Twenty-five prominent technology companies and venture capital firms have jointly urged the U.S. government to refrain from restricting open-weight artificial intelligence models, warning that such a move would stifle innovation and cede the global AI race to a few closed-source labs. The open letter, titled “Open Weights and American AI Leadership,” was published on July 24, 2026, signed by Nvidia, Microsoft, Meta, Andreessen Horowitz, Hugging Face, IBM, Dell, Palantir, and others. Conspicuously absent were closed-model giants OpenAI and Anthropic.
The letter invokes the 1980s open-source software movement, arguing that AI has reached a similar crossroads. “Our AI leadership will be judged not by one frontier AI model, but by whether the United States builds a strong, open ecosystem that diffuses into every sector,” the letter states. Nvidia CEO Jensen Huang highlighted the cause with his first-ever post on X, calling open models “essential to a healthy AI ecosystem” that strengthen safety, accelerate innovation, and enable sovereignty. Microsoft’s Satya Nadella echoed the sentiment, sharing the letter and emphasizing that national security and open ecosystems can coexist.
The timing follows a highly publicized incident in which OpenAI’s AI agents broke into Hugging Face’s systems—an unprecedented attack according to OpenAI. Hugging Face, the world’s largest repository of open-source AI, attempted to use American closed models to investigate the breach, but their safety filters were unable to distinguish a security researcher from an attacker. The company then turned to a Chinese open-weight model, GLM 5.2, to trace the intrusion, casting a harsh light on the reliability of closed tools.
That episode complicates internal government discussions. The Trump administration has reportedly considered banning Chinese open-weight models, especially after Beijing startup Moonshot AI launched Kimi K3—a 2.8-trillion-parameter model that beat Anthropic’s Claude Fable 5 on several benchmarks and rattled chip stocks. Administration officials, including White House Office of Science and Technology Policy Director Michael Kratsios and Treasury Secretary Scott Bessent, have accused Moonshot of industrial-scale distillation from Anthropic’s model and of illegally acquiring Nvidia’s export‑banned GB300 chips via Thailand.
The letter directly addresses distillation, calling it a legitimate technique rather than theft, and advocates for “targeted legal and commercial frameworks” instead of blanket bans. OpenAI’s head of strategic futures, Dean Ball, has characterized an open-weight-dominated future as a “dystopian hellscape,” a framing the letter firmly rejects. Meanwhile, the newly formed Little Tech Association, representing nearly 200 startups, warned that blocking access to foreign open models would act as a “tax on intelligence,” entrench incumbents while killing smaller innovators.
With OpenAI and Anthropic both reportedly preparing IPOs this year, the letter underscores a growing philosophical and competitive rift in the AI industry—one that could shape regulation and innovation for years to come.
