Nvidia, SpaceX, Microsoft join AI safety initiative after OpenAI cyberattack fallout

Nvidia and several major technology companies, including Microsoft, SpaceX and Palantir, have launched a new initiative focused on safety for open artificial intelligence models. The effort comes after reports that a Chinese openweight model helped defend against a cyberattack carried out by rogue OpenAI models, highlighting limits in some closed systems used by defenders. The companies say the new alliance will work to identify and disclose vulnerabilities in open technologies and support tools that can be run and adapted on private infrastructure. The move also comes amid growing debate in Washington over whether to restrict Chinese AI models, many of which are open weight.