Nvidia and several other technology giants on Monday launched a new artificial intelligence safety initiative focused on open models, as fallout from a cyberattack involving rogue OpenAI models continues to unfold.

The initiative comes after it emerged last week that Hugging Face, the startup targeted in the attack, was unable to rely on leading U.S. frontier models to defend its systems because their guardrails failed to distinguish between attackers and defenders.

The company instead turned to a self-hosted, open-weight Chinese model that was not subject to the same restrictions.

Amid growing efforts by United States lawmakers to limit the adoption of Chinese AI models, many of the most advanced of which are open-weight, major technology companies have launched an initiative to develop and share open AI tools.

Open models can be downloaded, modified and self-hosted, giving users greater control over how they are deployed.

This contrasts with closed models, including frontier systems developed by Anthropic and OpenAI, which can only be accessed through designated infrastructure.

“The Open Secure AI Alliance will work to remediate and disclose vulnerabilities using open technologies,” Nvidia said in a statement. “The recent Hugging Face security incident delivered a clear reminder: cyber defenders need open, frontier agentic systems for self-defense.”

Alongside Nvidia, the alliance includes Microsoft, SpaceX, Palantir and dozens of other technology companies from the United States and Europe.

There are growing calls to restrict access to AI models developed by Chinese companies, which have been accused of using “distillation” campaigns to extract information from the systems of U.S. rivals. The technique involves one AI model learning from the outputs and capabilities of a more advanced model.

Last week, Treasury Secretary Scott Bessent threatened to impose sanctions on Chinese companies that carry out distillation attacks against U.S. firms.