28th August 2026
Britain Helped Invent the Computer. So Why Is It So Difficult to Build Data Centres Here?
There is something slightly strange about Britain's position in the modern data-centre race.
We were here at the beginning.
In 1943, while Europe was still at war, a Post Office engineer called Tommy Flowers was quietly building a machine that would change computing history. Working at the Post Office Research Station at Dollis Hill in London, Flowers developed Colossus, an electronic machine designed to help Bletchley Park break the German Lorenz cipher.
It was an extraordinary achievement.
Colossus was the world's first large-scale programmable electronic digital computer. It used thousands of electronic valves and was designed and built in Britain at a time when the idea of a machine carrying out electronic calculations on this scale was still revolutionary.
By early 1944, Colossus was working at Bletchley Park. Eventually there were ten machines operating there.
The technology was so secret that much of the story remained hidden for decades after the war.
Britain therefore has a remarkable claim to being one of the birthplaces of modern computing.
So why, more than 80 years later, does Britain sometimes appear to struggle to build the enormous data centres needed to run the next generation of computing?
The answer is not that Britain has forgotten how to build computers. It is that modern computing has become an energy problem.
The computer has changed but so has its appetite
Colossus was enormous by the standards of 1943.
But today's artificial intelligence and cloud-computing data centres make it look tiny.
A modern hyperscale data centre can require hundreds of megawatts of electricity. The newest AI facilities being planned around the world can require power on a scale more commonly associated with industrial complexes than computer rooms.
That changes the question completely.
It is no longer simply:
Can Britain build the building?
It is:
Can Britain provide hundreds of megawatts of reliable electricity, connect it to the site quickly, provide cooling and construct the supporting infrastructure?
And increasingly there is another question:
Can it do all of that without undermining the country's climate objectives?
That is where Britain is encountering difficulties.
The electricity connection is becoming the bottleneck
The UK has a substantial amount of electricity generation, and Britain is investing heavily in offshore wind, nuclear power, batteries and other forms of low-carbon generation.
But electricity generation and electricity availability at a particular location are not the same thing.
A company can find a large piece of land and propose a data centre, but that does not mean the electricity grid nearby has the capacity to supply it.
And upgrading the grid takes time.
This is now becoming one of the biggest obstacles facing data-centre developers.
Ofgem is considering changes to the electricity connection system because the enormous increase in applications for new connections has created a queue containing projects which may never actually be built. Recent reporting suggests that speculative data-centre applications are contributing to the problem, creating what has been described as a "phantom" demand for electricity capacity.
That creates an extraordinary situation.
A genuine data centre can be ready to invest billions of pounds, employ people and bring computing capacity into Britain — but it can still find itself waiting for a grid connection.
Britain has created another problem for itself
There is an irony here.
Britain has spent years encouraging renewable electricity generation.
We now have enormous offshore wind potential around our coastline.
Scotland in particular has become one of Europe's major renewable-energy regions.
But the electricity network was not originally designed around this new geography of power.
Much of the electricity generation is located far from the biggest centres of demand.
The result is that Britain can simultaneously have enormous renewable-energy resources and difficulty supplying a particular new industrial customer with the power it wants.
That may sound ridiculous.
But electricity networks are physical infrastructure.
You cannot simply send an email to the grid and ask for another 500 megawatts.
You need substations, cables, transformers, transmission capacity and planning permission.
And those things take years to build.
The data-centre boom is enormous
The reason this matters is that demand for computing power is growing exceptionally quickly.
Artificial intelligence has changed the economics of data centres.
A traditional data centre might have been primarily concerned with cloud computing, websites, storage and business applications.
AI introduces a different level of electricity demand.
Large numbers of specialised processors need to operate simultaneously, and those processors generate enormous amounts of heat.
That heat has to be removed.
Consequently, a data centre is effectively an electricity-to-computing-to-heat conversion machine.
The more computing power we want, the more electricity we need.
And that is why the data-centre debate is rapidly becoming part of the energy debate.
Some developers are beginning to work around the grid
The most extraordinary consequence of the grid shortage is that developers are increasingly considering generating electricity themselves.
Two proposed data centres in England have attracted attention because of plans for on-site gas-fired generation after difficulties obtaining sufficient grid capacity. Analysis cited by the Guardian suggests that the two projects together could require around 1.3GW of electricity when fully operational.
That produces another irony.
Britain wants more data centres to support artificial intelligence and the digital economy.
But if the grid cannot provide the electricity quickly enough, developers may build their own power stations.
That could mean new gas generation being built specifically to power some of the world's most advanced computing facilities.
The debate therefore becomes much bigger than planning permission for a building.
It becomes a question about Britain's entire energy system.
And then there is water
Electricity isn't the only requirement.
Computers produce heat, and heat has to be removed.
Many large data centres therefore require significant cooling infrastructure and, depending on the technology used, potentially substantial quantities of water.
That creates a particularly difficult problem in parts of southern England where water resources are already under pressure.
A proposed data-centre campus in east London, for example, has attracted considerable attention because of its enormous projected electricity consumption and environmental footprint. Planning documents suggest that the proposed facility could consume around 2.65 billion kilowatt-hours of electricity annually.
At that point we are no longer talking about something comparable to an office block with a few computer rooms.
We are talking about infrastructure on the scale of a major industrial development.
Perhaps Britain is asking the wrong question
The usual argument is that Britain needs to attract data centres because they represent the future.
That is true.
Artificial intelligence, cloud computing and digital services are becoming fundamental parts of the economy.
If Britain doesn't have enough computing capacity, British companies may have to buy it from overseas.
That creates issues around economic value, national security, resilience and control over data.
But the opposite argument is also important.
Britain cannot simply say "yes" to every proposed data centre without asking where the electricity, water and infrastructure will come from.
