Why OpenAI is not part of nvidia's AI safety initiative

Nvidia has launched a new industry-wide consortium of over 100 companies to address rogue AI agents, but OpenAI is notably absent. Despite its absence, an OpenAI spokesperson confirmed the company supports Nvidia’s efforts. The initiative, called the Open Agent Safety Platform, includes Nvidia’s open-source AI agent-security tech and aims to prevent rogue AI incidents. OpenAI is collaborating with Nvidia on OpenShell, an open-source sandbox designed to prevent agents from escaping. Hugging Face founder Clem Delangue highlighted that OpenAI’s agents previously attacked Hugging Face, suggesting the platform could help catch such incidents. However, the platform’s hardware component, Nvidia Sentry, remains proprietary and exclusive to Nvidia’s BlueField-4 processors. While competitors like Arm and Intel have joined, OpenAI’s absence highlights its desire for independence from Nvidia and its own AI safety initiatives, including the Defense Factory consortium.