A stark warning has emerged about the structure of the American artificial intelligence industry, suggesting it has inadvertently created a system of "central planning" with corporate overlords. This critique comes as the US stock market, heavily buoyed by AI hype, marches towards record territory, with many speculating about a potential bubble.
The Singapore Perspective: A View from the Knife-Edge
The observations stem from a week spent in Singapore, a nation balancing carefully between US and Chinese influence, during the Singapore FinTech Festival in early December 2025. Conversations with central bankers, founders, investors, and diplomats painted a clear picture: China's position in the global AI race is accelerating exponentially, primarily driven by the proliferation of sophisticated open-source models.
This shift is quantified by data from Hugging Face. As of the beginning of December 2025, China holds 14 of the top 20 spots on the performance leaderboard for AI models, with the majority being open source. Notably, no US open-source models feature in that top tier.
The Open-Source Disruption: Flows of Capital and Compute
The practical impact of this is already being felt in Silicon Valley's own backyard. According to Martin Casado, a partner at venture fund Andreessen Horowitz, 80% of AI startups pitching to the firm are using Chinese open-source AI. Even high-profile tools like the "vibe-coding" startup Cursor are reportedly built on Chinese model cores.
This trend poses profound questions for the US investment thesis. If the global application layer is increasingly built on more efficient, cost-effective Chinese models, what does this mean for the projected revenue of giants like OpenAI? Furthermore, if these models require less computational power, how does this impact the hundreds of billions of dollars being poured into data centre construction—a key driver of current US GDP growth?
Beyond Software: The Hardware and Robotics Frontier
The divergence extends beyond pure software. While the US focuses on foundational models and a push towards Artificial General Intelligence (AGI), insights from Singapore suggest China is pulling ahead in applied AI, particularly in hardware and robotics. For instance, while Jeff Bezos recently raised $6.2 billion for manufacturing AI, Chinese firms have been deeply invested in this area for years.
The pragmatic Singaporean business community, known for its voracious curiosity, is now actively hedging its bets. DeepSeek is as common in their vocabulary as OpenAI, and entrepreneurs are routinely asked if they have drawn inspiration from Chinese AI applications—a question unheard of a decade ago.
The Perils of Corporate Central Planning
This global landscape highlights a critical vulnerability in the US approach. The author, Lewis Z Liu, argues that an overwhelming focus of capital on a select few companies like OpenAI and Microsoft has stifled innovation. Venture capital investors now routinely ask if OpenAI is planning a similar product, killing potential investments if the answer is yes. This, Liu contends, is the antithesis of the disruptive thinking that built Silicon Valley and mirrors a form of central planning—just with corporate, rather than government, rulers.
In contrast, China's vibrant ecosystem of open-source models fosters vast experimentation in business models and use cases, unburdened by concerns about future "national winners." The irony is palpable: cutting-edge AI is now being made freely available for anyone to build upon, largely through Chinese open-source efforts.
The conclusion is clear: the AI race is global. For the West to compete, it must diversify its thinking, embrace more open development, and avoid becoming a victim of its own centrally-planned version of market capitalism. The goal must be to ensure AI benefits all of humanity, not just a narrow corporate elite.
