NVIDIA Vera Rubin has officially entered full production with Samsung AI memory onboard. NVIDIA’s new-gen AI platform, which uses memory from Samsung and SK Hynix, will power agentic AI factories worldwide.

CEO Jensen Huang announced that the NVIDIA Vera Rubin AI accelerator has entered full-scale production, and the supply chain knows that it sources advanced LPDDR5X memory chips from Samsung, SK Hynix, and Micron.

At the keynote, NVIDIA CEO emphasized that Vera has dramatically improved the performance of AI agents. Huang also introduced Vera Compute, which features eight CPUs, and Vera BlueField, a storage-specialized variant.

“Vera Rubin is currently in full production,” Huang said during his keynote speech at GTC Taipei.

By becoming the first company to mass-produce HBM4 and then quickly moving to 12-layer HBM4E samples, Samsung Electronics has positioned itself at the center of Nvidia’s upcoming Vera Rubin transition.

Related article: Samsung unveils 12-layer HBM4E AI memory with faster speeds and improved efficiency

The post NVIDIA Vera Rubin AI era begins with Samsung memory appeared first on Sammy Fans.