Rackspace Technology has had a whirlwind history – the stock’s five-year beta sits just above 3.0. In the past year alone, RXT has hit a low of $0.39 and a high of $8.60. Similarly, the bonds have been a rollercoaster. Rackspace’s senior unsecured notes due December 2028 have spiked up to 84c on the dollar from the mid-teens earlier this year.

Cloud computing has never been so exciting.

Rackspace Technology operates two segments: Private Cloud and Public Cloud.

Private Cloud is the dedicated or managed infrastructure business. Rackspace offers dedicated, single-tenant infrastructure that Rackspace manages end-to-end. Customers in this segment enjoy higher security, reliability, and customizability.

The Public Cloud business helps customers run workloads on someone else’s cloud: almost always a hyperscaler (AWS, Azure, Google Cloud). Rackspace assists with migration, navigation, and deployment of customer applications on the hyperscaler’s cloud platform. Rackspace also resells capacity from these hyperscalers.

In short, Private Cloud is the more bespoke (and expensive) side of the business while Public Cloud holds the customer’s hand and “works alongside” the hyperscalers.

We’ll Start in 2016…

With an ill-fated LBO by Apollo. The PE firm valued the company at $4.3b, funding the transaction in part with $3.2b in debt. At the time, Rackspace posted $810mm in EBITDA, putting PF net leverage at ~4x – high but not unmanageable.

Apollo’s thesis was simple: Rackspace was a recurring revenue business with a healthy 20%+ FCF conversion and benefited from the tailwinds of the overall cloud industry.

Tailwinds? Sure. Did Rackspace ride any of those tailwinds? Not really.

Cloud Computing and Where Rackspace Fits in

When cloud computing was still fairly new, Rackspace operated in an important niche: a middleman that helped IT teams navigate the complexities of the cloud. Sometimes the customer chose Rackspace as the host, sometimes it was a third party. Either way, they needed Rackspace to advise on migration and deployment.

Fast forward a couple of years, and IT teams are more familiar with the technology and don’t need as much hand holding. Meanwhile, hyperscalers like AWS and Azure have built out their own tools to help customers navigate the environment. Suddenly Rackspace’s managed services started leaking customers.

From FY’23 through to FY’25, the company has posted consecutive revenue declines: (5.3%) in FY’23, (7.4%) in FY’24, and (1.9%) in FY’25. Simultaneously, margins have also taken a hit. Adjusted EBITDA margins slipped from 18.7% in FY’22 to 10.3% in FY’25. Specifically, the Private Cloud segment has seen its operating margin drop from 35.8% in FY’22 to 25.5% in FY’25.

Shrinking EBITDA Means a Lower Denominator…

And leverage has ballooned. Rackspace is currently sitting on $2.5b in debt and LTM EBITDA of $287.1mm; net leverage shakes out to ~8.3x. Remember, Apollo originally took Rackspace private at 4.0x.

Rackspace has been struggling with its debt load ever since the LBO. A shrinking income statement is one thing, but having a bloated balance sheet as well is the distressed world’s one-two punch. And with that comes some classic operational changes. Earlier this year, Rackspace laid off 750 employees (~15% of its workforce) and implemented massive price hikes.

On the financial engineering side of things, there was, most notably, a drop-down and debt exchange in ’24 that reduced debt by over $300mm and lowered cash interest by more than $11mm. Oh also… some non-participating lenders had their collateral stripped and primed. But hey, more runway to figure out a working go-forward strategy.

Which brings us to the future

It’s an AI story. No, seriously, it is.

Rackspace’s Private Cloud segment is known for doing a few things great – namely reliability and security (dedicated single-tenant hosting model). Industries like financial services and healthcare that want to utilize AI need assurance that their customer data is protected. They can’t just contract with an AI company like Palantir without that assurance… unless… Palantir has a relationship with a reliable hosting service like Rackspace!

Rackspace announced the partnership in February 2026, sending the stock from $0.42 to almost $1.40 in a day. Rackspace has already trained ~400 employees on Palantir’s platform and expects to ramp that number up.

Fast forward to May 2026 and Rackspace announces another AI partnership – this time with AMD. The stock spikes to $5.50.

Together, the two companies plan to work towards the creation of a global enterprise AI infrastructure and solutions provider. AMD GPUs will power the project while Rackspace will provide the operational piece (data centers, software stack). Both parties upgrade the MOU to a definitive agreement on June 16, 2026, and management hosts a call that lays out plans for a 30 MW footprint across Rackspace’s data centers. The stock jumps to $7.50.

On July 9, 2026, management hosts another special call to discuss the venture and provide some financial updates… including a $150mm downward revision of FY’26 revenue guidance. Management also announced a $250mm ATM offering to pay for AMD GPU-related initiatives. The stock drops to $4.40.

There’s a non-zero chance those equity-financed GPUs become collateral in a future asset-backed debt raise as well. There’s potentially one last slap in the face to equity holders.

The good news is management expects the AMD partnership to reach 15 MW by 2027 and 30 MW by 2028. Management is also estimating revenue per MW to be between $15mm and $20mm, so at maturity, the venture could be generating between $450mm and $600mm in revenue. Rackspace thinks EBITDA margins will fall in the 50% range, so that’s $225mm to $300mm in EBITDA. The company is targeting a small initial 2 MW deployment for completion by the end of 2026, which will cost ~$75mm in capex or $37.5mm per MW. Assuming Rackspace realizes some deployment efficiencies going forward and that capex number goes down to $30mm per MW, you’re looking at ~$900mm for the full 30 MW deployment. Not cheap.

If the AMD ramp up hits all of management’s targets, net leverage gets turned down 4x (at current debt levels); that’s a transformation. But that’s conditioned on Rackspace hitting the milestones on time and finding adequate financing for the rollout – the $250mm ATM covers maybe just a quarter of the needed capex.

The full deployment year of 2028 is also Rackspace’s $2.5b maturity wall. Any slight delay or issues with the AI ventures could send Rackspace back into the discount bin.