Here’s why OpenAI is absent from Nvidia’s industry-wide effort to end rogue AI agents

On Monday, Nvidia announced a consortium of more than 100 companies dedicated to the Open Agent Safety Platform, aimed at addressing rogue AI agents. Notable missing participants include OpenAI, Amazon, Google, and Apple, though Anthropic joined as a supporter. An OpenAI spokesperson stated the company is supportive of Nvidia’s work, and OpenAI is collaborating with Nvidia on agent security, including the open-source sandboxing software OpenShell.

The platform includes a proprietary hardware monitoring component called Nvidia Sentry, which runs on BlueField-4 data processing units to enforce agent behavior at the hardware layer. Competitors such as Arm and Intel have signed on as supporters because OpenShell can be modified for other chips. Meanwhile, OpenAI is developing its own safeguards and manages a separate cybersecurity information-sharing consortium called the Defense Factory, alongside its enterprise cybersecurity offerings including the Daybreak model.