The HPE Vultr order, disclosed on 30 September 2026, is worth $1.2 billion and is Hewlett Packard Enterprise’s first for the AMD Helios AI Rack system. HPE raised its fiscal 2027 networking growth outlook the same day. It has not said when the equipment ships or when the revenue lands.
What HPE Actually Sold
The order covers complete AI rack systems for Vultr’s data centres in the United States, not switches alone. Each rack combines 72 AMD Instinct MI455X graphics processors with AMD EPYC “Venice” central processors, AMD Pensando Vulcano networking chips and AMD’s ROCm software layer.
HPE’s own contribution is the fabric that ties the processors together: six HPE Juniper Networking QFX5252 scale-up Ethernet switch trays per rack, connecting all 72 graphics processors. A scale-up fabric is the high-speed interconnect inside a single rack, as distinct from the scale-out network that links racks to one another, and it is the part of the AI build-out HPE has been arguing it can win on Ethernet rather than proprietary links.
The equipment is destined for what HPE described as model training and inferencing workloads for service providers and enterprises. Juniper Networks, whose switches carry the traffic, became part of HPE through acquisition, and Vultr had worked with Juniper for close to three years before this order.
What an Order Is, and What HPE Has Not Disclosed
An order is a commitment to buy, not recognised revenue, and the gap between the two is where this announcement is thinnest. HPE has not disclosed how the $1.2 billion divides among networking hardware, compute, software and services, nor published a delivery schedule, nor said in which quarters the revenue will be recognised. Reporting by Fortune, Quartz and Dow Jones Newswires on 30 September and 1 October 2026 each noted the same absences.
Marie Myers, HPE’s executive vice president and chief financial officer, framed the task in those terms, saying the company expects “networks for AI to be a meaningful growth engine” and that “demand is continuing to outpace supply”. Supply constraints cut both ways: they support pricing, and they govern how quickly an order becomes a shipment.
As of 2 October 2026, the order stands as announced, with no delivery milestones in the public record. HPE’s financial year ends on 31 October, so fiscal 2027 runs from November 2026 to October 2027 — the window the raised outlook covers.
The Guidance HPE Raised
The forecast changes were larger than the order itself, and they were announced at the company’s securities analyst meeting rather than with quarterly results.
| Measure | Previous | New | Period |
|---|---|---|---|
| Networking segment revenue growth | 14% to 17% | High teens to low-20s percent | Fiscal 2027 |
| Networking revenue growth rate | 5% to 7% a year | High-teens percent compound annual rate | Fiscal 2026 to fiscal 2029 |
| Total company revenue growth | Not previously at this level | 13% to 17% | Fiscal 2027 |
| Non-GAAP diluted EPS growth | Not previously at this level | 16% to 20% | Fiscal 2027 |
| Juniper cost synergies | At least $600 million | $800 million annual run rate | By end of fiscal 2028 |
The networking operating margin target was held at a mid-to-high 20s percent range for fiscal 2027 through fiscal 2029. The compound annual growth rate revision is the steepest of these: the earlier framework assumed 5% to 7% growth a year, and the new one assumes roughly three times that.
Where the Networking Business Stood Before This
The raised outlook builds on a quarter that was already a record. In its fiscal 2026 third-quarter results, published on 2 September 2026 for the quarter ended 31 July 2026, HPE reported networking revenue of $2,893 million, up 74.9% on the prior-year period, at a 22.0% operating margin.
Within that, data centre networking revenue was $382 million, up 112.2% year on year — the smaller but faster-growing part of the segment, and the one the Vultr order feeds. Group revenue for the quarter was a record $12.2 billion, up 34%.
HPE has also said it expects cumulative fiscal 2026 orders for networks for AI to exceed $3 billion. Against that figure, a single $1.2 billion order is a material share of a year’s AI networking demand concentrated in one customer.
Who Vultr Is and How It Is Paying
Vultr is a private cloud infrastructure company, not a hyperscaler, and that matters to how the order gets funded. Founded in 2014 and based in West Palm Beach, Florida, it says it serves more than 1.5 million customers across 185 countries from 32 cloud locations. HPE described it as the world’s largest privately held cloud infrastructure company, which is the vendor’s characterisation rather than an audited ranking.
Its most recent disclosed valuation was $3.5 billion. It raised $333 million in December 2024 in a round led by LuminArx Capital Management in which AMD Ventures participated, and closed $329 million of conventional debt in June 2025 backed by Goldman Sachs and Citi. AMD Ventures is therefore an investor in the company now buying a rack built almost entirely from AMD silicon — a fact worth recording, though nothing in the announcement suggests the financing and the purchase are linked.
Vultr has reportedly been seeking at least $1 billion in fresh capital, working with Goldman Sachs, according to industry reporting that HPE’s announcement does not address. Whether that raise completes is one of the conditions behind a $1.2 billion equipment commitment, and it has not been confirmed by the company.
J.J. Kardwell, Vultr’s chief executive, said demand for high-performance AI infrastructure “continues to outpace available capacity” — the same supply argument HPE’s own executives make.
What to Watch Next
Three things would turn this announcement into results. The first is HPE’s fiscal 2026 fourth-quarter report, which will show whether the order appears in backlog and how management describes its timing. The second is any disclosure of the split between hardware, software and services, which determines the margin the order carries. The third is Vultr’s funding.
Antonio Neri, HPE’s president and chief executive, said that “AI is driving the largest infrastructure buildout in history and together with AMD and Vultr, we are building an open foundation to accelerate the next phase of AI.” Rami Rahim, who runs HPE’s networking business, framed it as AI “reshaping the technology stack, making the network more strategic”. Neither statement addresses delivery dates.
HPE shares rose about 3.9% on 1 October 2026, according to market roundups published that day. That is a single session’s move and not a measure of the order’s eventual value. For a comparable AI-infrastructure announcement with its own undisclosed conditions, see our report on Nvidia’s OpenShell and Sentry.
Frequently Asked Questions
How Big Is the HPE Vultr Order?
It is $1.2 billion, disclosed on 30 September 2026. It is HPE’s first order for the AMD Helios AI Rack by HPE system, and covers deployments in Vultr’s United States data centres.
When Will HPE Recognise the Revenue?
HPE has not said. No delivery schedule or revenue-recognition timing has been published, and an order is a purchase commitment rather than recognised revenue.
What Is in an AMD Helios Rack?
Each rack holds 72 AMD Instinct MI455X graphics processors with AMD EPYC “Venice” processors, AMD Pensando Vulcano networking chips and AMD ROCm software, connected by six HPE Juniper Networking QFX5252 scale-up Ethernet switch trays.
What Guidance Did HPE Change?
It raised its fiscal 2027 networking segment revenue growth outlook to a high-teens-to-low-20s percent range from 14% to 17%, lifted its fiscal 2026 to 2029 networking growth target to a high-teens percent compound annual rate from 5% to 7% a year, and raised its Juniper cost synergy target to an $800 million annual run rate by the end of fiscal 2028.
Is Vultr a Public Company?
No. Vultr is privately held, was last valued at $3.5 billion, and has said it is considering a stock market listing at some point in the coming years.
