At CES 2026, NVIDIA CEO Jensen Huang announced that the next generation of AI chips is now in full production, delivering AI computing performance that is five times greater than the previous generation. The new platform, scheduled for official release later in 2026, is already being tested in the labs of leading AI companies.
Central to the release is the Vera Rubin supercomputing platform, composed of six distinct chips working in synergy. The flagship server integrates 72 graphics units and 36 new central processors. Huang demonstrated that linking over 1,000 Rubin chips into a "chipset" can increase AI token generation efficiency by tenfold. This performance leap is achieved with only a 1.6x increase in transistor count, largely enabled by NVIDIA’s proprietary data types, which the company aims to establish as a widely adopted industry standard.
For enhanced AI inference, Rubin chips introduce contextual memory storage technology, significantly improving response speed for chatbots handling long conversations and complex queries. NVIDIA also unveiled a new generation of networking switches equipped with co-packaged optics (CPO), offering low power consumption and high bandwidth to interconnect thousands of machines—directly challenging comparable solutions from Broadcom and Cisco. CoreWeave has been confirmed as the first platform to deploy Vera Rubin systems, with Microsoft, Oracle, Amazon, and Alphabet also planning integrations.
Beyond hardware, NVIDIA announced it will open-source Alpamayo software and training datasets for autonomous driving, helping automakers optimize decision-making and development. Additionally, NVIDIA recently acquired startup Groq, incorporating core AI chip design talent from Google to further strengthen its technological moat. Huang also noted strong demand in China for the H200 chip, stating that export license applications have been submitted and are pending government approvals.
While NVIDIA remains dominant in AI model training, its latest platform release and ecosystem expansion aim to solidify its leadership amid intensifying competition from AMD, Google, and others.
ICgoodFind Perspective: The launch of NVIDIA's next-generation computing platform is set to reshape the AI supply chain landscape. We are committed to helping industry players connect with high-end chip resources and seize new opportunities in this evolving market.
