As debate intensifies over proposed U.S. limits on Chinese AI, Silicon Valley startups and NVIDIA chief Jensen Huang have pushed back, arguing that Chinese AI models are high‑quality and should remain accessible to American developers.
## Startups and industry leaders urge against broad bans
A coalition of roughly 200 small and mid‑sized U.S. tech companies — the Little Tech Association — sent a letter on the 22nd to President Donald Trump, Secretary of Commerce Howard Lutnick and Michael Kratsios, director of the White House Office of Science and Technology Policy, urging that access to Chinese AI models not be blocked. The group, whose members include encrypted email provider Proton and accelerator Y Combinator, said U.S. leadership requires both a world‑class domestic open‑weight model and continued access to open‑weight models available globally. It warned that banning downloads would not stop their spread and could weaken the U.S. startup ecosystem.
In public remarks, Suhail Doshi, founder of AI infrastructure company Particle, warned that a ban on Chinese open‑weight models would force “hundreds of companies” to shut down and leave only the largest AI firms advantaged.
## Jensen Huang: open models benefit the ecosystem
NVIDIA CEO Jensen Huang told Axios that “China’s models are excellent” and that “excellent open‑source models should be used.” Asked whether U.S. companies should be permitted to use Chinese AI models, he replied, “Of course.” Huang argued open‑source and open‑weight models expand the market by bringing more users into AI, while customers wanting high‑performance services will continue to pay for commercial offerings. He also dismissed worries that Chinese models would drive U.S. firms from the market, calling that possibility “zero,” and rejected claims that downloading such models necessarily creates security backdoors, saying global scrutiny makes open models safer.
Huang additionally countered Wall Street concerns that the release of models such as Moonshot’s Kimi K
