Why Devon is saying no to a data centre

The quiet market town of Great Torrington in Devon felt more like Morocco during the recent heatwave. Yet despite stifling conditions, dozens of sweaty locals amassed in the town hall on the evening of 25 June, troubled by an issue with far more lasting consequences than a hot spell. 

Earlier that day, many had heard for the first time, either by a very late letter or from their MP’s Facebook page, that a public consultation with a company called Xlinks was to be held at the town hall, at which they would have an opportunity to express their feelings about the proposal to build one of Europe’s largest AI data centres on their doorstep.

The thunderous baritone voice of Sir Geoffrey Cox MP signalled the start of proceedings, as he introduced Simon Morrish, director of Xlinks, and project director Richard Hardy, to the suspicious natives of one of the most deprived rural areas in Devon, dependent primarily on tourism and agriculture. 

Cox tried to calm the aggravated crowd as Morrish attempted to allay their concerns. Questions and accusations covered noise pollution, damage to local roads and increased traffic, light pollution, water pollution, harm to wildlife, the impact on tourism, rising local temperatures, the effect on house prices, and several other concerns. Morrish responded with a series of general assurances. However, he admitted that many of these would not be enforceable in the likely event, if it is approved and built, that Xlinks sells the data centre (he mentioned Meta and Google as potential buyers). 

The greatest rumblings from the crowd were prompted by the response to an enquiry from one resident as to whether the proposed 850 acre site for an AI data centre and an associated but separate application for a battery storage site (BESS) of some 2,700 shipping container-sized battery units, was intended to revive the £24 billion proposed Xlinks Morocco-UK Power Project which the Government withdrew support for in the previous year. Morrish replied sheepishly that this was a desired outcome. Since the delayed mailout notifications of the consultation listed a great many alleged benefits of the development, and reasons the site had been selected, but neglected to mention any hope of Xlinks salvaging their existing investment, the audience thereafter seemed dubious about Morrish’s following claims and assurances. 

One possible explanation for Morrish enduring the heat and fury of Torrington town hall that evening may therefore relate to what XLinks stood to lose. The company had previously invested heavily in a proposal for a 3,800km subsea cable, the longest of its kind in the world. The proposed interconnector would have carried solar and wind-generated electricity from Morocco’s Guelmim-Oued Noun region to a landing point on the North Devon coast near Saunton Sands. The ambitious project was formerly billed as a flagship of Britain’s renewable future, providing enough clean power, its backers claimed, for some seven million homes, about 8 per cent of Britain’s electricity.

Their plans were scuppered, however, by Ed Miliband’s department. In June 2025, the Government concluded, in a written statement to Parliament delivered by junior minister Michael Shanks, that the project was simply not in the national interest. Ministers were sceptical of the sustainability and security of reliance on power arriving from two and a half thousand miles away, at the far end of the world’s longest subsea interconnector, dependent on a single corridor over a 25-year period of the government-backed price guarantee they requested. Officials pointed to delivery risk, operational risk, and, in the wake of recent cable sabotage incidents elsewhere in Europe, the security vulnerability of betting so much of Britain’s electricity supply on a single line under the sea. Miliband’s team preferred to spend taxpayer money on wind, solar, nuclear and storage built on British soil — where the risk, and the supply chain benefit, stayed at home.

Xlinks chair and former Tesco chief executive Dave Lewis expressed “bitter disappointment”, noting that the government itself had recognised the project’s potential in 2023, when it was designated a Nationally Significant Project. Xlinks lost government support, and with it the only financing route that made the project possible, but they appear to have pivoted to a new strategy.

Morrish told the June meeting that Xlinks had already begun buying the land it needed. A second letter to locals, sent on 7 July by Richard Hardy, presented an alleged justification for why the site was chosen. Devon, he claimed, enjoyed a “Goldilocks” climate, neither too hot nor too cold, which would reduce “the need for heating and cooling” and cut energy and water use. The emphasis on heating seems peculiar. Servers generate so much waste heat once running that heating is rarely required at all, and only ever in conditions far colder than Devon experiences. Cooling, not heating, is the real and constant draw on power, which makes the choice of this popular holiday destination, known locally for its humid summers, somewhat questionable. Would it not have been more economical to select a colder location further north?

