The demand for NAND flash memory from NVIDIA's AI servers is entering a phase of explosive growth, and Samsung Electronics is leveraging its production capacity advantage to swiftly turn this trend into a strategic opportunity.
Samsung Electronics has reportedly initiated mass production of its 10th-generation V-NAND (V10) and commenced formal supply to NVIDIA. Concurrently, the company has allocated approximately 60% of its total V-NAND production capacity to its 9th-generation (V9) product lines. This move is aimed at meeting the demand for NVIDIA's upcoming new storage device, "Context Memory Storage" (CMX). Industry projections indicate that the NAND capacity required for CMX will surge from 35 million TB this year to over 100 million TB next year.
This surge in demand is exerting a profound impact on the supply dynamics of the entire NAND market. NVIDIA's large-scale locking of NAND supply implies a significant narrowing of available storage chip supplies for other enterprises and consumers, putting upward pressure on market prices.
Simultaneously, Samsung Electronics Chairman Lee Jae-yong is expected to meet with NVIDIA CEO Jensen Huang in San Francisco in the near future. The two leaders are anticipated to engage in in-depth discussions regarding the supply of HBM and next-generation NAND products, as well as potential cooperation on constructing data centers within South Korea.
CMX Fuels a Leap in NAND Demand
The development of NVIDIA's CMX stems from the dramatic expansion of "Key-Value Cache" (KV Cache) data volumes during AI inference processes. Previously, GPU servers could handle KV cache data internally. However, as data scales continue to grow, independent external storage devices equipped with thousands of terabytes of solid-state drives (SSDs) have become a necessary configuration for AI servers.
Reports indicate that a single NVIDIA CMX device is equipped with 576 SSDs, offering a total storage capacity of 9,600 TB. Industry estimates suggest the NAND capacity required for CMX will be around 35 million TB this year, surpassing 100 million TB next year. An independent semiconductor analyst noted on social media that a 100 million TB scale roughly equates to adding a demand source the size of Apple to the NAND market, an impact on the supply side that cannot be underestimated.
Samsung currently possesses a monthly production capacity exceeding 100,000 V-NAND wafers. It is reported that NVIDIA has requested an expansion of NAND capacity from Samsung ahead of CMX shipments, a request to which Samsung is actively responding. Samsung's total NAND production this year is estimated to reach approximately 250 million TB. The company plans to leverage this globally leading production scale to solidify its position as a core supplier for NVIDIA's CMX.
Three Generations of V-NAND in Simultaneous Production
In terms of product line strategy, Samsung is concurrently advancing the production processes of three generations of V-NAND products.
For V9, Samsung has increased its production share to about 60% of total V-NAND capacity, with yield rates stabilizing above 80%, marking its entry into a mature mass production stage. A year ago, V8 still accounted for a significant portion of V-NAND output; its share has now dropped below 40%, demonstrating Samsung's rapid response capability to AI server demand. Furthermore, over half of the NAND-dedicated cleanroom space at Samsung's Pyeongtaek P4 fab remains unutilized, providing ample room for future capacity expansion.
For V10, Samsung formally commenced mass production in the first half of this year. It currently accounts for less than 5% of total V-NAND output, indicating it is in the ramp-up phase. V10 utilizes an approximately 400-layer stacking architecture, offering a storage density increase of over 50% compared to V9's 286 layers. This allows for the installation of higher-capacity SSDs within the same CMX module. Notably, V10 will see the first large-scale commercial use of molybdenum material to replace traditional tungsten wiring, potentially reducing wiring thickness by 30% to 40%, a material Samsung had trialed on a small scale in V9.
For V11, Samsung initiated trial production earlier this year, targeting a stacking layer count between 400 and 500 layers, an increase of about 100 layers over V10. Samsung plans to double the trial production scale in the second half of the year to accumulate sufficient yield data and accelerate the establishment of mass production systems.
Supply Concentration Impacts Consumer Market
The deepening NAND alliance between Samsung and NVIDIA, while boosting Samsung's earnings outlook, is also intensifying supply bifurcation across the memory industry.
Reports suggest Samsung's projected profit for 2026 may exceed the cumulative total of its profits over the past 40 years, largely driven by the structural explosion in AI server storage demand. However, the flip side of this windfall is that as Samsung locks a large proportion of its capacity for major tech clients like NVIDIA, the supply of NAND for other enterprises and consumers will be significantly squeezed, raising the risk of upward price pressure.
Previously, the AI boom first triggered DRAM shortages and drove significant price increases. The NAND market is now facing a similar supply-tightening dynamic. As NVIDIA's CMX shipment volumes continue to expand, this pressure is expected to intensify further.
In the long term, the accelerated development of V11 NAND may offer some relief for the consumer market. The 500-layer stacking technology theoretically enables the mass production of high-capacity consumer SSDs at lower costs, but this is contingent on Samsung being willing to allocate capacity for it. Given the current direction of capacity allocation, this prospect remains uncertain.
Upcoming Leadership Meeting to Expand Collaboration
The collaboration between Samsung and NVIDIA has expanded from DRAM and HBM to foundry services, and now incorporates NAND flash memory into the alliance framework, continuously extending both its depth and breadth.
As mentioned, Chairman Lee Jae-yong is scheduled to meet with CEO Jensen Huang in San Francisco soon. The agenda for this meeting is expected to cover supply arrangements for HBM and next-generation NAND products, as well as cooperative plans for building data centers in locations such as Gwangju, South Jeolla Province, South Korea.
Industry observers note that the alliance between Samsung and NVIDIA spans the entire spectrum of memory and foundry businesses, and discussions between the two leaders will further deepen this collaboration. With NVIDIA's CMX shipments imminent, the strategic significance of this meeting is set to become even more pronounced.
Comments