Why OpenAI is absent from Nvidia's agent safety effort

Nvidia has organized an industry-wide push called the Open Agent Safety Platform, with the stated goal of reining in rogue AI agents. OpenAI, one of the most prominent builders of agentic products, is not listed among the effort's public supporters.

That absence is not the whole story. TechCrunch reports that OpenAI is working with Nvidia on the effort privately, even though it has not signed on in public. The report describes a gap between the public roster and the actual collaboration.

The phrase 'rogue AI agents' points at a now-familiar worry: systems that can call tools and run multi-step tasks may behave in ways their operators did not intend. Nvidia's choice to answer with a shared platform signals that agent safety is being treated as an ecosystem problem rather than a single-vendor one.

OpenAI's stance is the open question. As a major supplier of agent products, whether it joins, and on what terms, will shape how much of the ecosystem such a framework can actually cover. A public absence paired with private engagement also suggests that vendors' safety postures and their engineering work do not always move in step.

What to watch next: the specific mechanisms Nvidia puts behind the platform, which companies formally sign on, and whether OpenAI's private work becomes a public commitment. Those details will determine where autonomous agents can safely be deployed in real business settings.