Rackspace Secures Major AI Compute
AMD and Rackspace Partner on 30MW AI Compute, Offering Managed Enterprise Cloud Solutions
Key Points
Rackspace’s CEO Gajen Kandiah framed the pitch around an “operating model” and accountability, saying the collaboration “combines the right compute with the right operating model and delivers something the market hasn’t offered before.”
The partnership bundles four explicitly named capabilities into Rackspace’s governed offering: “Enterprise AI Cloud,” “Enterprise Inference Engine,” “Inference as a Service,” and “Bare Metal AMD Instinct.”
Rackspace says the AMD-powered capacity is intended not only for AI but also “general-purpose workloads,” using a combination of Instinct GPUs (including MI355X and MI350P) and EPYC CPUs for both workload types.
Bloomberg reported detailed credit-market impact: Rackspace’s first-lien term loan due in 2028 rose to roughly “93.5 cents to 95.5 cents,” gaining about 5 cents on the dollar, after Rackspace disclosed the late-2026 to 2028 30MW AMD deployment; Bloomberg also noted the rally followed a period when the debt had traded “at deeply distressed levels just two months ago.”
In a more detailed governance framing, Rackspace said enterprises need infrastructure “governed from the ground up,” with “one operator accountable for business outcomes,” rather than “a collection of vendors each owning a piece.”
AMD and Rackspace Technology have signed a definitive agreement to deploy 30 megawatts of AMD-based AI compute across Rackspace's global data centers, with rollout beginning in late 2026 and running through 2028. The deal formalizes an earlier memorandum of understanding and marks one of the largest governed enterprise AI infrastructure commitments tied to AMD's Instinct GPU lineup, according to Yahoo Finance.
Rackspace shares surged on the news. Its first-lien term loan due in 2028 jumped to roughly 93.5 to 95.5 cents on the dollar — a gain of about 5 cents — after trading "at deeply distressed levels just two months ago," Bloomberg reported. AMD shares fell slightly on the day.
What Rackspace Is Actually Selling Enterprises
The deal bundles four specific offerings into what Rackspace calls its "governed Enterprise AI Cloud" architecture. Those four products are: Enterprise AI Cloud, Enterprise Inference Engine, Inference as a Service, and Bare Metal AMD Instinct. The hardware will use AMD Instinct MI355X and MI350P GPUs alongside AMD EPYC CPUs, according to Technology Reseller.
Rackspace says the capacity is not just for AI. It will also handle "general-purpose workloads." The pitch targets tightly regulated industries — especially healthcare — where customers need clear accountability over their data and infrastructure. CEO Gajen Kandiah said the collaboration "combines the right compute with the right operating model and delivers something the market hasn't offered before."
Rackspace Bets on Accountability Over Hyperscale Reach
The core argument from Rackspace is simple: when an AI system fails in a hospital, someone needs to be responsible. Rackspace says enterprises need infrastructure "governed from the ground up" with "one operator accountable for business outcomes" — not "a collection of vendors each owning a piece." That framing takes direct aim at AWS, Azure, and Google Cloud, according to EE News Europe.
Both companies plan joint go-to-market efforts. The goal is to help enterprises move past early-stage AI pilots and into full production deployments using agent-based systems. Bare-metal access to AMD Instinct GPUs is central to that push, since autonomous AI agents need low-latency, dedicated compute — not shared cloud resources.
AMD's Debt-to-Credibility Lifeline for Rackspace
The credit market reaction tells a blunt story. Rackspace has carried heavy debt for years. Its term loan was trading in distressed territory as recently as April 2026. The 30MW commitment — a predictable revenue stream locked in through 2028 — gave debt traders a reason to reprice the risk, Bloomberg reported. That 5-cent gain on the dollar was immediate.
For AMD, the picture is more complicated. Thirty megawatts is meaningful for Rackspace but small compared to the gigawatt-scale builds being done by OpenAI or Microsoft. Some analysts suggest the AMD share dip reflects "sell the news" behavior, or doubt that this deal moves the needle against Nvidia's dominance in large-scale AI infrastructure, according to Yahoo Finance.
A Three-Year Rollout Built for Regulated Industries
Phase 1 deployment starts in late 2026, focusing on the initial Enterprise AI Cloud architecture. The remaining capacity rolls out in phases through 2027 and 2028, with full 30MW activation targeted by the end of that period, according to Technology Reseller. Rackspace's internal data center upgrades — including cooling and power management systems — were redesigned specifically for the high heat output of next-generation AMD Instinct chips.
The deal signals a broader shift in enterprise AI away from public hyperscale clouds and toward private, dedicated infrastructure. If it works, it could serve as a blueprint for other mid-tier providers looking to compete not on raw scale, but on governance, compliance, and accountability — especially for customers in healthcare, finance, and defense.
Sentiment
Mostly PositivePublishers
26
Articles
54
Reach
80
