Rackspace cuts 750 jobs and reinvents itself as a regulated-AI provider: the human cost of a necessary bet

By Momentum IA · June 28, 2026.

Rackspace Technology has just confirmed one of the most significant restructurings in its recent history: the elimination of approximately 750 jobs, 15% of its global workforce of about 5,000 people. The decision, approved by the executive committee on June 10 and communicated to those affected that same day, will generate between 75 and 85 million dollars in annual savings once completed. The company will incur costs of between 14 and 19 million dollars in severance during 2026. Most of the departures have already begun, and the process will extend over the next six months depending on each country's labor legislation.

Behind the numbers lies a narrative of strategic reconversion that the new CEO, Gajen Kandiah —in the post since September 2025, with prior experience at Hitachi Digital and 16 years at Cognizant— has placed at the center of the corporate story: the transition from the legacy public cloud business toward what the company calls 'governed enterprise AI', that is, secure and auditable AI infrastructure for clients in regulated industries such as healthcare, financial services and public administration. To support that shift, Rackspace has recently sealed alliances with AMD in AI infrastructure and with Palantir to deploy AI platforms in regulated environments, two partners that fit well with the institutional client profile it aspires to.

**Context matters: a company that has spent years searching for its place**

As sector context, Rackspace does not arrive at this moment from a position of strength. The company was acquired by Apollo Global Management in 2016, returned to public markets in 2020 and has since contended with pressure from AWS, Azure and Google Cloud, giants that have made the value proposition of generalist managed cloud enormously difficult. Its recent financial record reflects continued losses, although the company reported in early 2026 a narrowing of those losses, suggesting that the earlier cost adjustment is beginning to take effect.

In that context, the pivot toward regulated enterprise AI is not just a fad: it is, possibly, the only viable path of real differentiation left to a mid-sized company in a cloud infrastructure market dominated by three or four players with economies of scale impossible to replicate. Regulated sectors —banking, healthcare, defense, government— have specific needs for data sovereignty, model traceability and regulatory compliance that large cloud providers address in a more generic way. There may be a profitable and defensible niche there.

**What is not known, and is relevant**

The company has not published a geographic breakdown of the layoffs, which has generated noise on social media, where some users claim the cuts disproportionately affect U.S. workers while jobs are preserved in other regions —a claim that, as the article itself notes, remains unverified. Rackspace has approximately 1,800 employees in North America, 800 in EMEA and around 2,400 in Asia. No WARN notice has been filed in Texas, where it is headquartered, which either indicates that layoffs in that state do not reach the legal threshold, or that those affected are mainly workers outside the U.S. The opacity on this point is, in itself, a data point that deserves attention.

The suggestion is also circulating that the savings could be partially allocated to executive compensation. Rackspace has neither confirmed nor denied this with concrete data. Presenting it as a proven fact would be unfair; ignoring it as a legitimate concern would be too. In a cycle in which restructurings are routinely presented as 'investment in the future' while top-level bonuses remain intact, public skepticism has a rational basis.

**Our reading: the pattern that repeats, and what it reveals**

Rackspace is one more on an ever-growing list: tech companies that in 2025-2026 are cutting human capital to redirect resources toward AI. The difference here is not the move itself —we have already seen it at dozens of companies— but the proportional magnitude (15% is a high figure) and the starting point (a company that was already weakened, not one cutting from abundance).

That matters because the transition Rackspace is attempting is not guaranteed. Partnering with AMD and Palantir is a smart move on paper, but building credibility in regulated sectors takes time, involves long sales cycles and demands institutional trust that is not bought with a press announcement. The real risk is that the company ends up in no man's land: too small to compete with the hyperscalers on volume, and not yet specialized enough for a hospital or a financial regulator to choose it over providers with decades of compliance track record.

For the 750 affected workers, this strategic analysis is, at best, cold comfort. The promise that the savings will be reinvested in 'AI engineering teams' is real for those who have the right profile and manage to relocate inside or outside the company; for those who do not, the transition has a very concrete personal cost. There is no honest way to narrate this kind of restructuring without acknowledging that harm, however correct the strategic direction may be.

In the long term, demand for reliable AI infrastructure in regulated environments will grow steadily. The digitalization of healthcare, the automation of processes in banking and the regulatory requirement for auditable models in Europe and the U.S. are structural trends that will not reverse. If Rackspace manages to execute its pivot well —and that is a genuine 'if', not rhetorical— it will have found a viable place in the ecosystem. But the road between the bet and its validation is paved with uncertainty, and this week it is paid for, above all, by those who lose their jobs.