2nd September 2026
Scotland has spent years being told that renewable energy will transform its economy.
Now another industry has arrived that wants an enormous amount of it.
Artificial intelligence requires computing power on a scale that would have been difficult to imagine only a few years ago. Behind the impressive claims about AI assistants, automated systems and the next generation of digital technology lies something rather less glamorous.
Huge buildings filled with computers.
And those computers need electricity.
Lots of it.
That has suddenly made Scotland an attractive proposition.
The country has vast renewable-energy resources, large areas of land and an electricity system increasingly dominated by wind generation. For technology companies looking for locations for enormous data centres, Scotland can look like a very attractive place to build.
But there is a question that should be asked before we become too enthusiastic.
Who actually benefits?
There is nothing wrong with attracting investment.
Indeed, Scotland needs more of it.
A major data centre can generate construction work, create demand for engineering and technical services and provide business for contractors. It can also contribute to the wider digital economy and potentially attract other technology investment.
The problem is that the size of the building can give a misleading impression of the number of jobs that will ultimately exist inside it.
A hyperscale data centre can occupy an enormous area and consume huge amounts of electricity while requiring relatively few permanent employees compared with a traditional manufacturing plant of similar physical scale.
That does not make it a bad investment.
But it does mean that the economic case needs to be examined carefully.
If Scotland provides the land, the grid connection, the infrastructure and a large quantity of renewable electricity, it is reasonable to ask what Scotland receives in return.
The question becomes even more important when we consider the scale of what is being proposed.
Analysis of Scottish data-centre proposals has suggested that a large number of developments could ultimately require several thousand megawatts of electricity.
That is an extraordinary demand.
Scotland is building renewable generation at an impressive rate, but electricity generation and electricity demand do not automatically match. The electricity has to be moved around the country. The grid has to be strengthened. Connections have to be available. And the system has to cope with demand when the wind is not blowing.
A data centre cannot simply decide that it will close for three hours because the wind has stopped.
Its customers expect constant availability.
That creates another interesting question for Scotland's energy strategy.
If huge new data centres consume increasing amounts of electricity, how much of the country's renewable generation is effectively being reserved for global technology companies?
And what does that mean for everyone else?
This is not an argument against renewable energy.
Quite the opposite.
If Scotland has an abundant supply of clean electricity, using some of it to attract investment makes sense.
But electricity is also an economic resource.
It can attract manufacturing.
It can support hydrogen production.
It can help existing businesses expand.
It can reduce the cost of electrifying transport and heating.
It can help create entirely new industries.
So there is an opportunity cost.
Every large industrial consumer of electricity changes the balance of the system.
That is why simply saying that a data centre will run on renewable energy does not answer all the questions.
There is also the land.
There is the water.
There is the road infrastructure.
There is the grid connection.
There is construction traffic.
There may be noise.
And there is the visual impact.
These things matter particularly in rural communities, where the promise of economic development has sometimes been made before, only for local people to discover that the permanent economic benefit is much smaller than expected.
This is where Scotland should learn from its own history.
For generations the Highlands have supplied resources to the wider economy.
Timber.
Water.
Land.
Electricity.
Oil and gas.
Renewable energy.
The Highlands can become a very valuable resource base for the new digital economy as well.
But there is a difference between supplying a resource and building an economy.
If Scotland simply provides the electricity that allows international technology companies to operate enormous computing facilities, it risks repeating the same old pattern.
The resource is here.
The economic value is largely elsewhere.
That is not inevitable.
The terms can be negotiated.
If a company wants a massive electricity connection, perhaps the deal should include substantial investment in local skills.
If it wants to build in a rural area, perhaps it should contribute to infrastructure.
If it requires specialist workers, perhaps it should establish training and apprenticeship programmes.
If it is benefiting from Scottish universities and research expertise, perhaps there should be stronger links between the company and those institutions.
And if the facility requires a huge amount of grid investment, there should be a clear discussion about who pays for it.
None of this is anti-business.
It is what a sensible investment strategy looks like.
The same argument applies to tax.
Politicians often become excited when they hear that billions of pounds of private investment are coming into a region.
But investment is not the same thing as economic value retained locally.
A company can spend billions building something and still generate relatively little permanent employment.
Conversely, a much smaller investment in a manufacturing business might create hundreds of skilled jobs and generate a network of local suppliers.
The numbers need to be understood rather than simply celebrated.
There is another issue that is likely to become increasingly important: water.
Data centres can use water for cooling. At present their overall share of Scottish water demand may be small, but that tells us little about what happens if the industry expands rapidly.
The sensible question is not whether today's demand is manageable.
It is whether the system remains manageable when today's data centres become tomorrow's much larger digital infrastructure.
AI is likely to require increasing computing capacity.
The demand is not going away.
That means Scotland has a choice.
It can welcome the industry on almost any terms because it wants investment.
Or it can recognise that Scotland possesses something the industry desperately needs and use that position to negotiate a better economic return.
I would argue for the second approach.
Scotland should welcome the investment.
But it should not be dazzled by the word "AI".
A giant building containing thousands of computers is not automatically an economic transformation.
The real measure should be what it leaves behind.
Are there more skilled workers?
Are there more apprentices?
Are local companies winning contracts?
Are universities benefiting?
Is infrastructure being improved?
Are communities receiving something tangible?
Is the grid being strengthened in a way that also helps other businesses?
Are the economic benefits going to last longer than the construction project?
Those are the questions that matter.
There is potentially a particularly interesting opportunity for the Highlands.
The region has land, renewable energy and communities that need new economic opportunities.
A properly structured data-centre industry could bring investment into areas that have traditionally struggled to attract major technology projects.
But that will only happen if the benefits are deliberately connected to the communities in which the infrastructure is built.
Otherwise the risk is that we end up with a rather strange form of economic development.
The countryside provides the land.
The Scottish energy system provides the electricity.
Local communities provide the roads and infrastructure.
And international technology companies collect most of the value.
That would be a poor bargain.
The AI revolution is probably going to happen whether Scotland participates or not.
The question is whether Scotland participates intelligently.
We should not be afraid of data centres.
We should not automatically oppose them.
But neither should we assume that every billion-pound investment is necessarily good news simply because the headline figure is impressive.
Scotland has something valuable.
It has renewable energy.
It should use that advantage to build an economy, not merely to become a convenient electricity supplier for somebody else's.
The question facing ministers and local communities is therefore quite simple.
When the computers are switched on, the servers start running and billions of pounds of digital activity begins passing through these buildings, what will actually be left in Scotland?
If the answer is good jobs, new skills, stronger infrastructure, local businesses and long-term investment, then the AI boom could be a very good thing for Scotland.
If the answer is mainly a very large electricity bill and a very large building, we should perhaps have negotiated a better deal.