The July letter went on to state plainly that there were “no current proposals” to power the campus via the Morocco cable, and that the interconnector “remains on pause” as a wholly separate project, connected to the data campus only by the possibility (at some unspecified later phase) of a renewables interconnection, which would require its own planning process.

Yet the more plausible link between the two projects may run in the opposite direction to the one Xlinks describes. The development, comprising a data centre with an electrical load of up to 1.5GW and a 1.8GW battery energy storage facility, would require significant investment in and reinforcement of the electricity network. The data centre alone would rank among the largest electricity users in the country.

Following the Government’s decision to abandon support for the Morocco-UK interconnector, Xlinks faced the prospect of seeing years of planning, stakeholder engagement and infrastructure preparation in North Devon go to waste. The Devon data campus, therefore, might provide a means of keeping a project of national significance anchored to the same ground their investments are. A development that establishes a substantial new demand for power in North Devon also supports the argument that the region requires major new energy infrastructure. If political circumstances were to change in future, the Morocco interconnector could then be presented not as a speculative proposal seeking a customer, but as a ready-made answer to a demand that had already been created.

Political circumstances have already been changing in Xlinks’ favour. Since September 2024, the Government has classified UK data centres as Critical National Infrastructure, placing them alongside sectors such as energy, water and emergency services. This change, perhaps more significant to the project than many locals realised, reflected the ministers’ view that data centres had become essential to the country’s economic and public life.

In the initial June letter and in the meeting, Xlinks had informed residents of several Public Information Days (PIDs) to be held between 14-17 July, where they would meet them in village halls to discuss the plans. Hardy’s July letter, however, said these would now be delayed until after the summer, in order to give Xlinks more time to “talk with local residents beforehand.” 

This delay happened to coincide with two related developments in government policy. The first was a change to the Network and Information Systems (NIS) Regulations. It gave regulatory effect to the Government’s elevated view of data centres by designating qualifying facilities as essential services for cyber security purposes. It has nothing to do with planning law, and does not remove or alter developers’ consultation obligations. But such obligations are affected by the Planning and Infrastructure Bill, which by July had already passed, with its central change due to come into force on 24 July. From that date, developers of Nationally Significant Infrastructure Projects (NSIPs), a category that can now include very large data centres if ministers choose to designate them as such, will no longer be under any legal duty to hold the statutory pre-application consultations, such as the PIDs Xlinks had promised. Ministers said government guidance would still encourage developers to engage with communities. But an expectation is not an obligation, and the distance between the two is precisely where a project under pressure has room to manoeuvre.

Whether the Devon data campus will ultimately be designated as an NSIP, decided by the Secretary of State rather than Torridge District Council’s own planning committee, remains to be seen. But it seems fair to ask if this is what Xlinks is hoping for. After all, look at how the applications themselves are being structured. Two separate filings are due to land with Torridge District Council in October, one for the data centre, and another for the battery storage site (BESS), despite both the June and July letters to residents treating the two as a single project, “Devon data campus”. The July letter’s FAQ section doesn’t even mention the BESS by name even once.

The method of the applications also raises questions. The BESS is going in as a full application while the much larger data centre is going in as outline only — meaning the principle of the development is established now while the details are left for later. For a company hoping to begin enabling works in 2027 and full construction the following year, the choice is surprising: the smaller site is fully specified, while the larger one remains undefined. One possible explanation is that this approach helps navigate opposition at an early stage. Outline applications require less information, leaving fewer details open to challenge until the principle of development has already been established. 

The archaeology of the Huntshaw area is a subject I know from having made documentaries on Berry Castle and the barrows at Daracott Moor. Berry Castle, an Iron Age hillfort, lies immediately north of the proposed data campus site. To the east of the proposed BESS site, one of the Bronze Age barrows produced the Huntshaw dagger, one of the largest metal weapons in Northern Europe when it was made over 4000 years ago. There is also at least one listed building within the proposed BESS site itself, Wayhead Cottage.

