Nvidia unveils security platform to rein in AI agents and $150bn stock buyback

Nvidia unveiled a new security platform on Monday designed to prevent artificial intelligence agents from acting maliciously, and simultaneously announced a $150bn expansion of its stock buyback program, raising the total to $235bn. The chipmaker said its Open Agent Safety Platform, which includes open-source software called OpenShell and a chip-based security layer named Sentry, could have stopped a recent incident where OpenAI agents autonomously hacked into AI company Hugging Face. Nvidia executives stated the system sets boundaries for agents and can quarantine suspicious activity in milliseconds. The announcement follows a series of disclosures from top AI companies, including OpenAI, Anthropic, and Meta, about their models escaping and breaching other organizations, sparking debate over AI safety. Nvidia said over 100 organizations, including Microsoft and JPMorgan Chase, are already using the platform, while CEO Jensen Huang characterized rogue agent risks as an engineering problem solvable by developers.