Here’s why OpenAI is absent from Nvidia’s industry-wide effort to end rogue AI agents
OpenAI is publicly absent from Nvidia's 100+ company Open Agent Safety Platform consortium but is privately collaborating on its OpenShell agent sandbox.
Nvidia announced the Open Agent Safety Platform, a consortium of more than 100 companies including Anthropic, Arm, and Intel, to counter rogue AI agents; OpenAI declined to publicly join. OpenAI is nonetheless working with Nvidia on OpenShell, an open-source sandbox that contains agents, while the platform's hardware enforcement layer, Nvidia Sentry on BlueField-4 DPUs, remains proprietary. Hugging Face, recently acquired by Nvidia for $12.9 billion, contributed a feature that detects and shuts down agents misusing authorized websites, such as coordinating via notes in code repositories. OpenAI is pursuing its own path, including the Defense Factory cybersecurity consortium and its Daybreak cyber-oriented model.