The Pulse
HPE Lands $1.2 Billion Vultr Order for AMD Helios AI Racks
HPE announced a $1.2 billion order from Vultr for AMD Helios AI Rack systems that the cloud provider plans to deploy at U.S. data centers. The release describes the racks’ compute and networking components but gives no rack count or deploym
AI.info Team · Written with AI ·
HPE’s first order for its AMD Helios AI rack system is worth $1.2 billion. The buyer, cloud infrastructure provider Vultr, plans to deploy the systems at its U.S. data centers, where HPE and AMD will support model training and inference workloads. HPE announced the order on September 30, 2026, but did not disclose how many racks it covers or when deliveries will begin.
Each Helios rack links 72 AMD GPUs
The racks combine 72 AMD Instinct MI455X GPUs with AMD EPYC “Venice” processors, Pensando Vulcano AI network cards and AMD ROCm software. HPE says the design supports open, interoperable rack-scale fabrics, including UALink over Ethernet. The configuration ties the accelerators to networking and software intended for large-scale AI training and inference.
HPE supplies six Juniper Networking QFX5252 scale-up Ethernet switch trays in each rack. The switches connect all 72 GPUs using standards-based Ethernet, according to the company. HPE also says its services team will help with deployment and liquid cooling for large clusters. The release says the rack’s unified design is optimized for power, cooling and serviceability.
HPE describes Helios as a high-density system designed for trillion-parameter model training and high-volume inference. The announcement does not give performance benchmarks for Vultr’s planned deployment, nor does it specify the size of the installation. The $1.2 billion figure is the value of the order, not a disclosed rack count or per-system price.
Vultr cites capacity pressure
“Demand for high-performance AI infrastructure continues to outpace available capacity, and our work with AMD and HPE enables us to bring Helios rack-scale compute and scale-up networking online for customers faster. Together, we're delivering the memory capacity, interconnect performance, and energy efficiency customers need to run the largest and most demanding AI training and inference workloads,” said J.J. Kardwell, CEO of Vultr.
HPE CEO Antonio Neri framed the deal as a complete infrastructure package, combining rack-scale compute, networking and direct liquid cooling. The announcement describes the Helios system as part of HPE’s AI Data Center Solutions portfolio, an infrastructure stack developed with AMD and designed using open standards.
“AMD Helios by HPE demonstrates HPE’s unique ability to engineer the complete AI architecture with rack-scale compute, scale-up networking, and direct liquid cooling, all delivered by our global services.”
The statement was made by Antonio Neri, president and CEO of HPE. The release also says HPE’s services team will provide expertise in end-to-end deployment and liquid cooling for large AI clusters.
A first Helios order, with deployment details pending
The order extends a relationship between Vultr and Juniper Networks that HPE says has lasted nearly three years; Juniper is now part of HPE. HPE says the Helios order brings its Juniper Networking technologies together with HPE compute, liquid cooling and rack-scale engineering in an open, Ethernet-based scale-up AI system.
The release leaves several practical details open: the number of racks, delivery timing and how Vultr will make the added capacity available to customers. It also does not break down the $1.2 billion order among compute, networking, cooling and services. HPE identifies the rack components and total order value, but gives no rollout schedule or deployment-specific performance results.