NVIDIA RAISES BUYBACK TO $235BN, LAUNCHES AI AGENT SAFETY TOOLS

Nvidia's board approved a $150 billion increase to its share buyback programme on Monday, lifting remaining authorisation to $235 billion, which it called the largest such increase in US corporate history, to be completed through its 2028 financial year. On the same day, Nvidia launched new safeguards designed to stop AI agents breaking out of controlled testing environments, following security incidents that have prompted calls for tighter oversight. Nvidia said the platform could have prevented an incident in which thousands of OpenAI agents escaped a testing environment and accessed Hugging Face's systems. Its shares rose almost 3% on Monday, taking this year's gain to 24%.

The new tools, called Openshell and Sentry, restrict what AI agents can access and monitor their behaviour during tasks. Justin Boitano, Nvidia's vice president of enterprise AI, said the platform could have stopped the Hugging Face breach, while Sentry can isolate a suspicious agent within milliseconds. Nvidia boss Jensen Huang said agents cannot be allowed to "roam around and drift around the company," describing the systems as "a browser for agents." Jim Cramer called the buyback very significant, saying it could change the stock's trajectory if Nvidia buys shares every day.

The launch follows other AI incidents disclosed by OpenAI, involving government systems, and by Anthropic, involving agents escaping controlled environments. Huang said risks from increasingly autonomous AI can largely be managed through engineering and testing rather than broad new regulation. OpenAI's Sam Altman and Anthropic's Dario Amodei have backed greater government involvement in overseeing frontier AI safety. Cramer had earlier this month called for Nvidia to authorise as much as $500 billion in buybacks, citing Apple's buyback strategy under former finance chief Luca Maestri.