More than 20 major technology companies—including Nvidia, Microsoft, and Meta—urged U.S. policymakers Friday to resist "premature restrictions" on open-weight artificial intelligence models, warning that heavy-handed regulation could concentrate power in the hands of a few corporations and push innovation abroad.
The letter arrives at a critical moment. Chinese startup Moonshot AI released Kimi K3 earlier this month, a model that outperforms cutting-edge American offerings across some industry benchmarks. That achievement has rattled Washington, triggering a fierce debate about whether to restrict Chinese AI access in the U.S. and accusations of intellectual property theft that could reshape how America regulates the technology.
But the companies signing the letter—which notably excluded OpenAI and Anthropic—are making a different argument: open-weight models, which users can download, modify, and run on their own infrastructure, actually strengthen competition and distribute AI's benefits more broadly. They argue that relying solely on closed models isn't inherently safer. "They can be breached, misused, or fail in ways that outsiders cannot detect," the companies wrote, adding that concentrating advanced AI capabilities behind a small number of closed models compounds that risk.
The Competition Question
The stakes are enormous. OpenAI and Anthropic, each valued at nearly $1 trillion, are preparing potentially massive initial public offerings that could land as soon as this year. Anthropic confidentially filed its prospectus with the Securities and Exchange Commission in June, and OpenAI followed suit days later. That timing matters: both companies have strong incentives to shape policy in ways that protect their proprietary models from competition.
Yet OpenAI president Greg Brockman struck a different note Thursday, telling reporters that the company believes in broad access to AI technology. "I think that, that fundamentally, AI and AI usage is something that is actually very important to democratize," he said during a briefing in New York City. "And so, for me, at a sort of deep level, I think that having more models, more usage, that is a good thing." OpenAI CEO Sam Altman echoed that sentiment Friday, posting on X that he wants the U.S. to win with both open-weight and proprietary models.
Elon Musk, who runs an AI business under his rocket company SpaceX, amplified the letter on social media, writing that it has his "full support," though SpaceX did not officially sign it.
The Theft Accusation
Meanwhile, the Trump administration is drawing what it describes as a careful distinction between legitimate and illegitimate uses of open-weight technology. Treasury Secretary Scott Bessent told CNBC on Tuesday that the administration would investigate whether Chinese companies were stealing American intellectual property and said the government has "the ability to sanction them because of this theft."
White House advisor Michael Kratsios sharpened that framing Wednesday, claiming that China's Moonshot AI developed Kimi K3 by distilling Anthropic's technology. He acknowledged that legitimate AI distillation—creating cheaper, specialized models from larger ones—plays a vital role in open innovation. But he drew a hard line at "large-scale, covert industrial distillation aimed at stealing proprietary U.S. technology," calling it "unacceptable."
Treasury Secretary Bessent, U.S. Trade Representative Jamieson Greer, and State under secretary Jacob Helberg all made similar claims about intellectual property theft. The administration has previously considered adding Chinese AI labs to the Commerce Department's Entity List, and Bessent signaled Friday that sanctions and Entity List designations remain on the table when "PRC firms conduct covert, industrial-scale distillation attacks that cross the line into IP theft."
Real-World Consequences
The practical implications of these competing visions became visible earlier this month when Hugging Face, a major open-source AI platform, faced a cyberattack. OpenAI disclosed the attack on Tuesday, calling it an "unprecedented cyber incident." Yacine Jernite, head of machine learning at Hugging Face, revealed that the company initially tried to use Anthropic's Fable 5 to analyze and contain the attack, but the model's guardrails prevented it from recognizing that Hugging Face was defending itself. Jernite said Hugging Face then turned to Z.ai's open-weight model GLM 5.2 and was able to contain the attack "very quickly using this model."
That incident underscores the open-weight advocates' core claim: diversity in AI models serves a practical purpose. When one company's proprietary safeguards got in the way, an open-weight alternative solved the problem.
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. That assessment may complicate calls for outright bans.
Why This Matters:
The administration's emerging stance—supporting open-weight development while cracking down on what it characterizes as theft—reflects a fundamental tension in tech policy. Concentrating AI power in a handful of corporations raises serious questions about who controls transformative technology and who benefits from it. Open-weight models distribute that power more widely, but they also create security and accountability challenges that regulators are still learning to manage. How the U.S. resolves this tension will determine whether AI development remains competitive and accessible or becomes the exclusive domain of a few well-capitalized firms. The distinction between legitimate distillation and theft is real, but it's also one that powerful companies have strong incentives to blur. Policymakers face pressure from multiple directions: companies seeking protection from competition, civil society groups concerned about concentrated power, national security officials worried about China, and technologists arguing that open development drives innovation. The decisions made in the coming months will shape whether ordinary developers and smaller companies can participate in AI's future, or whether the technology becomes another domain where market power concentrates at the top.