NVIDIA GTC 2026: Jensen Huang Unveils Vera Rubin Platform, Dynamo Inference Software, and $1T Revenue Forecast
Tags Infrastructure · AI

At GTC 2026 in San Jose (30,000+ attendees), NVIDIA CEO Jensen Huang unveiled the Vera Rubin AI infrastructure platform — a five-rack system integrating Rubin R100 GPUs (336 billion transistors), Groq LPUs for low-latency inference, and Dynamo 1.0 inference orchestration software. Dynamo achieves a claimed 35x improvement in tokens-per-watt via disaggregated prefill/decode, routing prefill to Vera CPUs and decode to Rubin GPUs. Huang raised NVIDIA's revenue forecast to $1 trillion through 2027, up from $500B for 2026, citing surging inference demand. Vera Rubin ships H2 2026, with Rubin Ultra targeted for 2027. NVIDIA also announced the DSX AI Factory reference design, Omniverse DSX Blueprint for digital twin data centers, the NemoClaw agent stack, and the Nemotron Coalition for open models. The announcements mark the industry's decisive pivot from training-centric to inference-centric infrastructure as agentic AI workloads scale.