For months, the semiconductor industry watched closely as Samsung, SK Hynix, and Micron raced to qualify their HBM4 memory for Nvidia's next-generation AI platform. That waiting is now over.
Nvidia CEO Jensen Huang confirmed on June 5, 2026 that Samsung Electronics, SK Hynix, and Micron Technology have all passed certification to supply HBM4 high-bandwidth memory for the company's next-generation Vera Rubin AI accelerator platform marking the first public acknowledgment of approvals for all three major memory manufacturers.
What Is Vera Rubin?
Vera Rubin entered full production after being announced at the GTC Taipei keynote on June 1, 2026. Nvidia says it delivers 10x agent throughput at scale compared with the Grace Blackwell platform. It is designed specifically for what Huang calls "agentic AI" workloads the next frontier of large-scale AI deployment where systems act autonomously rather than simply responding to prompts.
Who Gets What Share?
All three suppliers are certified, but not equally positioned. Supply-chain analysts estimate SK Hynix holds roughly 60–70% of Vera Rubin HBM4 volume, with Samsung capturing approximately 25–30% and Micron supplying the remainder. SK Hynix entered the qualification process ahead of rivals, while Samsung began HBM4 mass production in February 2026.
Jensen Huang's South Korea Visit
The certification announcement was made during Huang's trip to Seoul, where the stakes could not be higher. Huang is expected to meet with chairs of major conglomerates including SK Group, Samsung, LG Group, Hyundai Motor Group, and Naver to discuss supply ramp commitments and physical AI partnerships.
The visit carries symbolic weight as well. On the opening day of Computex 2026, Huang visited the SK Hynix booth, wrote "Please Make More" on an HBM4E wafer on display, and signed his name alongside the message. It was an informal but pointed signal of just how tight AI memory supply remains globally.
On June 2, Huang publicly urged SK Hynix to produce more HBM chips, telling reporters that global semiconductor supply remains tight a statement underscoring the demand pressure Nvidia faces as it scales Vera Rubin deployments.
Why This Matters
Having all three major memory manufacturers certified for Vera Rubin significantly reduces Nvidia's supply-chain risk as it scales HBM4 production for what is expected to be enormous global AI infrastructure demand. It also marks a meaningful milestone for Samsung, which spent much of early 2026 working to close the qualification gap with SK Hynix and has now secured its position on the world's most strategically important AI platform.
You must be logged in to post a comment.