Vultr picks HPE and NVIDIA for next-generation AI cloud data centers

Vultr, described in the announcement as the largest privately held hyperscaler, said it has selected HPE and NVIDIA for large-scale AI data-center deployments. The decision points to continued demand from cloud providers for integrated infrastructure built to run high-performance AI training and inference workloads.

HPE made the announcement at HPE Discover Las Vegas 2026, where it said the Vultr deployment is aimed at next-generation AI infrastructure for cloud-scale data centers. Vultr did not provide additional public contract specifics in the brief announcement text, and HPE did not disclose cost, timeline, or the size of the initial rollout.

The companies framed the move as a path to scale AI infrastructure. NVIDIA and HPE did not spell out in the disclosed material whether Vultr will deploy a particular NVIDIA GPU generation or how systems will be configured across clusters, networks, and storage. As presented, the announcement mainly establishes that Vultr will pair HPE’s data-center platforms with NVIDIA technology for AI operations.

For NVIDIA, the announcement reinforces its role supplying accelerators and the associated software ecosystem that many data-center operators rely on to deliver AI services at scale. For HPE, the selection underscores how enterprise and cloud infrastructure providers are positioning server, storage, and systems-management offerings around AI use cases rather than general-purpose compute alone.

Vultr’s characterization as a cloud-scale hyperscaler also matters for the competitive backdrop. Hyperscalers and fast-growing cloud providers typically seek standardized hardware stacks that can be ordered and expanded repeatedly, while minimizing integration work for new AI capacity. In that context, a large-scale selection can announcement confidence in the ability of HPE and NVIDIA to support incremental growth over multiple refresh cycles.

Still, major details remain unreported. The announcement text available publicly does not include the specific hardware models, the number of systems or GPUs, target delivery dates, or performance benchmarks. It also does not indicate whether Vultr plans to use a single standardized reference design across regions or to vary configurations by workload type.

What to watch next is whether the deployment will be quantified over time, either through further HPE disclosures, Vultr communications, or NVIDIA customer references. Buyers in AI infrastructure also look for indicates on how such deployments will be operationalized, including management tooling, software support, and how quickly capacity can be commissioned for new AI projects.

For now, the announcement is a clear indicator of continued alignment among a major cloud operator, a large infrastructure OEM, and a leading AI chip supplier. Even without a disclosed scale or configuration, a large-scale AI data-center selection suggests that demand for GPU-accelerated infrastructure remains active in cloud environments.