Nvidia approves Vera Rubin HBM4 supply from three major chipmakers

Traders Union
All News
Financial News
Nvidia approves Vera Rubin HBM4 supply

Highlights

Nvidia cleared SK Hynix, Samsung and Micron to supply HBM4 for Vera Rubin.
The approval supports mass production of memory for Nvidia’s next AI accelerator cycle.
Samsung and Micron have both highlighted HBM4 production tied to Vera Rubin.
The decision reduces supply-chain uncertainty but does not remove capacity constraints.

Three suppliers approved

According to Bloomberg, Chief Executive Jensen Huang said Nvidia’s Vera Rubin platform is in full production and will use HBM4 from Samsung, SK Hynix and Micron. The confirmation ends months of speculation over whether Nvidia would rely mostly on South Korean suppliers or broaden the allocation to include Micron.

HBM4 is a critical part of Vera Rubin because AI accelerators increasingly depend on memory bandwidth, not only raw computing power. Nvidia’s Vera Rubin platform is designed as a rack-scale system combining GPUs, CPUs, networking and infrastructure into one integrated AI supercomputer architecture.

Samsung and Micron push into production

Samsung has said its sixth-generation HBM4 is in mass production and designed for Nvidia’s Vera Rubin platform. The company said the memory can deliver 11.7 gigabits per second, above the industry standard of 8 Gbps, with potential enhancement to 13 Gbps.

Micron separately said it had entered high-volume production of 36GB 12-high HBM4 for Vera Rubin. The company said the product delivers more than 2.8 terabytes per second of bandwidth and more than 20% better power efficiency than HBM3E, positioning it as a stronger competitor in a market long dominated by SK Hynix and Samsung.

SK Hynix remains a central supplier in the AI memory chain. Nvidia’s Huang met SK Group Chairman Chey Tae-won in Taipei this week, and SK Hynix has outlined plans to expand memory wafer capacity as AI demand keeps supply tight.

AI supply chain moves into its next phase

The approval matters because HBM supply has become one of the main limits on AI infrastructure growth. Large cloud providers and enterprise customers need more accelerators, but each new system requires advanced memory that is difficult to manufacture at scale.

For Nvidia, adding all three major suppliers lowers dependency risk and gives the company more room to ramp Vera Rubin shipments. For memory makers, the clearance confirms that HBM4 will be one of the most valuable battlegrounds in semiconductors as AI spending shifts from current Blackwell systems to the next generation.

As we previously reported, Nvidia aims to "reinvent the PC" with new processors.

This material may contain third-party opinions, none of the data and information on this webpage constitutes investment advice according to our Disclaimer. While we adhere to strict Editorial Integrity, this post may contain references to products from our partners.