Nvidia launches Rubin platform with six new AI chips.

Nvidia launched the Rubin platform, a coordinated set of six new chips designed to work as one AI supercomputer. The company said Rubin includes innovations such as updated NVLink interconnects, a Transformer Engine, Confidential Computing, a RAS Engine, and the Nvidia Vera CPU.

Nvidia announced the Rubin platform, a coordinated set of six new chips designed to function as a single AI supercomputer. The company said Rubin is intended to address rising demand for both training and inference workloads. Nvidia described five core innovations in the platform, including updated NVLink interconnects, a Transformer Engine, Confidential Computing, a RAS Engine, and the Nvidia Vera CPU. The announcement prompted some cloud providers to signal plans to adopt Rubin. Key points: - Nvidia launched the Rubin platform composed of six new chips. - Nvidia listed five innovations in Rubin, including NVLink, Transformer Engine, Confidential Computing, RAS Engine, and the Nvidia Vera CPU. - Nvidia's founder and CEO Jensen Huang said Rubin arrives as AI computing demand increases. - CoreWeave announced plans to add Rubin to its AI cloud and expects to deploy it in the second half of 2026. - One report noted that CoreWeave's shares moved in after-hours trading following the announcement. Summary: The Rubin launch introduces a new consolidated hardware platform intended to support larger and more complex AI workloads. Industry reaction includes CoreWeave's plan to deploy Rubin in the second half of 2026 and short-term market movement reported after the tie-up. Undetermined at this time.