Nvidia is heading up a new alliance of tech industry giants focused on keeping open-source AI models secure as Washington debates how and whether to restrict them.
“The right response is not to deny defenders access to capable open systems. It is to pair openness with strong safeguards, clear rules against malicious misuse, rigorous evaluation and rapid remediation,” Nvidia writes in a blog post Monday morning launching the new initiative, called the Open Secure AI Alliance.
A range of tech and cybersecurity companies are joining the effort, including Palantir, IBM, Crowdstrike, SpaceX, and Hugging Face — the company recently hacked by rogue closed models produced by OpenAI.
Nvidia says it plans to contribute open models, data, and other resources to the alliance to help “speed the development of new cybersecurity tools and techniques.”
The blog post urges policymakers to “recognize open models, harnesses and security tooling as defensive assets, not liabilities, in AI and cybersecurity policy” and argues that “blanket restrictions on open frontier AI systems would weaken defensive capacity and risk concentrating power, dependence and vulnerability in a few closed providers.”
Chinese open-weight models are gaining popularity against the leading American models, like those developed by OpenAI and Anthropic, which are largely closed and cost significantly more to use.
The White House is currently mulling potential limits on access to Chinese open-source models; Treasury Secretary Scott Bessent said last week that the administration would scrutinize Chinese models for potential intellectual property theft, and the administration later accused Moonshot AI’s Kimi K3 of stealing American tech.