LAS VEGAS: NVIDIA on Monday launched its next-generation Vera Rubin artificial intelligence computing platform at the Consumer Electronics Show, promising a dramatic leap in AI performance that could reshape everything from data centers to driverless vehicles.
The Rubin system — a tightly integrated suite of six chips combining the Rubin GPU, the Vera CPU and advanced networking and data-processing components — is designed to handle the increasing computational demands of modern AI. Company executives say it can deliver roughly five times the AI performance of its predecessor and cut operating costs for large models by up to 90%. Products built on the platform are expected to begin rolling out through cloud and enterprise partners in the second half of 2026.
“NVIDIA is doubling down on performance and efficiency with Rubin,” CEO Jensen Huang said in a keynote at the Las Vegas show, adding that the new platform will support expansive AI models and applications.
Alongside the hardware, NVIDIA introduced a family of open-source AI models designed for real-world uses such as autonomous vehicles and robotics. The Alpamayo model suite aims to give self-driving systems enhanced reasoning capabilities, enabling them to navigate complex environments and explain their decisions — a critical step toward broader deployment of autonomous technologies.
Industry experts see the Rubin launch as part of NVIDIA’s bid to maintain leadership amid rising competition in AI silicon and software. By pairing powerful chips with open model frameworks, the company hopes to accelerate innovation across sectors including healthcare, robotics and transportation.
As automotive partners adopt these tools, NVIDIA’s advances may translate into safer, more capable self-driving systems in the coming years, with broader commercialization still tied to testing and regulatory approvals.