Michelle Hawley

NVIDIA told its largest customers that AI server prices will rise 15%+ on Grace Blackwell and Vera Rubin systems, driven by soaring memory chip costs.

The Model Context Protocol's updated roadmap adds five priority areas, including agent identity for non-human callers and progressive tool discovery.

SpaceX will launch its first NVIDIA-powered Starmind AI satellite in Q4 2027. The company has filed to deploy up to one million of them.

OpenAI published Jalapeño's first benchmarks, showing up to 1.9x more AI work per watt than NVIDIA Blackwell. Deployment starts late 2026.

Former NVIDIA scientist Anima Anandkumar turned down Bezos-backed Prometheus to build a physics AI using neural operators instead of Transformers.

Google Cloud is adding pay-as-you-go pricing, hard spending caps and off-peak discounts up to 50% to Gemini Enterprise as companies push back on AI bills.

Bill Gates published his most detailed AI essay in three years, warning of mass job loss, new security threats and a world with "no plan" for the transition.

Anthropic will rent 460 megawatts at Nscale's West Virginia campus over six years using NVIDIA Vera Rubin chips. Capacity comes online late 2027.

OpenAI added WebMCP support to ChatGPT's desktop browser. AI agents can now call structured tools on compatible sites instead of scraping interfaces.

Anthropic's Model Hardware Standard lets AI agents operate microscopes, robotic arms and other equipment. AWS and Universal Robots are early partners.