Digital Sovereignty Trumps NIMBYism

The real fear Scotland should be facing is wholesale deIndustrialisation

I support the proposed hyperscale data centres (DCs) being sited in Scotland. If data is the new oil, then let it rain DCs here. With the right conditions in place, data centres could help drive growth, productivity and Scotland’s future prosperity.

I am no feral NIMBY, one major data centre is planned for my own backyard, on the site of the old Cockenzie Power Station. I am not against it per se. But I am against Scotland sleepwalking into a model where communities host strategic digital infrastructure while the benefits flow elsewhere and the risks remain local.

The artificial intelligence revolution is arriving at pace. As Otto von Bismarck said of another revolution, “It is better to start it ourselves than to suffer from it.” In a world of shifting political allegiances, digital sovereignty matters.

It matters because countries and businesses must avoid vendor lock-in. Critical systems should not become dependent on a handful of overseas software or cloud providers. It matters because jurisdictional control matters: sensitive public, commercial and personal data should remain subject to domestic law and democratic oversight. It matters because sovereign AI infrastructure will determine who controls access, terms of use and strategic capability. And it matters because national resilience is weakened when essential services depend on long, fragile chains of infrastructure stretching across borders.

AI is already transforming the business landscape. I run a software company, and complex tasks that used to take six months, such as building a data warehouse or an application, can now take weeks. Soon they may take days. It is hard to see the full impact unless you work close to it, but in the private sector AI is everywhere. It is being pushed from the workers up and from the bosses down. AI won’t fixed broken business processes but it does shine a glaring light on the hidden holes. Change is a happening and we need to be ahead of it otherwise our economy will be overtaken by faster movers.

Amara’s Law is worth recalling: we tend to overestimate the effect of a technology in the short run and underestimate it in the long run. In a few years’ time, nobody will call themselves an AI company. They will either simply be using AI as a matter of course, or they will be gone, overtaken by leaner competitors. Scotland cannot afford to spectate from the sidelines. We must be in the race if we want to benefit from it and that means building DCs here.

Concerns about noise, light pollution and water use are legitimate, but they are manageable through planning conditions, engineering standards and local accountability. These issues should be addressed seriously, not used as a blanket excuse by the Greens to block strategic development. The AI world will not wait for Scotland to produce the perfect national strategy. If we move too slowly, we will once again be left behind.

There are questions over land use that must be resolved. Nobody wants to see pristine countryside bulldozed for a Data Centre, the countryside has already suffered greatly from sprawling wind-farms. Equally, the risks must not all fall on working class communities. A balance needs struck but Scotland has no shortage of brownfield sites. Those expressing climate concerns should also remember that it is far more energy-efficient to process information in a purpose-built data centre than across millions of individual phones, laptops and devices.

My primary concern about the proposed data centres is not that they exist. It is whether Scotland has the firm electricity capacity to power them properly. The uncomfortable truth is that many of these projects may never be built because Scotland’s energy infrastructure is too weak, too expensive and too dependent on intermittent supply. Ironically, the political obsession with “clean” energy could make data centres far dirtier than they need to be if developers are forced to rely on vast banks of diesel backup generators.

For example, the proposed 300MW data centre at Larbert has been reported as requiring hundreds of backup diesel generators. A larger development at Cockenzie could require many more. No community hosting a data centre should be expected to tolerate what could, in effect, become a diesel backup power station during periods of stress or outage. At least the old coal-fired power stations had tall chimneys to disperse exhaust gases! Rows of diesel generators beside homes and schools are not an acceptable substitute for serious energy planning.

The solution is obvious: firm power should be built alongside strategic data infrastructure. In the United States, major data centre developers often invest directly in dedicated power supply. Scotland should expect the same. The Cockenzie site could host firm generation, whether through gas or future small modular nuclear technology The principle should be clear: continuous power, minimal diesel reliance and no dumping of environmental risk onto working-class communities.

Local benefits also need to be real. Offering nearby businesses discounted access to AI infrastructure, cloud services or training could help data-centre areas become genuine economic clusters. But on previous form, there is a risk that the money flows into government coffers while the burdens remain local.

Scotland should welcome data centres, but not as passive landlords. If these facilities are strategic infrastructure, they should be treated as such. That means firm power, minimal diesel reliance, enforceable local benefits and a share of the upside for the communities that host them. It’s not the limited development of AI data centres that we should fear but the wholesale deIndustrialisation facing Scotland.