[News] Huawei Enters Global AI Memory Race with Upcoming AI SSD on August 27

Shortly after releasing UCM (Unified Computing Memory)—an AI inference acceleration toolkit designed to reduce China’s HBM reliance—Huawei is making waves again in the memory sector as it plans to unveil a new AI SSD on August 27, according to Chinese media outlets guancha.cn and National Business Daily.

As the report from National Business Daily notes, this move shows Huawei is stepping into the global AI memory race with its latest AI SSD, joining memory giants like Kioxia and Micron. While Kioxia has laid out a mid-to-long-term plan centered on AI-driven storage innovation, SSD expansion, and capital efficiency to strengthen its NAND market position, Micron recently rolled out three AI SSDs, with the Micron 7600 SSD targeting AI inference and mixed workloads, the report adds.

As per Guancha, unlike consumer SSDs for PCs, Huawei’s product is designed for data centers handling large-scale AI model training and inference, aiming to tackle the so-called “VRAM wall” challenge in AI training and inference.

While traditional HBM faces capacity limitations, Huawei’s innovation provides large-capacity SSDs with enhanced data throughput efficiency, significantly boosting AI accelerator card performance and providing stronger support for AI applications, the report notes.

On the other hand, National Business Daily suggests that Huawei also plans to partner with integrated system vendors, embedding the new AI SSD into all-in-one AI servers.

Huawei’s Memory Breakthrough

Notably, according to Guancha, China’s DRAM and HBM technologies still lag behind foreign rivals—and face strict U.S. export and manufacturing controls—but domestic NAND flash has made significant strides, boosting the competitiveness of both consumer and enterprise SSDs. Huawei, as per the report, is combining the upcoming AI SSD with self-developed technologies like XtremeLink and SpeedFlex PCB.

It is also worth noting that this is not the only memory breakthrough Huawei has been made recently. In mid-August, Huawei launched UCM, AI inference acceleration toolkit centered on KV (Key Value) Cache technology. The system are said to be combining multiple cache optimization algorithms to intelligently manage KV Cache memory data produced during AI processing. According to Securities Times, UCM automatically distributes cached data across HBM, DRAM, and SSD storage based on memory heat patterns.

Read more

• [News] DeepSeek R2 Model Launch Reportedly Delayed Amid Huawei Ascend Chip Hurdles
• [News] Huawei Unveils UCM Algorithm to Cut HBM Reliance, Reportedly Goes Open-Source in September

(Photo credit: Huawei)