NVIDIA’s annual GPU Technology Conference (GTC) 2026 began on Monday, March 16, with CEO Jensen Huang set to deliver a keynote at the SAP Center in San Jose, California. The four-day event is being watched closely by investors and industry analysts as NVIDIA presents new developments in Artificial Intelligence, GPUs, data centers, robotics, and Artificial Intelligence agents. A central theme is NVIDIA’s strategy of reinvesting profits into the Artificial Intelligence ecosystem as competition in Artificial Intelligence chips intensifies.
Reports indicate that Huang will reveal a new Artificial Intelligence chip developed using technology from Groq, the Artificial Intelligence inference startup NVIDIA acquired for $20 billion last year. The non-exclusive licensing deal, which included recruitment of Groq’s key personnel, is NVIDIA’s largest technology purchase to date. NVIDIA is also expected to show how Groq’s inference technology works with its CUDA platform, and analysts expect a new class of servers that combine NVIDIA networking with Groq’s specialized chips to speed up Artificial Intelligence inference and lower costs.
NVIDIA may also present CPU-only servers, pointing to a broader focus on central processors alongside its established GPU business. An updated roadmap for next-generation GPUs, including the Vera Rubin family set to ship in the second half of 2026, is also expected. Together, those announcements would show NVIDIA expanding its hardware strategy across inference, networking, CPUs, and future GPU products.
Beyond silicon, NVIDIA is likely to highlight NemoClaw, an open-source Artificial Intelligence agent platform aimed at enterprise deployments with security, privacy protection, and scalable automation. The platform is expected to integrate with NVIDIA’s NeMo framework and support multiple hardware architectures, including Intel, AMD, and NVIDIA processors. That positioning suggests NVIDIA wants its software stack to remain relevant across a wider range of systems, not only its own chips.
Robotics and physical Artificial Intelligence are also expected to be major parts of the event. Those efforts build on last year’s keynote, which introduced Newton, a physics engine-powered robot, and the Groot N1 Artificial Intelligence foundation model. Analysts say physical Artificial Intelligence could emerge as a multi-trillion-dollar industry in the coming decades, with applications spanning autonomous systems, healthcare, and industrial automation. The event is set to present NVIDIA’s broader vision for computing through a mix of hardware, software, and robotics.
