Nvidia introduced its latest AI chip platform, "Rubin," on the eve of the Computex technology conference in Taiwan.
Nvidia Unveils Next-Gen AI Chip Rubin
In a keynote address at National Taiwan University on Sunday, Nvidia CEO Jensen Huang revealed this next-generation platform in development called Rubin, slated for 2026, and the Blackwell Ultra chip, expected in 2025. It followed the March announcement of the Blackwell model, which is still in production and set for release later this year.
The Rubin platform is poised to succeed the upcoming Blackwell platform. It features cutting-edge GPUs, a new Arm-based CPU called Vera, and advanced networking capabilities with NVLink 6, CX9 SuperNIC, and the X1600 converged InfiniBand/Ethernet switch.
Huang said Nvidia plans to introduce a new version of its flagship AI chips yearly, aligning with its strategy of pushing technological limits and offering data center-scale solutions. The company previously said it had a two-year update timeline for its AI chips.
Computex Keynote of Nvidia
In a press statement, Nvidia said its latest innovations were showcased during the Computex keynote. The keynote was created using AI tools from members of the Nvidia Inception startup program and built on Nvidia NIM and the company's accelerated computing.
"Our company has a one-year rhythm. Our basic philosophy is very simple: build the entire data center scale, disaggregate and sell to you parts on a one-year rhythm, and push everything to technology limits," Huang noted.
Join the Conversation