HPE Vultr AI Deal
Hewlett Packard Enterprise secures a $1.2 billion order for AMD AI rack systems from Vultr.
Anonymous
Key Points
1
HPE’s announced rack configuration includes AMD Pensando Vulcano AI network cards and ROCm, plus six HPE Juniper QFX5252 Ethernet switch trays. HPE says each rack package also includes liquid cooling and services for large AI clusters.
2
HPE CEO Antonio Neri described the broader AI buildout as “the largest infrastructure buildout in history,” saying HPE, AMD and Vultr are building an “open foundation” for the next phase of AI.
3
An analyst at Bank of America Securities said the Helios opportunity could eventually grow from several hundred million dollars in switch revenue to several billion dollars from full racks; he maintained a Buy rating and an $88 price target.
4
HPE also lifted its fiscal 2027 Networking segment revenue growth outlook to the high-teens-to-low-20% range and expects segment operating margins in the mid-to-high 20% range. It raised its target for Juniper cost synergies to $800 million in annual run-rate savings by the end of fiscal 2028.
5
HPE said it aims to outgrow market rates and gain share over the next three years in data-center networking, routing, and campus-and-branch businesses; its overall Networking segment is projected to grow at a high-teens CAGR from fiscal 2026 through fiscal 2029.
Hewlett Packard Enterprise landed its first order for the AMD Helios AI rack system — a $1.2 billion deal with cloud provider Vultr Fierce Network. The systems will deploy across Vultr's U.S. data centers to support AI model training and inference. Each rack bundles AMD computing power with HPE networking, liquid cooling, and software into an integrated package.
The deal marks a milestone in HPE's push to sell open, vendor-neutral AI infrastructure. CEO Antonio Neri called the broader buildout Pulse2 "the largest infrastructure buildout in history," with HPE, AMD, and Vultr creating an open foundation for the next phase of AI. HPE simultaneously raised its networking revenue growth projections, expecting data-center networking to expand 50% annually through fiscal 2029.
Vultr Orders Blade-Scale AI Systems for U.S. Cloud
Each Helios rack includes AMD Pensando Vulcano AI network cards, six HPE Juniper QFX5252 Ethernet switch trays, ROCm software, and liquid cooling TechZine. The systems are built for large AI clusters running model training and inference workloads. Vultr joins other cloud and infrastructure players racing to build out AI capacity amid surging demand for chips and networking gear.
HPE Raises Networking Revenue Growth to High Teens
HPE lifted its fiscal 2027 Networking segment revenue growth outlook to high-teens-to-low-20% range, up from prior guidance StreamlineFeed. The company expects segment operating margins in the mid-to-high 20% range. HPE also boosted its Juniper cost-synergy target to $800 million in annual run-rate savings by end of fiscal 2028, signaling confidence in the networking pivot.
Analyst: Helios Could Grow to Billions in Rack Revenue
A Bank of America Securities analyst said the Helios opportunity could balloon from several hundred million dollars in switch revenue to several billion dollars as full racks scale up DigiTimes. The analyst maintained a Buy rating and $88 price target on HPE stock, betting the company will grab share in the booming data-center infrastructure market. HPE aims to outgrow competitors over the next three years in networking, routing, and campus-branch businesses.
Open Infrastructure Play Deepens HPE-Vultr Partnership
The deal builds on an existing relationship between HPE and Vultr, with the cloud provider now buying complete rack systems rather than just components Fierce Network. HPE is betting that integrating AMD chips with its own networking and cooling will win over cloud builders looking for alternatives to locked-down, single-vendor stacks. The $1.2 billion order signals strong demand for open, composable AI infrastructure in 2025.
Sentiment
Mostly PositiveNEGATIVENEUTRALPOSITIVE
Publishers
18
Articles
52
Reach
70
A
AnonymousAuthor
