HPE Alletra Storage MP X10000 Release 4: Twice the Scale, Built for Production AI
HPE Community (Around the Storage Block), Wednesday, September 16th, 2026
HPE doubles X10000 cluster scale and adds NVIDIA RDMA and GPUDirect Storage for production AI.
HPE Alletra Storage MP X10000 Release 4 is now generally available, doubling single-cluster scale to 16 nodes and 16 JBOFs and roughly 23 PB of raw capacity with about twice the performance of the prior release.
Native-namespace NFS now sits alongside object storage, making file and object first-class protocols on one platform. NVIDIA integration extends RDMA acceleration from object to file, adds GPUDirect Storage enablement, and readies the platform for NVIDIA AI Enterprise certification; independent KV cache offload testing showed up to 20x faster time to first token.
Release 4 also adds KMIP external key management, TLS 1.3 enhancements, and air-gapped deployment support.