Nvidia unveils next-generation Rubin AI chip platform

At the Consumer Electronics Show today, Nvidia CEO Jensen Huang introduced the company's new Rubin computing architecture, calling it the pinnacle of AI hardware technology. The architecture is already in production and is set to scale up significantly in the latter half of the year.

"Vera Rubin was created to tackle a core challenge we're facing: the explosive growth in computational power required for AI," Huang told attendees. "I'm pleased to announce that Vera Rubin is now in full production."

First unveiled in 2024, the Rubin architecture represents the latest achievement in Nvidia's accelerated hardware development, which has propelled the company to become the world's most valuable corporation. Rubin succeeds the Blackwell architecture, which itself replaced the earlier Hopper and Lovelace designs.

Rubin chips have already been adopted by nearly every major cloud provider, including through Nvidia's prominent partnerships with Anthropic, OpenAI, and Amazon Web Services. Rubin systems will also power HPE's Blue Lion supercomputer and the forthcoming Doudna supercomputer at Lawrence Berkeley National Laboratory.

Named after astronomer Vera Florence Cooper Rubin, the architecture comprises six specialized chips engineered to work together. While the Rubin GPU serves as the centerpiece, the architecture also tackles increasing storage and interconnection bottlenecks through enhancements to Bluefield and NVLink systems. It also introduces the new Vera CPU, optimized for agentic reasoning tasks.

Nvidia's senior director of AI infrastructure solutions, Dion Harris, highlighted the advantages of the new storage system by pointing to the growing memory demands of modern AI systems related to caching.

"Emerging workflows like agentic AI and long-term tasks place significant pressure on your KV cache," Harris explained to journalists, referring to the memory system AI models use to compress inputs. "We've introduced an external storage tier that connects to the compute device, enabling much more efficient scaling of your storage capacity."

Join the Disrupt 2026 Waitlist

Secure your spot on the Disrupt 2026 waitlist for priority access when Early Bird tickets become available. Previous Disrupt events have featured industry giants like Google Cloud, Netflix, Microsoft, Box, Phia, a16z, ElevenLabs, Wayve, Hugging Face, Elad Gil, and Vinod Khosla among 250+ leaders delivering 200+ sessions designed to accelerate your growth and competitive advantage. You'll also connect with hundreds of startups driving innovation across all industries.

Join the Disrupt 2026 Waitlist

Secure your spot on the Disrupt 2026 waitlist for priority access when Early Bird tickets become available. Previous Disrupt events have featured industry giants like Google Cloud, Netflix, Microsoft, Box, Phia, a16z, ElevenLabs, Wayve, Hugging Face, Elad Gil, and Vinod Khosla among 250+ leaders delivering 200+ sessions designed to accelerate your growth and competitive advantage. You'll also connect with hundreds of startups driving innovation across all industries.

As anticipated, the new architecture delivers substantial gains in both speed and power efficiency. Nvidia's testing shows Rubin operates 3.5 times faster than Blackwell for model training and five times faster for inference tasks, achieving up to 50 petaflops. The platform also delivers eight times more inference compute per watt.

These advancements arrive during intense competition to build AI infrastructure, with both AI labs and cloud providers racing to secure Nvidia chips and the facilities needed to support them. During an October 2025 earnings call, Huang projected that $3 trillion to $4 trillion will be invested in AI infrastructure over the next five years.

Stay updated with all of TechCrunch's coverage of the annual CES conference here.

Watch Nvidia CEO Jensen Huang unveil what he described as the state of the art in AI hardware: the new Rubin computing architecture.

“Vera Rubin is designed to address this fundamental challenge that we have: The amount of computation necessary for AI is skyrocketing.” Huang… pic.twitter.com/MhGVqytX04

— TechCrunch (@TechCrunch) January 5, 2026