Samsung Reportedly Deepens Ties with NVIDIA, Scales Up NAND Flash Supply

According to South Korean media reports, Samsung Electronics is expanding its alliance with Nvidia into NAND flash supply, following partnerships in DRAM, high-bandwidth memory (HBM), and foundry (contract chip manufacturing). It is said that Samsung, leveraging its strong production capacity, is increasing shipments of its flagship ninth-generation V-NAND (V9) to Nvidia, while also beginning mass production and supply of its 10th-generation (V10) products. It is also accelerating development of its 11th‑generation (V11) products.

Soaring AI Storage Demand, NVIDIA CMX Becomes the Core Engine

NVIDIA's next-generation storage device, Context Memory Platform (CMX), set to launch in the second half of this year, which has prompted Samsung to swiftly restructure its production capacity and increase supply of high‑end NAND flash supply. With the rapid development of the AI industry, the volume of key‑value (KV) cache data generated during AI inference is exploding exponentially. Traditional GPU servers are no longer able to handle the massive contextual data storage pressure, making standalone storage devices equipped with high‑capacity SSDs a must‑have configuration for AI data centers.

NVIDIA's CMX platform is equipped with 576 SSDs, offering a total storage capacity of 9,600 TB, directly generating enormous demand for NAND flash.  The industry expects NAND demand for CMX to surge from 35 million TB this year to more than 100 million TB next year.  Samsung is expected to become a key supplier for NVIDIA's CMX platform.

Production Capacity Fully in Place, Samsung's V‑NAND in Volume Supply to NVIDIA

Industry insiders reveal that Samsung Electronics has built V-NAND production capacity of more than 100,000 wafers per month, and is mass-producing its 10th-generation V10 product for supply to Nvidia. To adapt to the new demand landscape for AI storage, Samsung has completed a core restructuring of its internal capacity allocation, allocating approximately 60% of its V‑NAND production capacity to V9, elevating it to flagship status. Just a year ago, V8 accounted for a significant portion of V-NAND output, but its production share has now fallen to below 40%. This rapid shift in capacity focus precisely matches the immediate flash storage needs of global tech leaders like NVIDIA for AI servers.

It is reported that Samsung's V9 flash yield has steadily improved to over 80%, entering a stable mass‑production phase, laying a solid foundation for medium‑ to long‑term supply stability between the two parties. In addition, More than half of the dedicated NAND cleanroom space at its Pyeongtaek Plant 4 (P4) also remains available, and the company plans to expand V9 production lines in line with future demand growth.

Accelerated Iteration of  V-NAND, Technology Improvement Builds Competitive Moat

In the first half of this year, Samsung began V10 production, raising it to less than 5% of total V‑NAND output. V10 has a stacking height of 400 layers, offering more than 50% higher storage density than the 286-layer V9. V10 NAND supports a higher number of stacked layers, ensuring that NVIDIA can assemble higher‑capacity SSDs within the same CMX module.

Samsung has begun trial production of the 500-layer-class V11 earlier this year, which is regarded as a high-stack V-NAND being developed to target at least 400 layers up to 500 layers. The company is increasing pilot runs to secure sufficient yield data and rapidly build a mature mass‑production system, with related output expected to double in the second half of this year. By increasing trial production volumes to secure sufficient yield data, the company aims to rapidly establish a mass-production system and plans to double related volumes in the second half.