Hewlett Packard Enterprise has signed its first commercial deployment of AMD's Helios rack system, a $1.2 billion agreement with cloud infrastructure provider Vultr to build out AI-focused data center capacity. The deal, disclosed alongside an upgraded long-term revenue forecast, signals how quickly the market for AI-ready hardware is moving and how hard established enterprise vendors are competing to capture that demand.
What the Vultr deal actually involves
Helios is AMD's integrated rack-scale system, combining GPUs, networking, and supporting infrastructure designed specifically for large-scale AI workloads. HPE's role is to manufacture and deploy that system inside data centers for customers like Vultr, a company that rents out GPU compute and cloud infrastructure to businesses running AI models. HPE CFO Marie Myers described the engineering involved in blunt terms: the switch unveiled alongside the deal packs 1,700 copper cables into a single unit, a detail that underscores how physically complex these AI data center builds have become. This is not simply a software or chip story. It is a systems-integration challenge, and HPE is positioning itself as the company that can solve it at scale.
Why HPE raised its revenue outlook
Alongside the Vultr announcement, HPE lifted its long-term growth projections for its data center networking segment, now guiding toward percentage growth in the low-to-high 50s through 2029. Myers pointed to two forces driving that revision. First, demand for AI data center buildout continues to accelerate, with the Vultr agreement following a prior deal with Oracle. Second, HPE sees momentum in self-driving, or autonomous, network management, an area where it says few competitors have the technical depth to deliver. ServiceNow, she noted, has set a goal of running a fully autonomous network by 2028, a target that depends on vendors like HPE building the automation layer underneath it.
A multi-vendor strategy, not an exclusive bet
Myers was careful to frame the AMD partnership as additive rather than exclusive. HPE has used chips and components from multiple silicon vendors throughout its history, and she described Helios as "one of many" platforms the company will support going forward. That matters for investors and customers alike: it suggests HPE is hedging against the risk of over-reliance on any single chipmaker in a market where GPU supply, pricing, and architecture choices are shifting fast. For enterprise buyers, it also means more flexibility in how they configure AI infrastructure rather than being locked into one vendor's roadmap.
The broader stakes for AI infrastructure buildout
Deals of this size illustrate how the AI boom has shifted capital spending from chips alone toward full-stack infrastructure - racks, cooling, networking, and automation bundled together. For companies like HPE, that creates a new competitive battleground against rivals building similar integrated systems. For customers such as Vultr, the appeal is speed to market: buying a pre-engineered rack system rather than assembling one from separate components. The risk, as with any fast-scaling infrastructure cycle, is execution. Complex deployments involving thousands of cable connections and tightly integrated hardware leave less room for error, and demand projections in AI infrastructure have proven volatile before. HPE's raised guidance reflects confidence in current order flow, but it also raises the bar for delivery over the next several years.