Combined with the area’s ecological and landscape value, these features help explain local concern about the introduction of data-centre buildings up to forty metres high and around 2,700 battery containers into what is currently a predominantly rural setting. The woodland surrounding Berry Castle, in particular, is an important green space for Torrington residents, combining natural beauty with a landscape shaped by thousands of years of human activity.

Noise is another concern. A BESS of this scale would materially alter the character of Huntshaw’s rural soundscape. Add the data centre and the combined noise pollution would be felt across a much wider area. Residents living directly across the street from a 30 MW data centre in Dowagiac, Michigan, have recorded noise levels typically around 61 dB at their property, with readings commonly in the 50–62 dB range and described as a constant 24/7 mechanical hum audible through closed windows and over the television (local reporting from MLive and Spectrum News, June 2026). Residents near data-centre and BESS clusters in Prince William County, Virginia, have reported comparable levels in complaints submitted to local authorities.

Manufacturer data for utility-scale BESS inverters and cooling fans, as submitted with UK planning applications, typically shows sound pressure levels in the range of 70–92 decibels at one metre from the equipment. Mechanical humming noise can travel surprisingly far, in certain conditions. Under BS 4142:2014, the standard commonly used to assess industrial noise in planning decisions, a rating level around 5dB above background sound is treated as potentially adverse, while 10dB or more indicates a likely significant impact. Ambient background noise in rural Devon is very low, so this threshold could easily be exceeded.

In their July letter, Xlinks pointed to the START Campus Sines data centre in Portugal as an example of a similar development that benefited the local area. But that site is in a developed area south of Lisbon near to a refinery, power station and industrial port. This is an already noisy, busy area, hardly comparable to rural Devon.

The desire to preserve this part of the Devon countryside isn’t merely NIMBYism on aesthetic grounds

Xlinks’ board is unlikely to have welcomed Torridge District Council’s decision in early July to reject a speculative application for a modest ten-acre battery storage site at Woodtown. This BESS was also supposed to be near the Alverdiscott substation, but was opposed by residents and councillors who saw it as the “creeping industrialisation” of the countryside around the substation. The same argument is now obviously being made against the much larger Xlinks proposal. A council that would not wave through ten acres at Woodtown seems unlikely to approve eight hundred and fifty at Huntshaw. This again raises the question of whether Xlinks might seek a legal means of bypassing the council. 

But the desire to preserve this part of the Devon countryside isn’t merely NIMBYism on aesthetic grounds. North Devon’s economy runs substantially on tourism and agriculture, and both depend on the very thing a data campus and a BESS would compromise: an unspoiled landscape carrying UNESCO Biosphere status, a designation awarded precisely because this stretch of Devon has kept its beauty and its ecology intact. Once forty metre buildings and hundreds of acres of battery containers are erected on the rolling fields beneath an ancient Celtic hillfort, that beauty is lost. So, in all likelihood, are many of the visitors who currently sustain the local economy on the strength of it.

Devon is not alone in reaching this conclusion. In the United States, where the proliferation of AI data centres has advanced far further and faster, opposition has become one of the few issues to unite Americans across the political divide. A Gallup poll found 71 per cent of American voters, including 75 per cent of Democrats and 63 per cent of Republicans, do not want a data centre built in their area, and the opposition is winning: at least 75 data centre projects worth around $130 billion were blocked or cancelled in the first three months of 2026 alone. Some of that American anger is, admittedly, about AI in general rather than any specific site. One survey found only 8 per cent of opponents actually lived near the facility they were objecting to. Devon’s case, on the other hand, is not so easily dismissed. Many of those objecting to the scheme live, work, or farm within eye or earshot of it. The complaint here is not an abstract anxiety about artificial intelligence. It is a specific one, about a specific hillfort and its views, a specific river made famous by Tarka the Otter, and a specific corner of rural Devon that still feels like England, and the people making it are the people who will actually live with the consequences.

More than 26,600 people had signed the “Leave Devon Alone” petition by 15 July, a remarkable figure in a rural district of this size. This is one of the few issues to unite communities across the political lines which increasingly divide British society. Xlinks may yet secure its outline permission, its NSIP designation, its second chance at Morocco. It certainly has the right to try. But it will do so against the wishes of a community that, within hours of finding out about the proposal, met Mr Morrish in the town hall to offer a clear and resounding “no”.