Hewlett Packard Enterprise has received a $1.2 billion order from cloud provider Vultr for AMD Helios AI rack systems, its first order for the platform, the company disclosed in a filing with the Securities and Exchange Commission Wednesday. HPE announced the deal alongside its Networking Investor Day, where it raised its targets for the networking business built around its Juniper acquisition.

What Happened

Vultr plans to install the systems in its U.S. data centers for AI training and inference, according to the announcement as reported by Investing.com. Each Helios rack combines 72 AMD Instinct MI455X GPUs with AMD EPYC "Venice" processors, AMD Pensando Vulcano AI network cards and AMD's ROCm software, along with six HPE Juniper Networking QFX5252 Ethernet switch trays that provide the scale-up fabric, including support for UALink over Ethernet. HPE Services will handle deployment and bring its liquid-cooling expertise to the installation. The announcement did not include a delivery schedule.

Vultr chief executive J.J. Kardwell said demand for high-performance AI infrastructure keeps running ahead of available capacity. HPE chief executive Antonio Neri described AI as driving the largest infrastructure buildout in history.

RelatedIran Security Chief Threatens Cables, Data Centers and AI Sites if Fighting Resumes

At the investor event, HPE guided networking segment revenue growth for fiscal 2027 to the high teens to low 20s percent, with operating margin in the mid-to-high 20s. Through fiscal 2029 it targets a revenue compound annual growth rate in the high teens for the segment. By business line, it expects data center networking to grow at a rate in the low-to-high 50s percent, routing in the low-to-high 20s, and campus and branch and security each in the high single digits. HPE also raised its Juniper cost-synergy target to $800 million in annual run-rate savings by the end of fiscal 2028, up from $600 million.

HPE shares rose about 4% on the day, according to Investing.com.

Why It Matters

The order is a significant reference customer for AMD's Helios, the company's rack-scale system designed to compete with Nvidia's 72-GPU NVL72 racks. Neoclouds such as Vultr, which rent GPU capacity to AI developers, have been among the first buyers willing to build large clusters on non-Nvidia hardware, and a $1.2 billion commitment signals that at least one is prepared to do so at scale.

For HPE, the deal ties its server business to its newer networking franchise. Every Helios rack in the order ships with Juniper switch trays, which is the kind of attached networking revenue HPE is counting on to reach its data center networking targets. Rami Rahim, who leads HPE's networking business and was Juniper's chief executive before the merger, said AI is making the network more strategic and creating substantial new demand.

Context

HPE completed its roughly $14 billion purchase of Juniper Networks in 2025 after a U.S. antitrust review, and has since reorganized to make networking a central growth segment. The data center networking forecast is the most aggressive in its new outlook, reflecting the shift in AI clusters toward Ethernet-based fabrics for connecting GPUs, both within racks and across them.

What to Watch

Key details still to come are the delivery timing for Vultr's racks and whether other cloud providers or enterprises place Helios orders through HPE. The networking targets will be measured against HPE's quarterly results over the next several reporting periods, starting with how much AI-related switching shows up in data center networking sales.