Wire · technology
NVIDIA Launches Open Agent Safety Platform to Secure Agents From Testing to Deployment
◆ Sectors
◆ Geography
◆ Source
◆ Verified
Fusion42 · 28 September 2026 · Fusion42 review
NVIDIA has launched the Open Agent Safety Platform, combining open-source software and hardware components to enforce full-stack governance and security for AI agents from testing to deployment. The platform includes OpenShell software, which sets secure runtime boundaries on CPUs, and Sentry, a hardware watchdog running on BlueField-4 DPUs that quarantines agents acting outside allowed parameters.
This Wire brief sits within Fusion42's coverage of AI Infrastructure, AI Agents and Cybersecurity, and 13 sources have reported it between 28 Sep 2026 and 3 Oct 2026.
◆ ◆ The Wire takeaway
You face new infrastructure tools that enforce AI agent boundaries at hardware and software levels. Integrate these to prevent rogue agent behaviour before deployment and meet rising safety expectations in AI systems.
◆ Coverage
13 sources · first reported 28 Sep 2026 · latest 3 Oct 2026
◆ Related on Wire
◆ Topics