Dominic Jainy stands at the forefront of a digital revolution that is physically reshaping the landscape of the United Kingdom. As a veteran IT professional with a deep focus on artificial intelligence, machine learning, and the decentralized potential of blockchain, he has spent years navigating the intersection of heavy infrastructure and high-level compute. With the UK government’s 2024 decision to classify data centers as critical national infrastructure, the sector has moved from the periphery of industrial planning to the very heart of national security and economic strategy. Dominic’s insights provide a rare bridge between the technical requirements of modern AI and the grit of large-scale engineering required to power it.
The following discussion explores the phenomenal growth of the British data center pipeline, which has recently surged by nearly 60% in a mere six-month window. We examine the shifting geography of these “digital factories,” moving from the traditional, dense clusters of the M4 corridor to the sprawling, high-capacity brownfield sites of Northern England and Scotland. The conversation also tackles the daunting logistical challenges of grid connectivity and the innovative, yet complex, timeline of integrating small modular nuclear reactors into the energy mix. Through this lens, we uncover the tension between the immediate hunger for AI processing power and the long-term reality of a power grid pushed to its absolute limit.
With the current pipeline reaching over 14GW across 173 projects, how do you perceive the sheer speed of this expansion compared to the digital landscape we navigated just a few years ago?
The acceleration we are witnessing right now is nothing short of breathtaking, moving at a pace that feels more like a frantic gold rush than a standard industrial build-out. When you look at the data, we’ve seen the pipeline grow from about 10GW in March to over 14GW by September of this year, which represents a massive influx of 4.5GW of capacity in just a six-month window. This isn’t just a marginal increase; it is a 60% jump in the number of projects entering the planning system, driven almost entirely by the insatiable demand for AI compute. You can feel the intensity in the market as developers race to secure planning consent and, perhaps more importantly, the physical grid connections required to keep these facilities breathing. It is a high-stakes game where being classified as critical national infrastructure has given the industry a new level of urgency, yet the physical reality of laying cables and building substations remains a grueling, slow-motion challenge against the clock.
The regional data shows a fascinating split where the South East leads in project volume, while the North East and Scotland are winning on the scale of individual sites. What is drawing these massive hyperscale developments away from the traditional London hubs?
While London and the M4 corridor still hold the crown for volume with 82 projects and 5.2GW of capacity, the real story is the sheer physical footprint of the new developments in the North. In the North East, the average scheme size has reached a staggering 228MW per project, which is nearly triple the 77MW average we see in the more congested South. This shift is driven by the hunt for massive brownfield sites—think former power stations and heavy-industrial land—where the skeletal remains of old industry provide the perfect foundation for high-capacity transmission infrastructure. In Scotland, we’ve seen a threefold increase in the committed pipeline in just half a year, jumping from 0.9GW to 2.4GW across 14 projects. There is a certain poetic irony in seeing the same sites that once hummed with coal-fired turbines now being repurposed to house the silicon brains of the AI era, utilizing those existing high-voltage connections that are simply too difficult to find in the crowded Home Counties.
The data highlights a “pre-planning” tail of 4.4GW that hasn’t even entered the formal system yet. How does this looming wave of projects complicate the already strained relationship between developers and the national grid?
This 4.4GW “queue” is essentially the next wave of the storm, consisting of 24 massive schemes like the 1GW campus at the former Cottam power station and the 600MW London Data Freeport. When you add this to the 14.6GW already in the pipeline, you’re looking at close to 19GW of capacity all knocking on the same door, asking for power from a grid that has a very finite supply. It creates a palpable tension in the industry because many of these projects are competing for the same limited transmission nodes, making the “first-come, first-served” nature of grid applications a life-or-death struggle for multi-billion dollar investments. We are seeing projects move through the “nationally significant infrastructure project” route just to bypass traditional bottlenecks, but even with political support, the physical reality of the copper and transformers in the ground doesn’t change overnight. The grid is no longer just a utility; it is the ultimate gatekeeper, and for the communities hosting these sites, the sight of massive new substations is becoming as common as the data centers themselves.
The Cottam project in Nottinghamshire is being billed as a potential landmark for small modular reactors, but the servers are expected to arrive long before the nuclear power does. How should we view the “nuclear-powered” label in the context of current energy shortages?
The Cottam project is a brilliant vision of the future, involving a collaboration between Holtec, EDF, and Tritax to utilize SMR-300 reactors, but the timeline reveals a significant “power gap” that we have to be honest about. Even though the design completed its Generic Design Assessment in March 2026, the licensing and construction phases mean we won’t see actual electrons flowing from those reactors until 2030 at the absolute earliest. In the interim, those data centers will still be built and will still need to draw their massive loads from the existing grid, which effectively adds another huge competitor to the current power pool. It’s a sensory experience of two different industrial speeds: the rapid, agile deployment of server racks versus the decade-long, methodical march of nuclear engineering. While the promise of “green” nuclear power is the headline, the immediate reality for the next few years is another large campus putting pressure on our traditional infrastructure while the SMRs remain a blueprint on the horizon.
With the methodology behind these figures becoming more sophisticated—accounting for floor area and statistical modeling—how confident can we be that these 14GW and 19GW figures represent the true ceiling of UK capacity?
Actually, I would argue that these numbers are a conservative lower bound because our modeling often excludes the smallest edge sites or specialized healthcare schemes that don’t publish their capacity data. The current methodology, which uses an adjusted R-squared of 0.74 to estimate capacity from floor area, is much more robust than what we had in the past, but it still can’t account for the rapid density increases we see within existing footprints. As AI chips become hotter and more power-hungry, the wattage per square foot is skyrocketing, meaning a building that was planned for 20MW three years ago might be pushing for 40MW today. We are tracking 173 committed projects, but the “pre-planning” list is shifting into the active pipeline so quickly—with sites like Cato and Spango Valley moving into planning just weeks after our last data extract—that the ceiling keeps moving higher. The 19GW total is a massive target, but given the current trajectory, it may well be the baseline for the next phase of growth rather than the peak.
What is your forecast for the UK’s data infrastructure over the next few years?
I anticipate a period of intense “infrastructure Darwinism,” where only the projects with the most creative power solutions will actually reach completion. We are going to see a significant divergence between the “announced” capacity and the “energized” capacity, as the grid bottleneck forces developers to either invest heavily in their own onsite generation—like the SMRs at Cottam or massive battery arrays—or face years of stagnation. By 2028 and 2029, the geographic shift will be complete; the North and Scotland will be the undisputed heavyweights of UK compute, hosting massive 500MW+ campuses that dwarf anything currently standing in the South East. However, the success of this 19GW pipeline depends entirely on whether the government can accelerate grid investment to match the speed of the private sector, otherwise, we risk having the world’s most advanced digital brain with no way to turn the lights on.