The challenge is therefore not whether Britain should have data centres.
It is where they should go.
And that is where Scotland — and particularly the Highlands — could have an interesting opportunity.
What if we stopped thinking of data centres as London developments?
Britain's biggest constraint is not necessarily a shortage of land.
It is the combination of land, electricity, connectivity and planning.
The Highlands have something potentially very valuable: enormous quantities of renewable energy nearby.
Scotland has some of Europe's best wind resources.
The North of Scotland also has major transmission infrastructure investment underway.
And there is another increasingly important factor.
Data centres need reliable communications.
They need high-capacity fibre connections to the rest of Britain and the world.
That means the ideal location isn't simply a remote field with a wind turbine beside it.
It needs to be an area where power, communications, land, cooling and infrastructure can all come together.
That is where places such as Caithness could potentially enter the conversation.
Could the Highlands become part of Britain's computing infrastructure?
This is where the debate becomes interesting.
Caithness has spent decades being associated with energy.
Dounreay demonstrated the importance of nuclear research and engineering.
The Pentland Firth and surrounding waters have enormous renewable-energy potential.
Offshore wind is becoming an increasingly important part of Scotland's energy system.
Transmission investment is changing the electricity infrastructure across the north.
So perhaps the question shouldn't simply be:
How do we get more electricity out of the Highlands?
Perhaps it should also be:
How do we use more of that electricity in the Highlands?
A large data centre could potentially become one answer.
It would not solve every local economic problem.
Nor would it necessarily create thousands of permanent jobs; modern data centres can be surprisingly automated once built.
But the construction investment could be enormous, while the facility would require engineers, technicians, security, maintenance, electrical specialists and other skilled workers.
More importantly, it could create a new type of economic infrastructure.
Instead of exporting renewable electricity south and importing digital services back north, some of the economic value of that electricity could potentially be captured locally.
That is a very different proposition.
Tommy Flowers offers an extraordinary lesson
There is another reason why the story of Colossus matters.
Flowers did not wait for somebody else to invent the future.
He saw a problem, understood what technology could do and built something that many people initially thought was too ambitious.
The National Museum of Computing records that Flowers' all-electronic approach was regarded with scepticism because of concerns about the reliability of the large number of valves required. Yet his machine worked, and Colossus became a landmark in computing history.
That is very British in its own way.
The country had the engineers.
It had the scientific expertise.
It had the communications infrastructure.
And it had an urgent problem that needed solving.
The result was one of the foundations of the computer age.
The question today is whether Britain can recreate that combination of engineering ambition and infrastructure planning.
Because the danger is that we become very good at inventing the technology while becoming increasingly bad at providing the physical infrastructure needed to operate it.
We could end up exporting the economic opportunity
This may ultimately be the biggest risk.
The companies developing the world's most advanced artificial intelligence systems do not necessarily need to locate their computers in Britain.
They can put them in the United States, Scandinavia, Ireland or elsewhere.
If Britain makes it too difficult, too slow or too expensive to provide the electricity and infrastructure they require, investment will simply go somewhere else.
The irony would be extraordinary.
Britain could have some of the world's best universities, engineers, scientists and AI researchers while the actual machines doing the computing are sitting in another country.
We would have invented the technology but imported the infrastructure.
And that would be a very expensive form of nostalgia.
Britain doesn't need to build data centres everywhere
There is a sensible middle ground.
Britain shouldn't allow developers to build enormous facilities wherever they can find a field and persuade somebody to provide a power connection.
But neither should we make it almost impossible to build them.
Instead, Britain could identify locations where the infrastructure makes sense.
Areas with renewable electricity.
Areas with transmission capacity.
Areas with available land.
Areas with appropriate fibre connectivity.
Areas where water and environmental constraints can be properly managed.
And areas where local communities can actually benefit from the investment.
That would turn data-centre planning from a battle over individual developments into a national infrastructure strategy.
Perhaps the biggest lesson is about thinking ahead
The British problem with data centres is not really a failure of computing.
It is a failure to connect several different infrastructure decisions.
Energy policy is being developed in one place.
Planning policy somewhere else.
Digital infrastructure somewhere else.
Water policy somewhere else.
Economic development somewhere else.
Yet a data centre needs all of them simultaneously.
Tommy Flowers did not have that problem.
He had one job: solve the problem.
He and his team built the machine, tested it and got it working.
Britain now faces a much larger version of the same challenge.
We know computing is going to become more important.
We know artificial intelligence will require enormous amounts of computing power.
We know data centres require huge quantities of electricity.
And we know Britain has substantial renewable-energy resources.
The question is whether we can connect those facts together quickly enough.
From Colossus to the next generation
There is something wonderfully ironic about the story.
In 1943, Britain built a revolutionary computer in a London laboratory using technology that seemed impossibly ambitious.
Today, Britain is trying to build the infrastructure required for computers millions of times more powerful.
We have more scientific knowledge, better engineering, better communications and vastly more powerful technology than Tommy Flowers could ever have imagined.
But we are discovering that the difficult part may no longer be building the computer.
It is building everything around it.
Electricity networks.
Substations.
Fibre.
Cooling systems.
Planning infrastructure.
Water supplies.
Roads.
Buildings.
And the political will to decide where these facilities should go.
That is why the data-centre debate matters.
It is not really about server rooms.
It is about whether Britain is prepared to build the physical infrastructure of the next economy.
And perhaps, just perhaps, the Highlands should be asking whether some of that infrastructure could be built here.
After all, Britain helped invent the computer.
It would be rather unfortunate if, 83 years after Tommy Flowers built Colossus, we discovered that we could no longer find somewhere to plug the next generation of computers in.
Read more about Tommy Flowers HERE