Nvidia and a group of U.S. and European technology companies launch the Open Secure AI Alliance, focused on building and sharing open-source tools to help defend AI models and AI-powered agents from cyberattacks. Multiple outlets report that the initiative is intended to provide security teams with resources for testing, auditing, identifying vulnerabilities, and helping patch and disclose security issues across infrastructure. The alliance’s membership includes companies such as Microsoft, SpaceX, IBM, Palantir, and others, bringing the total to several dozen founding participants. Several sources link the effort to recent concerns and incidents involving major AI systems, including fallout from OpenAI’s disclosed cyberattack and subsequent scrutiny of how “frontier” and freely available models are secured.
A key element of the alliance, repeatedly noted across coverage, is that major frontier AI companies are not part of the group. Reported participants say the alliance does not include OpenAI, Google, or Anthropic, and it emphasizes that defenders can use open models and tools they can run themselves. The new effort builds on earlier work and is described as aligned with broader open-source security initiatives.
