Nvidia Doubles AI Chip Revenue Forecast to One Trillion Dollars at GTC 2026

Nvidia Sets Sights on a Trillion-Dollar AI Chip Empire

Jensen Huang took the stage at the SAP Center in San Jose on Monday to deliver a blockbuster GTC 2026 keynote, doubling Nvidia's AI chip revenue forecast to more than one trillion dollars in cumulative revenue from its Blackwell and Vera Rubin platforms through the end of 2027. The previous outlook had pegged the opportunity at five hundred billion dollars through 2026.

Vera Rubin Platform Enters Full Production

The hardware centrepiece of the event was the Vera Rubin platform, now in full production, which Nvidia says delivers five times the inference performance and three and a half times the training performance of Blackwell. The Rubin GPU packs three hundred and thirty-six billion transistors and up to two hundred and eighty-eight gigabytes of HBM4 memory per chip. Nvidia also unveiled the Vera CPU, featuring eighty-eight custom Arm-based Olympus cores designed to handle the demands of agentic AI workloads.

Agentic AI and Open-Source Push

Nvidia launched NemoClaw, an open-source platform enabling enterprises to build and deploy autonomous AI agents with built-in security and privacy guardrails. The platform is notably hardware-agnostic, a departure from Nvidia's traditionally proprietary ecosystem. The company also formed the Nemotron Coalition to advance open frontier models alongside partners including Perplexity and Black Forest Labs.

Market Dominance and Future Roadmap

With hyperscalers expected to spend more than six hundred and sixty billion dollars on AI infrastructure this year, Nvidia commands an estimated eighty to ninety percent of the AI chip market. Huang revealed a three-generation GPU roadmap stretching through the Feynman architecture in 2028 and announced a gigawatt-scale deployment deal with Mira Murati's Thinking Machines Lab, the first major commercial commitment to Vera Rubin.