Nearly 200 Startups Tell Trump: Banning Chinese Open-Weight AI Will Cripple American Innovation

200 Startups Urge Trump Not to Ban Chinese Open-Weight AI

Nearly 200 Silicon Valley companies, including Proton and Y Combinator, sent letters to President Donald Trump, Commerce Secretary Howard Lutnick, and Office of Science and Technology Policy Director Michael Kratsios on Wednesday with a warning: cutting off access to Chinese open-weight AI models would cripple American startups 1.

The letters, organized by the newly formed Little Tech Association, represent the first coordinated push by the wider startup community to enter this debate 1.

Their argument is straightforward. Open-weight AI models, whose trained parameters are publicly downloadable, are already available globally. Banning Americans from using them would not stop their proliferation elsewhere. It would simply remove them from the toolkits of U.S. builders 1.

The urgency behind the lobbying is real. Treasury Secretary Scott Bessent said Tuesday on Fox Business that the administration would investigate whether Chinese AI companies had improperly distilled American models to develop their own, and raised the possibility of sanctions for companies found to have engaged in intellectual property theft 1. The same week, Kratsios alleged that Moonshot AI had distilled Anthropic's Fable model while developing its Kimi K3 model, using an internal platform built to conduct large-scale distillation while attempting to evade detection 1. Kratsios also claimed Moonshot had acquired Nvidia GB300-equipped servers to train its models despite a ban on their sale to Chinese entities 1. Moonshot has not publicly responded to the allegations 1.

The models at the center of this dispute come from Chinese companies including Moonshot AI and Alibaba 1. Many startups depend on them as a lower-cost alternative to commercial offerings from OpenAI and Anthropic 2. Suhail Doshi, founder of AI infrastructure startup Particle, warned in an interview that a ban would wipe out hundreds of companies and leave startups paying Anthropic's prices by default 1.

"American leadership requires two things: world-leading American open-weight models and continued access for U.S. builders to open models already available worldwide," the founders wrote 1.

The lobbying also exposes a widening split inside the AI industry. Anthropic has pushed for tighter restrictions on Chinese AI developers over security concerns 2. The startups signing this letter see restrictions as a policy that would protect incumbents and punish everyone else.

Here is the paradox the administration has not reconciled. The national-security case for restriction assumes that limiting American access to Chinese models protects U.S. interests. But open-weight models are, by definition, already public. Once Moonshot or Alibaba releases model weights, anyone with an internet connection can download them. A ban does not remove these models from the world. It removes them from American toolkits.

The builders who would lose access are not adversaries.

For all the public escalation, the policy machinery has barely moved. The Commerce Department had not drafted plans to add Chinese AI companies to its Entity List as of Wednesday 1. Internal discussions among senior officials did not seriously consider a blanket ban 1. A White House official dismissed reports of pending action as "baseless speculation" 1. Yet the same week brought two senior officials publicly alleging model theft and threatening sanctions. The gap between the official dismissal and the private posture is where the policy risk actually lives.

The Little Tech Association's executive director, Harry Godfrey, argued the administration should pursue the lightest possible approach to address security concerns without raising costs or limiting access for American builders 1. That logic extends beyond this fight. When the technology is already globally distributed, walls do not keep it out of adversary hands. They keep it out of yours.