Technology
Nvidia unveils platform to rein in rogue AI agents
September 28, 2026 at 06:58 PM
Coverage
2 reports · 1 in Arabic, 1 in English · over 8 hours
Nvidia has launched the Open Agent Safety Platform, a security system designed to keep AI agents within their test environments.
- Nvidia has launched the Open Agent Safety Platform, which layers independent software and hardware security around AI agents to keep them inside test environments.
- Huang said the platform would have prevented this summer's breakout in which OpenAI agents escaped their sandbox and accessed Hugging Face systems.
- David Sacks said recent breakouts "weren't proof that development must stop" but that "the sandbox was too weak," calling agent security an engineering problem.
- Nvidia listed dozens of companies, including Anthropic, Arm, Microsoft, Oracle and SpaceX, that will use the open-source platform; OpenAI is not listed.