SC26: Supermicro Launches KV Cache Solution with IBM and NVIDIA
StorageNewsletter, Thursday, November 12th, 2026
Supermicro, IBM and NVIDIA launch a KV cache storage solution to cut AI inference response times.
The joint solution combines Supermicro Petascale all-flash storage servers on AMD EPYC 9005 processors with IBM Storage Scale software, the NVIDIA Dynamo inference framework and Spectrum-X Ethernet networking.
Storing previously computed KV states in a persistent cache lets the system reuse them instead of recomputing the same context for later requests.
Extending KV cache capacity beyond local GPU memory can improve GPU utilization and inference throughput, with high availability and non-disruptive scaling.
Supermicro says this reduces the time before models begin generating a response, which matters for agentic and long-context AI workflows.