Nvidia recently announced a new consortium of over 100 companies dedicated to solving the challenges posed by rogue AI agents. However, one notable omission from the Open Agent Safety Platform is OpenAI, even though major industry players like its rival Anthropic have joined the initiative.
Although OpenAI has not publicly signed on as a supporter, a company spokesperson stated that it supports Nvidia’s work and continues to collaborate on key open-source components like the OpenShell sandbox. OpenAI’s absence is widely seen as an effort to maintain independence from its major investor, Nvidia, especially since the platform relies on proprietary hardware features, such as Nvidia’s BlueField-4 processing units, for full functionality.
While Nvidia leverages its hardware ecosystem to enforce agent safety at the silicon level with Nvidia Sentry, OpenAI is carving its own path. The company is actively developing proprietary cybersecurity safeguards and information-sharing consortia, aiming to establish its own leadership in enterprise AI security.
- Nvidia launched an alliance of 100+ companies to tackle rogue AI agents.
- OpenAI is notably absent from the public roster despite technical cooperation.
- The safety platform relies heavily on Nvidia’s proprietary hardware components.
- OpenAI continues to develop its own independent cybersecurity solutions and consortia.
Sources:
