Nvidia unveiled the Vera Rubin AI computing platform at CES 2026, claiming up to 10x lower inference token costs and faster training for MoE models.

Nvidia used the Consumer Electronics Show (CES) as the backdrop for an enterprise scale announcement: the Vera Rubin NVL72 server rack platform for AI data centers, featuring new concepts and technology like “context memory” storage, zero downtime maintenance, rack-scale confidential computing, and several other advancements.

The balance of power in high-performance computing is tilting toward tightly fused CPU-GPU designs, and two old rivals are suddenly aligned around that future. Intel and Nvidia are now chasing a mega chip that marries general-purpose compute with accelerated AI,

Nvidia CEO Jensen Huang gave an outlook on the upcoming AI server DGX Vera Rubin with in-house ARM processor cores and new GPU architecture at CES.

Nvidia’s $20 billion strategic licensing deal with Groq represents one of the first clear moves in a four-front fight over the future AI stack. 2026 is when that fight becomes obvious to enterprise builders.