Key Takeaways
- Jensen Huang, CEO of NVIDIA, announced that Samsung, SK Hynix, and Micron have successfully qualified to provide HBM4 memory chips for the upcoming Vera Rubin AI accelerator system.
- The Vera Rubin platform has commenced full-scale production and is scheduled for customer shipments in the third quarter of 2026, offering 10 times the agent throughput of Grace Blackwell.
- Industry analysts project SK Hynix will supply 60–70% of Vera Rubin’s HBM4 requirements, while Samsung captures 25–30%, leaving Micron with the balance.
- Micron shares fell 7.7% on June 5 despite certification news, weighed down by tech sector sell-off triggered by strong employment data and Broadcom earnings.
- The announcement came during Huang’s visit to South Korea, where he engaged with executives from SK, Samsung, LG, Hyundai, and Naver regarding supply expansion and AI collaboration.
On June 5, 2026, NVIDIA’s chief executive Jensen Huang publicly verified that the three leading memory chip producers—Samsung, SK Hynix, and Micron—have completed the qualification process to deliver HBM4 high-bandwidth memory for the company’s next-generation Vera Rubin AI accelerator lineup.
Shares of NVDA closed at $218.66, registering a 1.82% gain for the session, whereas Micron (MU) dropped 7.74%.
Speaking to journalists after landing in Seoul, Huang stated: “All three vendors have been qualified. All three vendors are in production, and they’re all racing to support Vera Rubin.”
Vera Rubin represents the next evolution beyond NVIDIA’s Grace Blackwell GPU platform. Production ramped to full capacity following Huang’s presentation at GTC Taipei on June 1, 2026, where NVIDIA revealed the architecture delivers ten times the agent processing capability of its Grace Blackwell predecessor.
“Agentic AI is a new kind of workload,” Huang explained. “One prompt can launch a thousand-step journey of reasoning, retrieval, tool use and response generation. Vera Rubin was built for this moment.”
The company expects initial shipments of the new platform to reach customers during the third quarter of 2026. Vera Rubin server configurations combine NVIDIA’s Vera central processors with Rubin graphics processors and massive volumes of HBM4 memory.
This represents the first public confirmation from Huang that all three memory manufacturers have achieved certification for the platform, putting to rest months of uncertainty surrounding the supply chain.
Projected Distribution of HBM4 Supply Volumes
Though NVIDIA has not released official allocation figures, industry supply-chain experts believe SK Hynix will capture between 60% and 70% of the HBM4 memory volume for Vera Rubin systems. Samsung is anticipated to secure approximately 25–30% of the allocation, with Micron providing what remains.
SK Hynix initiated the qualification process before its rivals. Samsung launched HBM4 volume production in February 2026. On June 2, Huang publicly encouraged SK Hynix to accelerate production even further, emphasizing that worldwide semiconductor availability continues to face constraints.
NVIDIA is simultaneously establishing a new research and development facility in South Korea and has begun hiring personnel for the operation.
Micron Shares Decline Despite Qualification Confirmation
The favorable supply chain news failed to boost Micron shares. MU tumbled 7.74% on June 5, caught in widespread technology sector selling pressure that followed an unexpectedly robust U.S. employment report, which unsettled rate-sensitive technology holdings.
The decline also occurred after Broadcom released earnings results Wednesday evening, which seemed to soften investor enthusiasm across the AI chip sector.
While in Seoul, Huang conducted meetings with leadership from SK Group, Samsung, LG Group, Hyundai Motor Group, and Naver to address production capacity increases and partnerships around physical AI applications.
NVIDIA disclosed that it has arranged discussions with all significant South Korean technology and industrial conglomerates as part of its strategy to expand Vera Rubin deployment on a worldwide scale.
