As Scotland debates the future of hyperscale AI data centres, a much bigger question is emerging. Can we build the infrastructure needed for the AI era in ways that are technically credible, environmentally responsible and trusted by the communities they serve?
When we published Are We Sleepwalking into a Grid Crisis? in February and The Hidden Cost of Intelligence in April, we argued that the rapid growth of artificial intelligence would eventually force us to confront an uncomfortable reality. AI may feel virtual, but it depends on very real infrastructure.
At the time, much of the discussion focused on future demand. Could electricity grids cope with the explosion in AI workloads? Would renewable energy generation keep pace? Were governments thinking strategically enough about the infrastructure an AI-powered economy would require?
Only a few months later, those questions have moved from technology journals into mainstream politics. What began as a debate about energy and sustainability is rapidly becoming a debate about something even more fundamental: trust. Can governments deliver on their promises? Can developers substantiate ambitious environmental claims? And can communities have confidence that the infrastructure being built in the name of AI will genuinely serve the public interest as well as commercial ambitions?
Reports that the SNP’s (Scottish National Party) National Council has backed calls for a pause on new hyperscale data centre developments in Scotland have intensified a debate that is now being repeated across Europe and beyond. The Scottish Green Party has argued that the scale of proposed developments threatens energy supplies, water resources and climate commitments. Business organisations, on the other hand, warn that a moratorium risks deterring investment and weakening Scotland’s competitiveness in one of the world’s fastest-growing technology sectors.
Both sides make valid points, but focusing solely on whether Scotland should build more data centres risks missing the bigger story. The real question is not whether we need AI infrastructure, we undoubtedly do, but whether we are building it in ways that are technically credible, environmentally responsible and capable of earning the trust of the communities that will host it.
That debate is only just beginning.
AI has become an infrastructure story
Artificial intelligence is often presented as a software revolution, but software is only the visible layer. Every AI model, chatbot, recommendation engine and intelligent business application ultimately depends on physical infrastructure: servers, storage, fibre networks, substations, cooling systems and enormous quantities of electricity.
As organisations race to adopt AI, that infrastructure is expanding at an extraordinary pace. Scotland currently has around two dozen proposed hyperscale data centre developments and estimates suggest that, if every proposal proceeded, they could collectively require more electricity than Scotland’s current peak demand.
Whether every proposal is eventually built is almost beside the point. Some projects will inevitably be delayed, redesigned or abandoned as commercial realities, planning decisions and energy constraints take effect. The more significant issue is that AI infrastructure is beginning to compete directly with housing developments, manufacturing, transport electrification and other industries for finite resources. Electricity generation, grid capacity, planning approvals, water supplies and skilled labour are no longer background considerations, they are becoming strategic assets.
This is not uniquely Scottish. Across Europe, North America and Asia, governments that have enthusiastically embraced AI are discovering that the physical infrastructure needed to support it cannot be delivered at the same speed as software innovation. AI may evolve in months; substations, transmission lines and renewable energy projects often take years.
That mismatch between digital ambition and physical reality is becoming one of the defining challenges of the AI era.
When ambition meets reality
The project was announced with ambitious promises of generating up to one gigawatt of renewable energy to power the site – roughly equivalent to the output of a small nuclear reactor. Yet documents obtained through Freedom of Information (FOI) requests suggest that, behind the scenes, developers and government officials were already acknowledging significant challenges around securing sufficient power. Questions have also been raised about whether the renewable generation required could realistically be delivered within the proposed timescales.
DataVita has responded by stating that its energy strategy combines new renewable generation, private-wire infrastructure and intelligent grid connections, with delivery subject to commercial agreements, planning approval and grid availability.
None of this necessarily means the project will fail, nor should it be taken as evidence that Scotland’s AI ambitions are fundamentally misplaced. However, it does highlight a broader issue that extends far beyond a single development.
As governments compete to attract AI investment, announcements increasingly describe what projects aspire to become rather than what has already been secured. Renewable generation, grid capacity, planning approvals and infrastructure delivery are often presented as though they are guaranteed, when in reality many remain dependent on future decisions and lengthy approval processes.
Political ambition is moving at AI speed, while infrastructure still moves at the pace of engineering, regulation and construction. That gap matters because public confidence depends upon credibility. If communities begin to feel that environmental claims or economic benefits have been overstated, trust quickly erodes, not just in individual ‘local’ projects but in the wider global AI agenda.
The economic case remains compelling
None of this should be interpreted as an argument against data centres. Modern economies depend on them, and that dependence will only increase as AI becomes embedded across every sector.
Financial services, healthcare, government, manufacturing, telecommunications and scientific research all rely on resilient digital infrastructure. AI simply accelerates that trend by increasing demand for high-performance computing and low-latency networks.
Business leaders are therefore right to describe data centres as strategic national infrastructure. Countries that fail to invest risk becoming consumers of AI rather than contributors to its development, importing innovation rather than creating it. The economic opportunities are significant, extending well beyond the facilities themselves into construction, specialist engineering, software development, cybersecurity and wider digital ecosystems.
