NVIDIA Unveils Rubin AI Platform, Alpamayo Autonomous Driving Models, and Physical AI Advancements at CES 2026

Jensen Huang, the founder and CEO NVIDIA, presently the super-semiconductor company in the world, delivered the opening keynote at CES 2026 in Las Vegas, announcing the Rubin AI platform now in full production, the Alpamayo family of open reasoning models for autonomous vehicles, new physical AI models and tools for robotics, and updates across gaming and personal AI systems.

Huang introduced the NVIDIA Rubin platform as the company's first extreme-codesigned, six-chip AI platform and successor to the Blackwell architecture. Components include Rubin GPUs delivering 50 petaflops of NVFP4 inference, Vera CPUs for data movement and agentic processing, NVLink 6 scale-up networking, Spectrum-X Ethernet Photonics scale-out networking, ConnectX-9 SuperNICs, and BlueField-4 DPUs. The platform incorporates the NVIDIA Inference Context Memory Storage Platform, powered by BlueField-4, which provides a KV-cache tier for long-context inference with 5x higher tokens per second, 5x better performance per TCO dollar, and 5x better power efficiency. Rubin reduces inference token generation costs to approximately one-tenth of the previous platform and trains mixture-of-experts models with 4x fewer GPUs compared to Blackwell.

Pic: BlueField<...