Rackspace Technology is cutting its 2026 financial guidance while investing $250 million（約400億円） to accelerate its shift into enterprise AI infrastructure for regulated industries. The company is exiting lower-margin cloud resale businesses and redirecting that capacity toward GPU deployments, which it expects to generate higher revenue and margins. Rackspace plans to scale to 30 megawatts of AI compute capacity by the end of 2028, driven by partnerships with Palantir, AMD, and other AI technology providers.
Summaries like this, in your inbox every morning.
Sign up free →What happened
Rackspace Technology announced a $250 million（約400億円） at-the-market equity offering and lowered its full-year 2026 revenue outlook by $150 million（約240億円） and EBITDA outlook by $20 million（約32億円）. The company is exiting lower-margin businesses in public and private cloud to redirect capital toward GPU-based AI infrastructure deployments.
Why it matters
Rackspace is positioning itself as an operator of enterprise AI infrastructure for regulated industries and sovereign markets, backed by partnerships with Palantir, VMware, Uniphore, Rubrik, and AMD. The company expects AI deployments to generate $15 million（約24億円） to $20 million（約32億円） of revenue per megawatt of deployed capacity on average, with EBITDA margins above 50%—significantly higher than the low-margin services being abandoned.
What to watch
Rackspace aims to reach 15 megawatts of cumulative GPU capacity by the end of 2027 and 30 megawatts by the end of 2028. The first AMD-powered deployment is expected to be nearly 2 megawatts and cost approximately $75 million（約120億円） in capital expenditures, with completion targeted for the end of 2026. The company expects roughly two-thirds of GPU demand to come from new customers.
Rackspace is undergoing a strategic shift away from commodity cloud services toward higher-margin, specialized AI infrastructure. The company's pivot reflects a calculation that it can better compete by focusing on deployment, operations, and regulatory expertise rather than infrastructure resale. The Palantir partnership is central to this strategy: since entering that partnership in February, Rackspace employees have earned more than 400 Palantir certifications, and the companies' first joint deployment for a solar tracking manufacturer closed in 41 days and reduced quote cycle times by 94%.
The near-term financial pain—a $150 million（約240億円） revenue cut and $20 million（約32億円） EBITDA reduction in 2026—reflects the cost of exiting low-margin segments (cloud resale, colocation, basic hosting) while investing ahead of demand in GPU capacity and engineering staff. The company's updated guidance implies a 7% revenue decline at the midpoint rather than the 1% decline previously expected. However, management has signaled that EBITDA benefits from AI deployments are expected to arrive in 2027, meaning investors are funding what the company sees as a temporary margin compression before higher-return AI work scales.
The $75 million（約120億円） capital commitment for the first 2-megawatt AMD deployment and the ambition to reach 30 megawatts by 2028 suggest Rackspace is betting heavily on sustained enterprise demand for regulated-environment AI infrastructure. The fact that two-thirds of expected GPU demand is projected to come from new customers (not existing Rackspace clients) indicates the company views AI deployments as a growth lever rather than simply cross-selling to its installed base.
No discussion yet for this article
Get curated AI news from 200+ sources delivered daily to your inbox. Free to use.
Get Started FreeFree · takes 30 seconds · unsubscribe anytime
1 minute a day. The AI essentials.
200+ sources · Email / LINE / Slack
