Less Is Never More in China
Shallow capital markets make China's AI efficient - not dominant
China’s AI strategy revolves around building more efficient models—lower capital expenditure, fewer parameters, and performance that is arguably competitive with Western large language models (LLMs). This isn’t because Chinese founders are uniquely obsessed with elegance or have lower ambitions for intelligence. It’s because they simply don’t have the money that their American counterparts do. If they had it, they would spend it. They aren’t choosing frugality; frugality is forced upon them.
This is why the U.S.–China AI contest is not a single race. The two countries are running on parallel tracks, optimizing for entirely different objectives.
American AI companies are playing a global game. Their goal is to become the trusted, universal substrate of intelligence—the “token factory” that powers a cognitive operating system for the world, both digital and physical. As Elon Musk keeps emphasizing, the winning bet is on maximum truth-seeking AI (Grok’s north star), not generic AGI (OpenAI), not ideologically slanted AGI (Google), and certainly not “safe but lobotomized” AGI (Anthropic). Global users will ultimately reward only the system that delivers the most useful, least compromised tokens. Everything else will be relegated to niche or regional use.
Chinese AI companies, by contrast, remain severely capital-constrained. The reason is straightforward: China’s capital markets are thin because capital hates control. Deep, liquid markets require freedom—freedom of thought, freedom of speech, and freedom from arbitrary political interference. China’s ruling class maintains an iron grip on the country, and the price of that grip is shallow pools of risk capital. Even European countries that many of us criticize for creeping overreach (the UK, Germany, France) still have far deeper and more functional capital markets than China. Freedom, measured in dollars of patiently investable capital, pays dividends.
Faced with this reality, Chinese tech firms have adapted by bootstrapping aggressively from cash flow. Two big consequences follow:
They are forced to be myopic. Payback periods must be short; moonshots that burn cash for years are off the table.
They cannot credibly build for a global audience. No one outside the Great Firewall is going to adopt a Chinese cognitive OS, no matter how cheap or efficient. So they optimize for the domestic market—a huge market, to be sure, but still only ~18% of world GDP and shrinking in relative terms.
The real bottleneck, then, is not chips. High-end compute is available to anyone determined enough (legally or otherwise), and export controls are porous in practice. The binding constraint is capital.
When we talk to Chinese AI founders, they readily admit they chase efficiency because they have no other choice. The interesting question is whether this constraint is secretly an advantage—forcing them to innovate in ways that wasteful Americans never will.
Our answer: no. Access to abundant capital is always better than scarcity, just as access to abundant energy is always better than scarcity. What matters is what you choose to do with it. Tesla, for example, has raised tens of billions yet remains maniacal about efficiency—“tokens per dollar” and “tokens per watt” are constant refrains from Elon. Having the option to spend big doesn’t mean you must waste; it means you can choose to be disciplined while still swinging for the fences when the moment is right.
Countries, like companies, evolve within the constraints they’re given. Italy built the world’s greatest culinary tradition on sunshine and fresh produce; Norway became the planet’s finest fish exporter despite the cold and dark. No one talks about an “Italy discount” on salmon or a “Norway discount” on tomatoes—each simply plays its hand.
The same is true in tech:
The U.S. has deep capital → bets on massive, disruptive, long-horizon platforms.
Europe has capital but regulatory straightjackets → narrow, compliant applications.
China has neither deep capital nor regulatory freedom → ruthless efficiency and fast cash-flow loops.
None of this makes China “better” or “worse”—it just makes it different. But let’s not romanticize the difference. Less is never more when the “less” is forced upon you. Capital scarcity is biting Chinese AI ambitions hard, and the proof is in the continuing one-way flow of elite talent: the best young minds in China still dream of building in America, where capital is abundant and the mandate is still, at least relatively, to seek truth and ship fast.
That’s the real scoreboard.