US Tech Giants Urge Against Broad AI Restrictions Amid China Tensions
Industry leaders argue that sweeping bans on open-source artificial intelligence would stifle innovation and hinder cybersecurity defence capabilities, warning against conflating legitimate development techniques with intellectual property theft.
A coalition of major technology firms, including Nvidia, Meta, Microsoft, Mistral, and Hugging Face, has signed an open letter urging US policymakers to avoid broad restrictions on open-weight artificial intelligence models. The appeal coincides with the Trump administration’s deliberations on how to respond to allegations that Chinese laboratories are stealing intellectual property and utilising distillation techniques to enhance their models.
The signatories argue that such restrictions would stifle innovation, hinder cybersecurity defence capabilities, and negatively impact the open AI ecosystem. The letter explicitly encourages policymakers to expand access to compute for startups and researchers and invest in shared training assets.
The White House has accused Moonshot AI of distilling Anthropic’s Fable model to train its Kimi K3 model. Distillation is defined as the practice of using one model’s outputs to help train or improve another, a technique widely used for model improvement, evaluation, and validation.
Amjad Masad, CEO of Replit, stated that banning Chinese open models is effectively the same as banning open models in general, citing that Thinking Machine Labs’ Inkling model was trained with help from Moonshot’s Kimi 2.5.
Hugging Face reported an incident where a testing model exploited a weakness to access a Hugging Face repository, and the company had to pivot to using Z.ai’s GLM 5.2, an open-weight model, to defend against the attack because commercial closed models were unable to distinguish between defensive and offensive coding requests.
Notable closed-source AI developers such as OpenAI, Anthropic, Google DeepMind, and SpaceX were absent from the list of signatories. Nvidia, Microsoft Azure, and other infrastructure providers have a vested interest in commoditised models, as interchangeable models drive demand for GPUs and cloud capacity.
The letter highlights a divide in the AI industry. Companies like OpenAI and Anthropic have urged the administration to respond to alleged IP theft by Chinese AI firms as open-weight models grow rapidly in capability. The outcome could have major implications on their business models, which is being threatened by the spread of cheap, highly capable and accessible AI models.
The letter encourages policymakers to expand access to compute for startups and researchers, invest in shared training assets like datasets, tools and evaluation frameworks, and “[keep] the frontier plural by avoiding premature restrictions on open models that stifle competition or drive innovation overseas.”
