Nutanix said its Unified Storage platform has secured certification from Nvidia for use in artificial intelligence infrastructure. The designation means the storage system has been tested and validated against Nvidia’s requirements for AI workloads, a step that could make it easier for enterprises to run machine learning and deep learning tasks on Nutanix gear.
What the certification covers
The certification applies to Nutanix’s Unified Storage product, which combines file, object, and block storage in a single platform. Nvidia’s AI infrastructure typically involves high-performance computing clusters, GPU servers, and fast data pipelines. Storage that meets Nvidia’s standards is supposed to deliver the throughput and low latency needed to feed data to GPUs without bottlenecks.
Nutanix is the latest storage vendor to pursue Nvidia’s stamp of approval. Other companies, including Dell and NetApp, have previously announced similar certifications for their storage arrays. The trend reflects the growing demand for infrastructure that can handle the data-intensive nature of AI training and inference.
Companies building AI models often struggle with storage performance. If the storage layer can’t keep up, GPU clusters sit idle. A certification from Nvidia gives IT teams a benchmark to trust that a storage system can handle the workload. For Nutanix, it opens the door to more AI-focused deals, especially among customers already running its hyperconverged infrastructure.
Nutanix has been pushing into the AI market over the past year. The company introduced a set of AI-related features for its platform in 2024, including support for Kubernetes-based AI workloads and integration with Nvidia’s AI Enterprise software. The new certification builds on that effort.
No specific deployment numbers or performance benchmarks were provided with the announcement. It remains unclear how many customers have already adopted Nutanix Unified Storage for AI use cases, or how the certified platform compares to competitors in real-world tests.
