From Undersea Cable to AI Campus
Xlinks, the British company that spent years promoting a 3.6GW interconnector to carry solar power from Morocco to the UK, has reinvented itself as a data center developer. On June 29 it emerged that the firm has filed proposals for a 1.5GW data center campus, branded Valeon, in Alverdiscott, Devon, paired with a 1.8GW battery energy storage facility. The two planning applications to Torridge District Council carry a combined price tag of 12 billion pounds, roughly 16.4 billion dollars, making it one of the largest single AI infrastructure proposals in the UK.
The pivot is a sign of the times. Founded in 2018, Xlinks built its identity around long-distance clean power transmission, but the UK government withdrew support for the Morocco project last year, leaving the company with engineering expertise, grid connection ambitions, and no flagship. Management has framed the data center turn as a way to maximize value for all parties in a different way. Translated, that means the same grid connections and capital relationships that would have moved Moroccan electrons are now being aimed at feeding GPUs in Devon.
What Is Actually Being Proposed
Site renderings show five buildings and an on-site substation across the Alverdiscott parcel. The headline 1.5GW of data center capacity would place Valeon among the most ambitious campuses in Europe, comparable in nameplate terms to the gigawatt-scale sites being built in Texas and Ohio. The accompanying 1.8GW battery storage facility is the clever part: it positions the campus to firm up its own supply, smooth grid draw, and potentially arbitrage power markets rather than simply consuming whatever the grid can spare.
That integration of compute and storage is exactly where serious developers are heading. A 1.5GW load in rural Devon would be impossible to serve from the existing grid without major reinforcement, so co-locating nearly equivalent battery capacity is less a green gesture than an engineering necessity. For enterprise buyers watching European capacity, projects that bundle their own power flexibility are the ones most likely to actually get built and energized this decade.
The Backlash Arrives Early
Opposition has materialized before public consultation has even begun. A Change.org petition against the project has already gathered more than 2,250 signatures, with critics framing the campus as industrialization of a rural landscape. As one objection put it, North Devon is not an empty space waiting to be industrialised, it is a living landscape of farms, villages, wildlife and communities. That language will sound familiar to anyone tracking the wave of data center moratoriums and rezoning fights sweeping the US.
The formal consultation runs July 14 to August 11, and Torridge District Council now holds the decisive role. We have watched enough of these battles to know that the technical merits, grid integration, jobs, and clean-power pairing rarely settle them. What settles them is whether the local community feels the benefits land locally or flow to distant hyperscalers. Xlinks has a clean energy pedigree to lean on, but a 1.5GW campus in a region of farms and villages is a hard sell no matter how green the battery.
The Grid Question Nobody Has Answered
Even if Torridge approves the campus, the hardest problem remains the grid. A 1.5GW load is larger than the consumption of many British cities, and the South West is not historically a region of abundant transmission headroom. Xlinks' interconnector background gives it more credibility than a generic developer on this front, but credibility is not the same as a connection agreement. The pairing with 1.8GW of batteries is the mechanism by which the project hopes to make a load this size tolerable to the network, smoothing peaks rather than slamming the grid with constant maximum draw.
This is where European AI ambitions repeatedly collide with reality. Nameplate capacity is easy to announce and hard to energize, and the queue for grid connections in the UK has been a chronic bottleneck for renewables and data centers alike. We would want to see a firm connection offer and a credible reinforcement plan before treating Valeon as anything more than an aspiration. For now it is a serious proposal from a serious team, but the electrons it needs are not yet spoken for.
A Pattern Worth Watching
Valeon is a textbook case of stranded energy infrastructure being repurposed into AI compute. Companies that spent the last decade chasing interconnectors, mining, or crypto are discovering that their grid connections, land options, and capital relationships are worth far more aimed at GPUs than at their original purpose. We have seen the same pivot from former crypto miners and power developers across multiple continents this year, and Xlinks is simply the British version.
For CTOs evaluating European AI capacity, the signal is that the UK pipeline is broadening beyond London and the established M4 corridor into rural regions with spare land and grid headroom. The catch is execution risk. A 16 billion dollar proposal from a company that just lost its previous flagship is a long way from energized racks. Treat Valeon as a marker of where European capacity could grow, not as supply you can count on for any workload before the late 2020s.
