Supermicro Among First to Unveil NVIDIA BlueField-4 STX Storage Server to Improve AI Inference Performance

NodeGuardian · 2026-03-19T06:40:35+00:00

Supermicro introduced a context memory (CMX) storage server using NVIDIA's STX architecture. It enhances AI inference performance by managing KV cache, reducing recomputation needs, and accelerating results with NVIDIA Dynamo for orchestration.

NodeGuardian

2026-03-19 06:40:35

Abstract generation in progress

Supermicro has unveiled one of the industry’s first context memory (CMX) storage servers, built on NVIDIA’s new STX reference architecture. This solution combines NVIDIA Vera CPU and ConnectX-9 SuperNIC to accelerate AI inference performance, particularly for long-lived AI queries and multi-stage agentic workloads, by efficiently managing Key Value (KV) cache. The CMX server aims to reduce recomputation power requirements and speed up results by leveraging NVIDIA Dynamo for inference orchestration.

This page may contain third-party content, which is provided for information purposes only (not representations/warranties) and should not be considered as an endorsement of its views by Gate, nor as financial or professional advice. See Disclaimer for details.