Building a common language and shared understanding
Unless you are one of the world’s growing group of news avoiders, it cannot have escaped your notice that people are talking about ‘data centres’. They’re making videos about them, mapping them, and it seems, raising more questions about them than ever before. Up and down the UK, and across the world, these once apparently innocuous IT systems are hitting the headlines.
At Open Data Manchester (ODM), this growing debate about if and why these sites are needed, for whose benefit, at what cost and for who, hasn’t escaped our notice. That’s why we’ve been working with Adi Kuntsman, Reader in Digital Politics at Manchester Metropolitan University, to try to understand what’s going on here. But as we started digging, we couldn’t help but wonder whether everyone was really speaking the same language.
What is a data centre anyway?
Despite all the debate online, one thing you can quickly find out is that there isn’t a formal, agreed definition of a data centre. In the UK particularly, data centres do not have a standard industrial classification (SIC) code, which are used by the Office for National Statistics and HMRC so they can understand what a business does.
Perhaps understandably then, claims about their number vary:
-
In 2024, 800 “data centre operators” were identified by the Department for Science, Innovation & Technology including a claim that the “10 largest companies” are taking 80% of revenues in the UK, though this figure lacks a source or clarity on definitions.
-
The Parliamentary Office of Science and Technology (2026) claims 450 “large data centres” without a source.
-
The House of Commons Library (2026) cites a techUK report (2024) claiming 450 data centres.
-
The techUK report estimates there are 450 data centres in the UK, based on the number of “colocation, cloud, edge and hyperscale” sites they know have Climate Change Agreements in place so they can get a discount on an environmental tax, plus some, but not most enterprise sites for the sole use of one business.
Yet, each report, each explanation, each figure seems to inevitably demand knowledge of the reader that may not be readily available. Do ‘data centre operators’ operate single data centres, only data centres or something else? Does ‘large’ mean ‘large’ for a data centre or ‘large’ compared to something else?
What we can see is that data centres are a government policy priority in the UK being driven by the Department for Science, Innovation & Technology, and the actual or anticipated growth in computer processing power needed to meet demand for digital, and increasingly AI, tools. They are also a policy priority because many, or most, public and consumer services now rely on data centres to process or store data, which presents risk when things go wrong. They should also be a priority because they take significant land, need energy to run, water for cooling and backup power generation, and can also cause significant environmental harms, though the evidence on government commitment to this concern is less clear.
These priorities, which have developed over successive governments, mean this discussion spans at least six government departments, at least four regulators, a tech industry body and potentially all of the UK’s local authorities. Relevant legislation, sometimes very recently updated, dates back to at least 1991, and spans issues relating to water, planning, energy, the environment and national security, and perhaps more besides.
So what are we talking about?
From the work we have done on this so far, we can say:
“Data centres are large electrical computing systems that are used to process and store information. They require land, power, and heating, ventilation and cooling (HVAC) systems to control their temperature. Data centres may be categorised by what they are used for (e.g. enterprise use only) or be placed in particular locations for specific reasons, such as to be close to other sites so work can be offloaded if one site fails. They are a government priority because of innovation policies that are geared for economic growth, which are changing the UK’s approach to local planning, and due to national security concerns about to their widespread and growing use for processing data for public and consumer services. The extent of their current and potential environmental and social harms has been less clearly considered.”
Why is everyone talking about them?
In June, the Science, Innovation and Technology Committee, which is made up of Members of Parliament and holds the relevant government department to account, released an analysis of the current approach to “digital government”.
Including the evidence of experts in the field, they highlight there may be a lack of public consent for and limited trust in this area of government. Key concerns raised include “hype” around how much benefit these systems can bring, a lack of oversight of old systems, “vendor lock-in” with companies like Palantir, AWS and Microsoft, and an associated lack of “sovereignty”, or control, over newer digital systems. They also highlight that digital spending is not routinely monitored or shared in standard form.
The committee said that an Information Security Review done in 2023, which had not been made public until this investigation was done, showed “institutional failure and an incoherent approach to data hygiene”. They evidence this by the example of UK Biobank health data appearing on a Chinese ecommerce website. They also highlight concern with the new Labour government’s decision to move the Government Digital Service out of the cross-cutting Cabinet Office department responsible for effective governance and into the Department for Science, Innovation & Technology.
Overall the committee said it is unclear how the government plans to turn the AI promise, which is driving much of the demand for new data centres, into “tangible outcomes” and its approach lacks metrics for success or failure.
Few outside government will have read this report in full. But we may all be getting a sense of things not quite making sense. Suddenly ‘data centres’ are appearing as if from nowhere and it is not clear how or why. The next step for our work is to bring people together who are interested in this issue to help us work on agreeing a common language for talking about these developments.
We have already learned from our research that conversations about data centres can require an understanding of: physics, mechanics, electronics, computing, politics, law and regulation, economics, business, energy and environmental science. They are often about many of these things at the same time.
Without a common language and shared understanding, we may not be able to understand what a ‘data centre’ really means.