NVIDIA's Memory Configuration Adjustment Triggers Volatility in the Storage Market

iconKuCoinFlash
Share
Share IconShare IconShare IconShare IconShare IconShare IconCopy
AI summary iconSummary

expand icon
Market volatility hit the storage sector on June 6, 2026, after SemiAnalysis reported that NVIDIA may reduce the SOCAMM memory on its Vera Rubin NVL72 platform from 192GB to 96GB. The adjustment could lower per-rack costs by 10%, triggering a sharp sell-off in global storage stocks. Analysts quickly clarified that this change is a temporary optimization and does not affect HBM demand for GPU computing. Despite the volatility, many remain bullish on AI-driven storage demand and potential price gains.

BlockBeats news: On June 6, global memory sectors experienced a collective pullback following a report from SemiAnalysis regarding adjustments to the memory configuration of NVIDIA’s next-generation Vera Rubin NVL72 server platform. On June 5, SK Hynitz plunged 9.92%, while China’s A-share memory index fell approximately 4%, with leading declines seen in stocks such as Biwin Storage,江波龙 (Wave Memory), Memblaze, and GigaDevice. The previous trading day, Micron Technology closed down 7.74%.


Reports indicate that NVIDIA plans to reduce the SOCAMM system memory capacity paired with the Vera CPU from 192GB to 96GB, expected to lower the cost per rack from $7.6 million to $6.8 million, saving approximately 10%. The market initially worried that demand for AI server memory might be weakening.


However, several institutions subsequently pointed out that this adjustment only involves the detachable SOCAMM memory modules on the CPU side, representing a temporary, flexible configuration optimization rather than a permanent hardware downgrade; demand for HBM high-bandwidth memory related to GPU core computing power remains unaffected.


Previously, Morgan Stanley estimated that the price of NVIDIA’s new Vera Rubin (VR200) rack would be approximately $7.8 million, with memory-related material costs increasing by over 435% compared to the previous generation. Industry insiders also note that the AI hardware bottleneck is gradually shifting from compute power to memory and interconnects.


This year, the global storage sector has continued to strengthen, with the market gradually shifting its valuation logic for the storage industry from that of a traditional cyclical stock to that of an "AI infrastructure asset." However, as valuations rise rapidly and trading becomes more crowded, any news related to changes in storage demand could trigger market volatility.


Despite short-term corrections, multiple companies across the industry chain remain optimistic about sustained industry momentum. Biwin Storage noted that AI computing demand will keep the supply-demand gap difficult to alleviate in the short term, leaving room for further price increases; some manufacturers expect the supply shortage to persist until 2027. According to CFM Flash Market data, the global DRAM and NAND Flash market reached $137.14 billion in the first quarter of 2026, setting a new all-time quarterly high.

Disclaimer: The information on this page may have been obtained from third parties and does not necessarily reflect the views or opinions of KuCoin. This content is provided for general informational purposes only, without any representation or warranty of any kind, nor shall it be construed as financial or investment advice. KuCoin shall not be liable for any errors or omissions, or for any outcomes resulting from the use of this information. Investments in digital assets can be risky. Please carefully evaluate the risks of a product and your risk tolerance based on your own financial circumstances. For more information, please refer to our Terms of Use and Risk Disclosure.