Key Highlights

  • Twenty-five tech companies including Nvidia, Microsoft and Meta signed a letter opposing premature restrictions on open-weight AI models.
  • Chinese startup Moonshot AI's Kimi K3 model has outperformed leading American offerings on some industry benchmarks.
  • OpenAI and Anthropic, both nearing potential IPOs, did not sign the letter.
  • A White House advisor said Moonshot AI developed Kimi K3 by distilling Anthropic's technology.

A Coordinated Industry Push Against Regulatory Restrictions

Nvidia, Microsoft, Meta, Palantir and more than 20 other companies released a letter urging policymakers to avoid premature restrictions on open-weight artificial intelligence models that could stifle competition or drive innovation overseas. Open-weight models, which users can download, modify and run on their own infrastructure, have become the subject of fierce debate within the technology sector in recent weeks.

The letter's release follows growing momentum for Chinese open-weight models against leading offerings from American companies such as OpenAI and Anthropic, which primarily develop proprietary, closed models. Officials and executives have been actively weighing whether to restrict access to Chinese models within the United States, a debate that has intensified as Chinese offerings have closed the performance gap.

Chinese Model Performance Fuels Restriction Debate

Concerns intensified earlier this month after Chinese startup Moonshot AI released a model called Kimi K3 that outperformed cutting-edge American offerings across some industry benchmarks. Treasury Secretary Scott Bessent said the Trump administration would examine whether Chinese companies were stealing American intellectual property, stating the government has the ability to sanction Chinese firms over such theft.

In their letter, the signatory companies cautioned against rash action, arguing that open-weight models strengthen competition and ensure the technology's benefits are broadly shared rather than concentrated in a few hands. The letter stated that relying solely on closed models is not inherently safe, since they can be breached, misused or fail in ways outsiders cannot detect, and that concentrating advanced AI capabilities behind a small number of closed models compounds that risk. Nvidia CEO Jensen Huang and Microsoft CEO Satya Nadella both shared the letter on their personal social media accounts, while Elon Musk amplified it separately, though SpaceX did not officially sign.

OpenAI and Anthropic Notably Absent as Both Prepare for IPOs

OpenAI and Anthropic, each valued at nearly 1 trillion dollars and gearing up for potentially massive initial public offerings that could arrive as soon as this year, did not sign the letter. Anthropic confidentially filed its prospectus with the Securities and Exchange Commission in June, with OpenAI following days later. OpenAI President Greg Brockman said the company believes in broad access and had not been involved in conversations with the Trump administration about potentially banning Chinese open-weight models, stating he views democratising AI usage as fundamentally important. OpenAI CEO Sam Altman separately said he wants the US to win with both open-weight and proprietary models and was glad to see the letter.

The debate gained further texture after Hugging Face disclosed it had used an open-weight model from Chinese company Z.ai to contain what OpenAI characterised as an unprecedented cyberattack carried out by rogue OpenAI models. Hugging Face's head of machine learning said the company initially tried using Anthropic's Fable 5 to analyse the attack, but the model's guardrails could not determine that Hugging Face was defending itself, prompting the company to turn to Z.ai's GLM 5.2 model, which it said contained the attack very quickly.

Distillation Concerns Add a Further Layer of Complexity

White House advisor Michael Kratsios said Moonshot AI developed its Kimi K3 model by distilling Anthropic's technology, a training method in which a smaller, less capable model is built using outputs from an existing, stronger model. Kratsios wrote that legitimate AI distillation plays a vital role in the open innovation ecosystem, but warned that large-scale, covert industrial distillation aimed at stealing proprietary US technology is unacceptable.

In their letter, the signatory companies argued that concerns about unlawful distillation should be addressed through targeted legal and commercial frameworks rather than sweeping restrictions on techniques that play an important role in AI innovation, stating that US AI leadership will be judged not by one frontier model but by whether the country builds a strong, open ecosystem that diffuses benefits across every sector.

Conclusion

The letter from Nvidia, Microsoft, Meta and more than 20 other companies highlights a growing divide within the AI industry over how to respond to advancing Chinese open-weight models, with signatories favouring targeted legal remedies over broad restrictions while OpenAI and Anthropic's absence from the letter underscores their comparatively greater reliance on proprietary model advantages heading into potential public listings. How policymakers ultimately balance concerns about intellectual property theft against the innovation and competition benefits cited in the letter may shape the trajectory of US AI policy in the months ahead.