Nvidia CEO Huang confirms Vera Rubin in full production with Korean HBM chips - ANI News

NVIDIA CEO Jensen Huang confirmed that the company's next-generation AI accelerator, Vera Rubin, has entered full production at the GTC Taipei conference. The platform will incorporate high-bandwidth ... Read original →

Jensen Huang’s confirmation of Vera Rubin’s mass production signals a paradigm shift in AI chip design—from compute-centric to memory-bandwidth-driven. With HBM4 costs surging 435%, NVIDIA’s multi-sourcing strategy mitigates supply risk while cementing SK hynix’s pricing power (60–70% share). Samsung, shipping early HBM4E samples, aims to reclaim lost ground by targeting the late-2027 Vera Rubin Ultra. Technically, co-packaging 3nm logic with 7th-gen HBM intensifies competition for CoWoS capacity between TSMC and Korean OSATs. Geopolitically, Huang’s Seoul visits underscore NVIDIA’s push to build a U.S.-Korea-aligned, China-excluded HBM supply chain. Over the next 18 months, HBM availability will dictate AI accelerator margins; any U.S.-ROK export controls on advanced packaging could leave Chinese AI firms stranded with chips but no memory.