Nvidia GTC 2026 Kicks Off With Asia's Chip Giants in the Front Row

Jensen Huang takes the stage in San Jose today for Nvidia's annual GTC conference, where he has promised to unveil "a chip that will surprise the world." The spotlight falls on the Vera Rubin platform, Nvidia's next-generation GPU architecture reportedly packing up to 288GB of HBM4 memory. SK Group chairman Chey Tae-won is attending GTC for the first time to discuss expanding SK hynix's supply of high-bandwidth memory for Nvidia's new chips. In a parallel move, AMD chief Lisa Su visits Seoul on 18 March to meet Samsung Electronics and Naver executives about HBM supply, underscoring just how central Korean memory makers have become to the global AI hardware race.
Why it matters for Asia
South Korea's semiconductor heavyweights are no longer just suppliers - they are strategic partners shaping the trajectory of AI compute. SK hynix and Samsung are the only two companies capable of manufacturing HBM4 at scale, giving Seoul outsized leverage in the AI infrastructure buildout. For enterprise buyers across Asia-Pacific, the pace of next-generation chip rollouts will directly determine when more powerful AI workloads become commercially viable in the region.^


