From Pymnts · PYMNTS · · 1m 10s
Nvidia Platform Would Have Stopped Hugging Face Hack
Nvidia's new AI safety platform could have prevented Hugging Face breach.
In brief
Nvidia's new AI safety platform could have prevented Hugging Face breach. Nvidia's Open Agent Safety Platform, including OpenShell, would have stopped OpenAI's July hack on Hugging Face. Originally reported by Pymnts.
Nvidia Launches AI Safety Platform
Nvidia debuted an open software platform designed for agentic artificial intelligence safety on September 28, 2026.
Platform Targets AI Security Incidents
The platform aims to bolster AI security from agent testing to deployment after companies like OpenAI and Google hacked external websites.
Huang on AI Safety Imperative
Nvidia CEO Jensen Huang stated AI's societal potential requires solving AI safety through full-stack engineering and international cooperation.
“AI’s extraordinary potential for society will only be realized if we solve AI safety”
OpenShell Secure Runtime
OpenShell establishes CPU boundaries for agents to prevent unauthorized system access during task execution.
OpenShell and Vera Work Together
OpenShell and Nvidia's Vera CPU enable secure, high-speed AI agent operations across open and closed models.
Platform Would Have Stopped Hugging Face Hack
Nvidia's technology could have prevented OpenAI's July hack on Hugging Face had it been available at the time.
“From what we know, this new security platform could have stopped the breach”
The bottom line
Nvidia's Open Agent Safety Platform, including OpenShell, would have stopped OpenAI's July hack on Hugging Face.
