NVIDIA's Open Agent Safety Platform ties governance to its own silicon
As seen on the 24/7 Wall St. homepage on September 28, 2026.
NVIDIA Launches Open Agent Safety Platform to Secure Agents From Testing to Deployment
More than 100 organizations, including Anthropic, Microsoft, JPMorganChase, Palantir and Salesforce, are already building on the new safety stack, which ties agent governance to NVIDIA's own Vera CPUs and BlueField-4 DPUs. Every enterprise that adopts these controls has another reason to keep buying NVIDIA silicon.
Continue ReadingShow less
NVIDIA announced the Open Agent Safety Platform, an open software platform and reference system design built to strengthen AI security across the full lifecycle of an agent, from initial testing all the way through live deployment. The platform delivers full-stack governance and control spanning both software and the underlying hardware, compute, and robotics systems that run those agents.
More than 100 organizations are already building on the new safety stack, a roster that includes Anthropic, Microsoft, JPMorganChase, Palantir, and Salesforce. That breadth of early adoption signals that the platform is being treated as infrastructure rather than a niche security add-on, which matters for how quickly governance standards could consolidate around NVIDIA's architecture.
The strategic edge for NVIDIA as a hardware company lies in how the platform is wired together. Agent governance is tied directly to NVIDIA's own Vera CPUs and BlueField-4 DPUs, meaning that enterprises choosing to adopt these safety controls have a concrete hardware dependency on NVIDIA silicon.
For investors watching NVIDIA's ability to extend its data-center moat, this is worth noting. Every enterprise that standardizes on the safety platform has a built-in reason to keep purchasing NVIDIA hardware, turning a software governance layer into a recurring demand driver for the underlying chips.
Mentioned: NVDA