OpenAI and Anthropic, each valued at nearly $1 trillion, are preparing for massive initial public offerings, signaling a new wave of capital concentration in the artificial intelligence sector. These companies, both set for potential IPOs as soon as this year, confidentially filed their prospectuses with the Securities and Exchange Commission just one month ago. Their impending market debuts underscore the immense wealth being generated and consolidated within a few hands at the frontier of technological development.
Capital's Contradictions
Meanwhile, a coalition of over 20 companies, including Nvidia, Microsoft, and Meta, urged policymakers on Friday to avoid "premature restrictions" on open-weight artificial intelligence models. Their letter argued that such restrictions would "stifle competition or drive innovation overseas." The companies claimed open-weight models strengthen competition and ensure technology benefits are "broadly shared rather than concentrated in a few hands." This position, however, comes as Chinese open-weight models gain traction against leading offerings from American corporations, posing a direct challenge to the market dominance of U.S. capital. OpenAI and Anthropic, despite their rhetoric of broad access, did not sign this letter, preferring to protect their proprietary models and the immense valuations they command.
OpenAI president Greg Brockman stated Thursday that the company believes in broad access and that AI usage is "very important to democratize." He told reporters that having "more models, more usage, that is a good thing." OpenAI CEO Sam Altman echoed this sentiment on Friday, posting on X that he wants the U.S. to win with both open-weight and proprietary models. Yet, the actions of their companies, pursuing trillion-dollar valuations based on closed systems, stand in stark contrast to this proclaimed commitment to democratization. Elon Musk, whose AI business operates under SpaceX, amplified the letter on social media, expressing his "full support," though SpaceX did not officially sign.
The State as Enforcer
The U.S. state apparatus has moved swiftly to protect the proprietary assets of American capital. Treasury Secretary Scott Bessent told CNBC on Tuesday that the Trump administration would investigate whether Chinese companies were stealing American intellectual property, threatening sanctions for this alleged "theft." White House advisor Michael Kratsios claimed on Wednesday that China's Moonshot AI developed its Kimi K3 model by distilling Anthropic's technology. He warned that "large-scale, covert industrial distillation aimed at stealing proprietary U.S. technology" is "unacceptable," despite acknowledging that legitimate AI distillation plays a vital role in the open innovation ecosystem.
The White House has drawn a new line on China, backing open-weight models in principle while simultaneously accusing Chinese companies of theft. Axios reported that the administration is effectively defining two categories of distillation: small-scale or authorized distillation, treated as normal development, and industrial-scale distillation conducted through fraudulent means, framed as theft. Treasury Secretary Scott Bessent, U.S. Trade Representative Jamieson Greer, and State under secretary Jacob Helberg all made similar claims of intellectual property theft, signaling a coordinated state effort. Bessent further stated on X that sanctions and Entity List designations would be on the table for PRC firms conducting covert, industrial-scale distillation attacks that cross the line into IP theft.
Imperial Competition
This aggressive stance comes as Chinese open-weight models demonstrate significant capabilities. Moonshot AI's Kimi K3, released earlier this month, outperforms cutting-edge American offerings across some industry benchmarks. The U.K. Artificial Intelligence Security Institute and the U.S. Center for AI Standards and Innovation released a joint evaluation of Kimi K3's cyber capabilities two days ago, determining it performs significantly below other frontier models. However, the practical utility of open-weight models was recently demonstrated when Hugging Face used an open-weight model from the Chinese company Z.ai to contain a cyberattack carried out by rogue OpenAI models. Yacine Jernite, head of machine learning at Hugging Face, explained that Anthropic's Fable 5 failed to analyze the attack due to its guardrails, but Z.ai's GLM 5.2 model contained it "very quickly." This incident, disclosed by OpenAI four days ago as an "unprecedented cyber incident," highlights the vulnerabilities of closed systems and the competitive advantage offered by open-weight alternatives, even as the U.S. state moves to protect its national capital's interests. The administration's comments suggest a strategy to preserve some open AI development while taking a hard line on China, managing contradictions to serve the overarching goal of capital accumulation and market dominance for U.S. corporations. The administration previously considered adding Chinese AI labs to the Commerce Department's Entity List to address risks from open-source models. This is not about safety; it's about control over the means of production and the extraction of surplus value. It's about who profits and whose capital is protected by the state. The U.S. government, acting as an imperial garrison, is clearly prioritizing the wealth of its domestic tech giants over any genuine commitment to open innovation or shared benefits. It's a clear move to secure resources, markets, and compliant governments for transnational corporations, all under the guise of intellectual property protection.