China is exploiting the collectivist and bureaucratically stifled scientific environment in the West. It actively recruits top international scientists by offering them what they can no longer easily get at home: freedom to pursue their ideas, significant funding, and prestige, presenting a major brain-drain risk.

The proliferation of powerful open-weight models from Chinese entities is not just a commercial move. It's a calculated geopolitical strategy to commoditize the AI model layer. By reducing the technological gap and preventing US companies from establishing an unassailable lead, China aims to dilute America's economic dominance in a field potentially worth trillions.

Historically, America's strength grew by attracting top academics and scientists fleeing oppression, leading to major innovations. Now, by attacking universities and international students, the U.S. is causing this flow of elite human capital to reverse. Top researchers are seeking asylum elsewhere, draining the country of future breakthroughs.

A major vulnerability in the US strategy to block China's AI progress is its reliance on Chinese nationals for top talent. With nearly 40% of leading AI researchers coming from China, any serious attempt to restrict Chinese access to technology creates a self-defeating talent crisis for American tech companies.

The dynamic between the two superpowers has fundamentally shifted. Economic and technological issues, from AI and semiconductors to biotech and pharma, are no longer seen as purely commercial. Instead, both Washington and Beijing view them through a national security lens, treating them as potential weapons in a broader strategic conflict.

This argument posits that China's strategy isn't about open collaboration but is a state-subsidized effort to release unprofitable open-weight models. The goal is to flood the market, eliminate competition from US AI labs by making them unprofitable, and then control the market once competitors are gone.

Proposing an open-source model that quickly follows the US frontier is a flawed strategy. It antagonizes the US on two fronts: it threatens national security by promising to release dangerous capabilities to the world within months, and it commercially undermines trillion-dollar US companies by open-sourcing their technology.

The closed nature of leading US AI models has created an information vacuum. Sridhar Ramaswamy notes that academia is now diverging from US industry and instead building upon published work from Chinese companies, which poses a long-term risk to the American innovation ecosystem.

The core ethos of open source—unrestricted proliferation of powerful capabilities to everyone—is fundamentally incompatible with the national security state's mandate to control dangerous technologies. As AI models become more potent, this irreconcilable conflict will escalate into a major policy battle.