SanDisk has announced that an additional demand of 75-100 exabytes (EB) of NAND will arise by 2027 due to the KV cache in artificial intelligence (AI) inference processes. The KV cache serves as intermediate data that large language models (LLMs) use to store context, which cannot be efficiently handled by GPU's high-bandwidth memory (HBM) alone. SanDisk expects that 35% of the NAND workload in AI data centers by 2030 will come from KV caches. The KV cache directly impacts AI inference costs and response times, indicating a need for changes in data storage methods. SanDisk mentioned during its earnings announcement on January 29, 2026, that the demand for KV caches will double by 2028. Additionally, on July 2, 2026, SanDisk announced the sampling of BiCS10 1Tb TLC 3D NAND, stating that this product meets the requirements of AI data centers. SanDisk and SK Hynix signed a memorandum of understanding on August 6, 2025, to standardize high-bandwidth flash (HBF), which is expected to provide bandwidth similar to HBM. SanDisk aims to have the first samples of HBF in the second half of 2026 and samples of AI inference devices by early 2027.
This content is provided for general informational purposes only and doesn't constitute financial, investment, legal, or tax advice. Any events, rewards, online promotions, or related information mentioned herein should not be considered a recommendation, solicitation, or invitation to purchase, sell, trade, or otherwise deal in any crypto assets. Crypto assets are highly volatile and may result in loss. The availability of WEEX services, products, and related events may vary by region. You are responsible for ensuring that your participation is in accordance with applicable local laws and regulations.





