At the same time, communities are equally justified in asking difficult questions. How many long-term jobs will these facilities actually create? Who benefits from the enormous demand they place on electricity networks? What happens if local infrastructure must be upgraded at public expense? Can renewable energy commitments be independently verified? How will water consumption be managed, and who bears the environmental cost if those commitments fall short?
These are not anti-technology questions. They are precisely the questions responsible societies should ask before approving infrastructure that will shape local economies and environments for decades.
Moving beyond jobs versus the environment
Perhaps the most encouraging development this week came not from Scotland but from Italy.
As reported by IT Brief UK, Equinix and the energy company A2A have announced one of Europe’s largest data centre heat recovery projects in Milan. Instead of allowing the vast quantities of heat generated by servers to dissipate into the atmosphere, the excess heat will be captured and fed into the city’s district heating network. Once fully operational, the scheme is expected to provide heating for more than 21,000 homes while reducing carbon emissions and improving overall energy efficiency.
The technology itself is not revolutionary. Scandinavian countries have been integrating waste heat from industrial processes into district heating systems for decades. What is significant is the change in mindset. Rather than treating data centres simply as consumers of electricity, projects such as this recognise that they can also become contributors to wider community infrastructure.
That raises a more constructive question than the increasingly polarised debate over whether AI data centres should be built at all.
What if every major AI data centre was expected to demonstrate not only how it would minimise its environmental impact but also how it would create measurable local value? Waste heat could support nearby housing developments, renewable generation could be genuinely additional rather than competing for existing capacity, battery storage could strengthen local grid resilience, and partnerships with universities and colleges could help develop the skills needed for the digital economy.
The conversation changes. Instead of asking communities simply to accept the costs of AI infrastructure, developers and governments would be challenged to demonstrate the benefits in ways that are visible, tangible and locally relevant.
Why organisational leaders should care
It would be easy for business leaders to see this as a debate for planners, politicians and energy companies. That would be a mistake.
Every organisation pursuing AI has become part of this infrastructure story. Whether your AI applications run on Microsoft Azure, Amazon Web Services, Google Cloud or another platform, somewhere there is a physical data centre consuming electricity, requiring cooling and competing for scarce resources.
For years, cloud computing allowed organisations to abstract away the physical reality of digital infrastructure. Computing capacity appeared almost limitless and concerns about where workloads actually ran became someone else’s responsibility. AI is exposing the limits of that assumption.
Boards are beginning to ask different questions. Can our AI strategy scale if computing infrastructure becomes constrained? How resilient are our cloud providers? Are suppliers’ sustainability claims supported by credible evidence? Will growing regulatory scrutiny affect future costs or availability? What happens if communities increasingly resist major infrastructure developments?
These are no longer operational questions for IT departments. They are becoming strategic questions about resilience, governance, ESG commitments, supply chain risk and long-term competitiveness.
The organisations that succeed with AI will not simply deploy more sophisticated models than their competitors. They will also understand the infrastructure that underpins those models, ask harder questions of technology partners and recognise that responsible AI extends beyond algorithms to include the environmental and social impact of the systems on which those algorithms depend.
In that sense, the debate over data centres is really a debate about digital leadership itself.
Building trust as well as infrastructure
This week’s headlines demonstrate that the AI data centres debate is not going away. While Scotland considers whether to pause new hyperscale developments, new multi-billion-pound proposals continue to emerge elsewhere in the UK. According to the BBC, Alderney is exploring whether a data centre could transform its economy, while projects such as Milan’s show that alternative models are possible. Around the world, governments continue to compete aggressively for AI investment even as public scrutiny becomes more intense.
Perhaps that growing scrutiny is exactly what is needed.
The AI revolution will require substantial new infrastructure, but its long-term success will depend on far more than engineering expertise or investment capital. It will depend on trust: trust that energy promises are achievable, trust that environmental claims are backed by evidence, trust that local communities will share in the benefits rather than simply carrying the costs, and trust that governments are balancing economic ambition with responsible planning.
For organisational leaders, there is an equally important lesson. Digital transformation has always been about more than technology, and AI is no exception. Every AI strategy is connected to a network of physical assets, energy systems, supply chains and communities. Ignoring those connections is no longer an option because they increasingly shape business resilience, corporate reputation and the ability to innovate at scale.
The debate over AI data centres is therefore not really about buildings. It is about what kind of digital future we are choosing to build, how we communicate the trade-offs involved openly and honestly, and whether businesses, governments and communities can work together to create infrastructure that is both economically valuable and socially sustainable.
If the first phase of the AI revolution was about discovering what the technology could do, the next phase will be about proving that we can build the infrastructure to support it responsibly. That may ultimately prove to be the more difficult challenge and the one that determines whether AI earns the public trust it will need to fulfil its extraordinary potential.
————————————————————————————————————-
If you found this article of interest, please don’t forget to sign up for our NEWSLETTER for the latest industry news and insights delivered direct to your mailbox.
Image of Scottish Parliament Chamber by Maja R. on Unsplash
Comments are closed.