable forms, and subtly embed fees into infrastructure. Hyperliquid didn’t win by initial institutional backing but by creating faster, cleaner venues for offshore leverage trading, enabling liquidity to aggregate before competitors react. Polymarket did something similar in event-driven information markets, launching elegant standardized “yes/no” tokens. Many excellent consumer fintech companies are following the same path.

Money flows like water, along the path of least resistance. So, if you track global capital flows, you’ll see huge pain points—opportunities that are not captured by existing giants, and which cannot be solved by AI alone. Teams with prominent backgrounds often don’t know how to conduct local field research. As early as 2023, I wrote about the astonishing data showing USDT and USDC have extremely low overlap, and most U.S. stablecoin founders are completely unaware of this gap. That’s why I remain structurally optimistic about crypto, fintech, and open financial infrastructure—these are fields where young founders have unique advantages. They are the best channels for highly financialized and globalized assets, and young founders will have a clear edge here. The future of the dollar is outside the United States.

History is often measured in 20- to 30-year cycles—social movements and asset cycles follow this pattern. Each generation has its own bullish logic. During a person’s roughly 40-year peak (from age 20 to 60), there are about 2 to 3 real opportunities to catch the wave of the era: institutional and production method revolutions. If lucky, you encounter each once. Our parents experienced the Cold War (or China’s reform and opening) and the first wave of the internet; our generation faces crypto and AI.

Many suffer because they judge paths that shouldn’t succeed by the success functions of AI labs and deep tech outsiders.

Yes, some tracks have become less likely to succeed. Yes, some paths are more closed off than before. But the frontier, by definition, belongs to outsiders; the true frontier doesn’t care much about who you know or whether you pass the old rules’ pre-screening. The real challenge is whether young founders are honest enough to choose the right frontier for themselves and recognize their battlefield.

In every generation, choice is more important than effort. Now, choosing the right game is more critical than ever. If the field you’re building involves reducing friction in capital flows, making it easier for global users to earn, creating new trading targets or methods, then we always look forward to founders building the next-generation financial infrastructure for a more open, programmable, globally accessible economy.