I was going to subtitle this post: “I went to five data centre panels at UKREiiF, so you didn’t have to”. However, I only made it to four. The fifth was too full. This sums up the moment we’re in- the momentum around data centres is real and seemingly growing. Across the four panels (and one closed door), I heard recurring themes, sharp debate and cautious optimism about the UK’s role in this sector. Here are some reflections.
The wave
Data centres aren’t new—but we are riding a new wave of interest, driven by the AI revolution, the digitisation of business, and investor appetite for resilient infrastructure.
The UK, already a European leader, has much to build on: strong market knowledge, mature regulation, and government support.
But how long will this wave last? Is this another “flash in the pan” conversation like gigafactories—or something more enduring?
The consensus is that demand will remain strong. AI, cloud, and connected technologies are only accelerating, and they need a home. However, whether the UK captures the full value of this growth depends on how we navigate a set of complex challenges. The situation is constantly evolving, but the current view is that whilst the UK may be well-placed for edge, inference and cloud centres, training centres—larger, denser, more power-hungry—may go elsewhere.
The challenges
I’ve heard the key challenges expressed as the “three Ps” at some data centre talks. I saw it redefined as the “five Ps” at UKREiiF: Power, Planning, Policy, People and Perception. Each merits more than its own article, but I have a few observations on each here.
Power
Whether it is three or five “Ps”, power always seems to come first.
It’s a complex subject, and I’m grateful I have colleagues who dedicate their entire legal practices to this specific area.
It is one of the great challenges of our time that the future is more energy-intensive, at a time when supply is constrained by various factors including (rightly) sustainability concerns.
It was interesting to hear from a range of practitioners on how they are trying to meet that challenge.
On the one hand, I heard Europe’s largest engineering firm discuss using sophisticated AI tools to redesign in virtual spaces, making data centres up to 40% smaller, while increasing output by 40%. On the other hand, data centres developers and hyperscalers raised that large-scale facilities are becoming the norm, seeking fewer but larger campuses of 100+ MW to facilitate economies of scale.
Some operators said they’re exploring partnerships with energy firms for private supply.
Most believe we’re generating enough power and that it’s distribution that needs improvement. “The choke point is the last mile.”
The “cheapest” energy is what we waste. Old, inefficient data centres need upgrading. We must be bolder with re-use: heat recovery, district networks, local integration.
A further theme: data centres as models of resilience. What can the broader ecosystem learn from them?
And how can centres be more flexible with energy use—up and down from a base load?
Nuclear will likely need to move inland. SMRs (Small Modular Reactors) may help and are being talked about as a “silver bullet”,but are unlikely to be widely used imminently.
While we debated this in Leeds, significant grid reform and investment was announced — a timely and welcome move.
That’s enough on power for now - I’m sure we’ll write on zonal energy pricing another day…
Planning
I was amused at one hyperscaler panellist’s plea to the gathered real estate professionals not to continue treating data centres as a traditional B8 use. I can imagine how often well-intentioned developers have approached her with schemes that are inappropriate for the end user’s required specifications. Even an outline permission can overlook many critical items, such as permitted heights. That’s before you consider other requirements, such as acoustics and fibre infrastructure.
More generally, I was struck by how much time was devoted to planning at two of the three panels. There are some encouraging signs that the system is improving at handling data centre applications - particularly the reference to data centres in the London Plan and some special arrangements, such as SEGRO’s simplified planning zone in Slough. However, the need for greater policy literacy across local authorities was a repeated theme.
The focus of planning conversations was a debate about the form applications should take - NSIP vs DCO vs TCPA.
I am left with the feeling that the TCPA route should be sufficient for most applications. However, moving into the dedicated infrastructure routes might have some merit if certain sensitivities (e.g. size and green belt) mean you’d likely end up in an appeal anyway. In those circumstances, you can accept the extra time these processes take in return for the greater certainty they would give you.
