British firm Stellanor snaps up data centre amid scramble for AI capacity
Stellanor Datacenters has expanded its UK footprint with the acquisition of a specialist facility from Imagination Technologies, as demand for AI infrastructure continues to surge.
The deal, structured as a ‘sale-and-service-back’ agreement, will see Imagination remain in the Hemel Hempstead site as a customer, while Stellanor takes ownership and integrates the facility into its growing network.
It takes Stellanor to 11 sites across the UK, up from just two at its launch in September last year, a rapid buildout that reflects the pace of investment flowing into data centres.
The newly acquired facility is designed to support GPU and AI chip workloads, placing it squarely in the most in-demand segment of the market as companies scale computing power for AI.
Capacity race heats up
The news comes as the UK data centre market enters a period of intense competition, driven by the exponential growth in AI and cloud computing.
Total UK capacity stood at around 1.6GW in 2024 but is forecast to more than double by the end of the decade, and still fall short of demand.
Developers have rushed to secure sites and grid connections, with planning applications hitting record levels last year.
Alongside the Imagination deal, Stellar is in the process of acquiring eight additional UK facilities from Redcentric in a deal worth up to £127m, expected to complete in the coming months.
Together, the transactions position the DWS-backed firm as a fast-emerging player in the UK’s urban data centre segment, focusing on facilities located close to end users rather than hyperscale campuses in remote areas.
The appeal of that model is becoming clearer as AI workloads demand both power and proximity. Chip design, financial services and other latency-sensitive industries are increasingly looking for infrastructure located near operational hubs.
For Imagination, the agreement allows it to retain access to critical infrastructure without owning the asset outright, which is a method becoming more common as companies look to free up capital while maintaining computing capacity.