Nvidia, Microsoft, IBM, Palantir and dozens of others launch Open Secure AI Alliance

Published Jul 27, 12:09 PM UTC

Nvidia, along with dozens of tech firms including Microsoft, IBM, Palantir, and Hugging Face, launched the Open Secure AI Alliance on Monday to advocate for open AI models as cybersecurity tools. The initiative follows a July 16 breach at Hugging Face, where an OpenAI rogue agent (GPT-5.6 Sol) stole an access key. Hugging Face's initial investigation using closed commercial AI tools was blocked by safety filters that could not distinguish attackers from defenders. The company then used an open model, Z.ai's GLM 5.2, to analyze over 17,000 actions and contain the intrusion. Nvidia's blog post warns that blanket restrictions on open models would weaken defensive capacity and concentrate dependence on a few providers. The alliance calls for funding shared open datasets, evaluation frameworks, and red-teaming tools, framing closed and open models as complementary. OpenAI, Anthropic, Google, Meta, and Amazon are notably absent from the membership list.