On June 5, NVIDIA CEO Jensen Huang confirmed for the first time that three major memory chip manufacturers—Samsung Electronics, SK Hynix, and Micron Technology—have obtained certification to supply NVIDIA's U.S. company with the most advanced high-bandwidth memory product, HBM4, for its artificial intelligence accelerators. HBM4 is a key component of NVIDIA's next-generation AI platform, Vera Rubin. Huang revealed at Computex that Vera Rubin, scheduled for delivery in the third quarter of this year, has entered full production. The new system is built on NVIDIA's Vera CPU and Rubin GPU clusters, with each server system equipped with terabytes of HBM4. Additionally, Huang plans to further meet and have dinner with top South Korean business leaders on Friday.
