An American field study on China’s artificial intelligence ecosystem has been circulating widely within Chinese technology and venture capital circles this week, sparking pointed self-reflection among investors and founders. The study, conducted by researcher Macedo and analyzed by Xia Ri in Eurasia Review, argues that China’s AI bottleneck is not a deficit of talent or capital, both of which are abundant, but rather a structural screening process that systematically rewards high-level standardized performance over the kind of perilous, non-consensus originality that drives foundational breakthroughs.
The Talent Paradox
Macedo’s central observation is striking: the vast majority of entrepreneurs he encountered in China are exceptionally diligent, maintaining an intense work ethic and presenting impeccable credentials, elite academic backgrounds, experience at firms like ByteDance or DJI, and impressive arrays of papers and patents. Yet the specific archetype he was seeking, the “independent thinker” characterized by intellectual rebellion, obsessive focus, and the ability to formulate entirely new questions, remained relatively scarce. He attributes this not to any inherent cultural failing, but to the combined forces of education and venture capital.
China’s educational system, Macedo argues, excels at training high-functioning executors: people who optimize known problems, accelerate validated paths, and replicate successful paradigms with extraordinary efficiency. Chinese venture capital firms reinforce this tendency by prioritizing pedigrees, endorsements, and “Big Tech” labels when selecting founders to back. The result is a mechanism that produces a high volume of founders who are undeniably excellent, yet fundamentally similar, a market that systematically reinforces what the analysis calls “non-innovation.”
Where the System Works, and Where It Doesn’t
The analysis is careful to distinguish between sectors where China’s approach succeeds and where it struggles. Hardware continues to yield strong results precisely because it relies on supply chain density, engineering iteration, and organizational execution, areas where standardized high-achievers naturally excel. China’s dominance in electric vehicles, consumer electronics, and, increasingly, robotics hardware reflects this strength. The MERICS report on China’s embodied AI ambitions similarly notes that China’s manufacturing scale is its primary competitive advantage in humanoid robotics.
Foundational AI models, native software, and new-paradigm products, however, depend far more on non-consensus intuition, the endurance of long-term isolation, and the willingness to pursue directions that are initially misunderstood. These are precisely the qualities that the current Chinese ecosystem struggles to accommodate. The true deficit, Macedo concludes, is not a lack of intelligent individuals but the absence of a mechanism that allows outliers to persist and thrive over the long term.
The Valuation Problem
The field study also raises concerns about valuation dynamics in the Chinese AI market. Macedo notes that many early-stage AI projects command premium valuations well before establishing product-market fit or generating revenue, while the late-stage market drives a handful of scarce targets to levels that defy fundamental economic logic. Capital is increasingly gravitating toward “replicable success narratives,” “prestige labels,” and the “speculative space of imminent IPOs,” funding templates rather than teams with the potential to shift paradigms. The humanoid robotics sector is specifically flagged as a potential valuation bubble, a concern that echoes the broader debate over whether China’s record-breaking VC investment in Q1 2026 is being allocated to the right kinds of bets.
The study’s circulation within Chinese VC circles suggests that at least some investors recognize the tension it describes. As Xi Jinping’s call for “original innovation” at the Shanghai basic research symposium makes clear, the Chinese government is also acutely aware that fast-following and iterative improvement are insufficient for long-term technological leadership. The question is whether the structural incentives of the education system and the capital market can be realigned quickly enough to produce the outliers that foundational AI research demands.
What Would Change the Equation
Macedo’s analysis is not entirely pessimistic. He identifies several conditions that, if met, could shift the dynamic. First, the emergence of a new generation of founders who have spent significant time outside China, not just studying abroad, but building products in adversarial, resource-constrained environments where pedigree counts for nothing and results are the only currency. Second, the development of a small but influential cohort of investors willing to back non-consensus bets over multi-year horizons, accepting the higher failure rate that genuine exploration demands. Third, and perhaps most importantly, a change in the cultural valuation of failure: in the current Chinese ecosystem, a failed startup is a reputational liability; in the ecosystems that produce foundational breakthroughs, it is a credential.
Some of these conditions are beginning to emerge. DeepSeek’s Liang Wenfeng is frequently cited as an example of the archetype Macedo describes, a founder who pursued a non-consensus direction for years before the market caught up. The open-source AI strategy that has made Chinese models competitive globally was itself a non-consensus bet at the time it was made. These examples suggest that the ecosystem can produce outliers, but does so despite its incentive structure, not because of it. Closing that gap is the structural challenge that Xi Jinping’s “original innovation” directive seeks to address, and Macedo’s field study, circulating quietly through the VC community, is helping to diagnose it.