Policy
It’s easy to see why the government is currently supportive of the sector. Data centres align with consumer and business trends, contribute to economic growth, and—compared to traditional infrastructure—can be delivered quickly, with capital front-loaded. That timing works well with election cycles.
Much has changed since 2020, when data centres were hastily added to the list of key workers in biro after urgent lobbying. The recent designation of data centres as Critical National Infrastructure (CNI) is a significant step forward. Still, there was a sense among panellists that we are only beginning to understand what this designation will mean in practice.
In particular, the government's ambition to re-industrialise former industrial areas as “AI Growth Zones” is drawing attention. Key questions remain: How can we reduce power prices and expedite delivery to stimulate demand while managing locational latency and rising land costs? These projects won’t be built speculatively—so demand as well as supply must be nurtured.
Still, the opportunity is real. It makes little sense from a resilience perspective for so much critical digital infrastructure to remain concentrated in one place (west London). Encouragingly, local government representatives were out in force, making strong pitches for their communities’ digital futures—not just through data centres, but also through network development.
People
None of this will work without people who understand the sector across the ecosystem. including planners, surveyors, valuers, consultants and engineers. Training them—amid tight public and private resources—is a challenge we’ll have to meet.
One panellist raised an interesting point: in a high-stakes environment where time to market is key, is it always wise for cost consultants to scrutinise every line item for overspend? A few percentage points in cost could pale in comparison to the commercial benefits of speed.
A London Borough leader shared that it took their team two years to become data centre literate. That underlines the need for patient, long-term investment in skills—not just short-term fixes.
Looking further ahead, we must attract a younger, more diverse workforce into the sector. The current workforce is ageing, and the skills gap will only grow unless the industry can inspire the next generation.
Cybersecurity, too, is ultimately about people. Technical safeguards matter, but human error remains the most significant risk. Someone without a lanyard, someone clicking a phishing link—these are the likely weak points. Ongoing investment in education and awareness is critical.
Perception
Data centres affect everyone, whether they realise it or not. So, public perception matters.
There was a thought-provoking moment in one panel: Should we learn from the housebuilders and stop spending so much energy trying to win over those who will never support us?
One panellist revealed that its surveys show most people in the UK either have a neutral or favourable view of data centres once they understand them. Only around 7% expressed active opposition. And when they grasped the local economic benefits, support tended to grow.
Larger developers are generally doing a good job on community engagement—consulting on local priorities such as education, aesthetics, and biodiversity. But this is work that needs constant attention and reinvestment.
In places like Ireland, debates about power and water usage have raised genuine concerns. But context matters. For example, while data centres use 20% of metered electricity in Ireland, they account for just 3% in the UK. Furthermore, the UK’s climate and more efficient cooling systems mitigate many of the risks associated with water use. Better education and myth-busting are needed to shape a fairer social contract around digital infrastructure. And where people do have concerns, we all need to consider how our digital behaviours contribute to the load.
The Opportunity
So why take all this on?
Because these challenges are worth solving, yes, resources are finite, and the UK has multiple pressing infrastructure needs—from homes to heat networks, manufacturing to grid upgrades. But the digital economy touches all of them. And data centres are foundational to its growth.
We need a joined-up approach—one that doesn't treat digital infrastructure as a niche interest but recognises its role in supporting everything else.
The rewards are substantial. While some growth projections appear optimistic given current constraints, techUK estimates that—with the right conditions—the sector could grow from £4.7 billion in Gross Value Added (GVA) annually to £44 billion by 2035. That’s a tenfold increase, potentially creating over 40,000 jobs across the supply chain.
Outside the panels, it was a pleasure to meet so many people pushing this sector forward—from forward-thinking local leaders to the design consultant behind what two panellists agreed was a “beautiful data centre” (!) in London.
I left UKREiiF tired yet energised—for our team to play its part in helping the sector navigate the challenges and seize the opportunities ahead.
Find out more about our Data Centres and Digital Infrastructure team here.