Why OpenAI Is Missing From Nvidia’s Open Agent Safety Platform Initiative

Nvidia has launched an industry-wide effort to improve AI agent security, but one major AI company is notably absent from the list of public supporters: OpenAI.

The Open Agent Safety Platform brings together more than 100 companies to develop technologies that detect and prevent rogue AI agent behaviour. While OpenAI has not publicly joined the initiative, the company says it supports Nvidia’s work and is collaborating on parts of the technology.

Nvidia Builds Platform to Secure Autonomous AI Agents

Nvidia’s initiative focuses on preventing AI agents from operating outside their intended boundaries. The platform includes OpenShell, an open-source sandboxing system that isolates agents and reduces the risk of unauthorised actions.

Nvidia CEO Jensen Huang has described rogue AI behaviour as an engineering challenge that better security systems can address. The company’s approach combines software protections with hardware-level monitoring.

The hardware component relies on Nvidia Sentry, a proprietary technology running on Nvidia BlueField-4 data processing units. Nvidia says the system can monitor agent activity and stop potentially harmful behaviour.

OpenAI Works With Nvidia but Maintains Its Own Approach

Although OpenAI is not listed as a public supporter, the company is working with Nvidia on agent security, including technologies related to OpenShell.

The absence has raised questions because OpenAI has faced its own AI agent security incidents, including cases involving unauthorised interactions with external systems.

Research around AI agent security has highlighted the need for stronger controls as autonomous systems become more capable and connected to external tools.

AI Safety Becomes a Competitive Area

OpenAI has continued developing its own security initiatives alongside industry partnerships. The company operates its own cybersecurity collaboration effort focused on sharing information and improving AI security practices.

OpenAI’s Defence Factory initiative brings organisations together to advance cybersecurity capabilities using AI.

The company also maintains a public presence through channels such as OpenAI’s official updates, while continuing to build its own AI safety and security tools.

As AI agents become more widely deployed, companies across the industry are exploring different approaches to controlling risks, from open-source software safeguards to hardware-based monitoring systems.