Rackspace Technology and the business of staying on call

In 2018, a hotel company acquired something no guest had asked for: a deadline. After its spin-off, Wyndham Hotels & Resorts had four months to leave a shared corporate data center. Its applications had accumulated across decades. Previous migration attempts had failed. The tidy story about moving everything to the cloud ran into the untidy business of keeping hotels working.

A hotel company runs out of time

Wyndham chose Rackspace and used colocation as a bridge. It could relocate hardware while deciding which workloads belonged elsewhere. The company reported completing the exit on time and cutting first-year run costs by 40-45%, according to Rackspace’s case study.

The revealing detail is what changed the plan: older applications needed more work than the deadline allowed. A staged move made room for that reality. This is a useful way to understand Rackspace. Its customer often owns a collection of dependencies, each with a different clock.

“Can I check in?”

Scott Strickland, Wyndham CIO, on what the guest actually cares about

The phone was part of the product

Rackspace dates its managed-hosting beginnings to 1998. Richard Yoo, Dirk Elmendorf and Pat Condon were the original technical founders; Graham Weston supplied early capital and became CEO. Morris Miller was also a co-founder and early investor. The opportunity was straightforward: businesses wanted websites without becoming experts in the machinery underneath them.

The company made service its calling card, branding the approach Fanatical Support. Today it uses Fanatical Experience. Employees are called Rackers, a small linguistic clue to a culture built around the people doing the operating. The mission links technical learning to customer outcomes. Those words become commercially meaningful only when an engineer understands the application well enough to fix the right problem.

There was also an ambitious contribution to the tools themselves. In 2010, Rackspace and NASA helped launch OpenStack. Rackspace brought object-storage technology; NASA’s work contributed compute foundations. An infrastructure operator helped create software that other operators could use. That history matters because Rackspace’s role has never been limited to answering support tickets.

Buy the team, keep the context

Today the menu includes public-cloud operations across AWS, Microsoft Azure and Google Cloud, private and hybrid infrastructure, application modernization, data work and cyber resilience. A business can engage Rackspace to plan a migration, build the environment and keep improving it after launch. For a regulated organization, that work can include meeting requirements for where data resides and who may access it.

Elastic Engineering makes the labor model unusually visible. Customers buy fractional access to a continuing pod of architects and engineers through monthly hours-based tiers. The pod works alongside the internal team in sprints. Keeping the same people involved is the point: less time explaining yesterday’s decisions, more time acting on them.

This is an enterprise services business. Revenue comes from managed operations, infrastructure consumption and resale, projects and recurring engineering engagements. Buyers should compare the complete operating bill: cloud capacity, service fees, licenses, migration work and the staff they still need. A lower infrastructure price can coexist with a higher total cost.

Consider the buyer who has inherited a database, an aging application and a new analytics project. Each may need a different environment, but somebody must still connect access controls, incident response and spending decisions. That coordination is where an outside operator can help. The customer still needs enough internal knowledge to challenge the proposed architecture and recognize when service is falling short.

Its alternatives include cloud consultancies, infrastructure service firms and an internal team buying directly from a hyperscaler. Rackspace’s case rests on coordinating across environments and staying involved in operations. That is a useful distinction, though breadth alone cannot establish better execution. A simple application supported by a capable internal team may have little need for another services layer.

A promise with a repair bill

A service reputation does not prevent a service failure. In December 2022, ransomware disrupted Rackspace’s Hosted Exchange email business. The company retired that platform and moved many customers to Microsoft 365. Its filings recorded incident-related expenses of $5.9 million in 2022 and $5.2 million in 2023, before insurance recoveries. Those figures describe company expenses, not the full consequences for customers.

The financial picture also resists a simple turnaround story. Rackspace reported approximately $2.686 billion in 2025 revenue, down 2%, and a $226 million net loss. In the second quarter of 2026, revenue rose 0.6% to $670.1 million, while gross margin fell from 19.4% to 17.1%. More sales did not translate into a stronger gross margin.

The next thing somebody has to run

Under Gajen Kandiah, CEO since September 2025, Rackspace is emphasizing governed enterprise AI. Foundry for AI by Rackspace, launched in 2023, organizes work into readiness, incubation and industrialization. The distinction is practical: a working demonstration still needs data controls, integration, monitoring and an owner when its behavior changes.

The infrastructure announcements grew more concrete in 2026: an AMD agreement for an initial 30 megawatts of compute phased from late 2026 through 2028; a VMware Cloud Foundation 9.1-based platform; and NVIDIA Cloud Partner Program membership with Blackwell infrastructure announced. Planned capacity should be judged as planned capacity. Customer adoption and dependable operation are separate tests.

For buyers, the test is concrete. Ask who handles a failed deployment, which team owns a security alert and how quickly data can be restored. Put the answers beside the commercial proposal. An impressive partner list establishes access to technology; it does not tell you how your particular incident will be handled. That judgment requires examining the service you are actually purchasing.

The lesson a reader can borrow is smaller than an AI strategy. Inventory dependencies. Decide what can be replaced, adapted or relocated. Give every system an operational owner. Rehearse restoration before declaring a migration complete. Rackspace’s enduring proposition is that the work continues after the technology arrives. The buyer’s job is to make that continuing work explicit.