Huawei Launches OceanStor M900 to Enhance AI Efficiency

Huawei has unveiled its OceanStor M900 Context Memory Storage at the HUAWEI CONNECT 2026 event in Shanghai. This innovation targets hyperscale data centers and accelerates AI inference by offering a shared memory space with PB-scale capacity and TB/s-level performance. This marks a shift towards deeper integration of compute, network, and storage, crucial for advancements in AI infrastructure.

As AI scales from chatbots to agents that perform complex tasks, the demand for efficient memory storage grows. The OceanStor M900 addresses memory constraints with a UnifiedBus network, achieving global multi-tier KV cache with one-hop connections. It significantly expands memory capacity and enhances the computing power of SuperPoDs, a preferred AI infrastructure choice.

The solution provides up to 64 PB capacity, boosts performance with integrated architecture, and reduces access latency drastically. Moreover, its KV-aware adaptive storage technology extends SSD lifespan and lowers token costs, reducing the overall expenses associated with AI infrastructures. This enables broader AI adoption across industries.