Dominic Jainy is a seasoned IT professional whose expertise sits at the intersection of artificial intelligence, machine learning, and infrastructure evolution. With the UK government setting a bold target of 6GW for AI-ready data center capacity by 2030, the industry faces a high-stakes race against time and technical obsolescence. In this conversation, we explore the logistical hurdles of the current 4.9GW projected pipeline, the shifting geographic focus toward northern power hubs, and the engineering challenges of retrofitting a legacy estate where 77% of sites currently operate below 10MW.
The target for 6GW of AI-ready capacity by 2030 appears ambitious since current approved projects only reach 4.9GW. What specific regulatory bottlenecks must be cleared to bridge this gap, and how can the three-year construction timeline be compressed to meet the 2030 deadline?
To bridge the 1.1GW gap between our current 4.9GW approved trajectory and the 6GW target, we have to address the sheer inertia in the planning system. Right now, there is a staggering 6.2GW of unapproved capacity sitting in the pipeline that is intended for completion by 2030, but it cannot move forward without a “rash” of rapid approvals. Developers are staring at a hard deadline; they must break ground by 2027 to accommodate a standard three-year construction phase, yet the lack of explicit labeling for data centers in planning applications often slows the process down. We need a more streamlined, “nationally significant” designation for these sites to bypass localized friction, especially since many of these projects are massive AI growth zones. Accelerating the timeline requires the government to treat these as essential infrastructure, much like energy projects, to ensure that the 18 critical sites that will eventually account for 68% of our capacity aren’t delayed by paperwork.
Many existing facilities are over a decade old and operate under 10MW capacity. Given that AI racks now require liquid cooling and massive power draws, what are the primary technical hurdles in retrofitting older sites, and when does it become more cost-effective to build entirely new high-capacity zones?
The technical hurdles for our legacy estate are profound, considering that 42% of our existing 600MW of data center stock is more than six years old, and 14% is over a decade old. These older sites were designed for a different era of computing, whereas modern AI racks are pushing toward a 100kW draw per rack today, with a roadmap to reach a massive 1MW per rack in just two or three years. Retrofitting becomes a losing game when you realize that as power and cooling needs spike, the physical footprint of the racks you can actually host decreases significantly. It becomes more cost-effective to build new when you consider that AI requires specialized liquid cooling systems that most 10MW legacy sites simply cannot support without a total skeletal overhaul. We are seeing a shift where 5.5GW of our future pipeline is concentrated in just a few “giant” projects because building from scratch allows for the massive power densities that 190 of our smaller, older sites just can’t match.
The geographic focus is shifting from the London-M4 corridor toward the M62 axis near Manchester and Leeds. What infrastructure advantages do these northern regions offer for high-capacity projects like the 1GW Elsham park, and how will this regional shift affect the national power grid’s stability?
The shift toward the M62 corridor is a strategic necessity because the London and M4 region is becoming incredibly crowded, currently holding 850MW of our 1.59GW installed capacity. The north offers the physical space and energy headroom required for “mega-projects” like the 1GW Elsham Tech Park in Lincolnshire and the 500MW site in Blyth. By 2034, the Hull-Manchester-North Wales axis is actually projected to overhaul the South East in terms of capacity, which helps decentralize the load on the national grid. While this shift reduces the “all eggs in one basket” risk for London’s power stability, it places immense pressure on northern grid nodes to support staged capacity additions. If managed correctly, these northern hubs provide a more balanced national energy draw, utilizing regions that historically had the industrial infrastructure to handle heavy electrical loads.
With over 6GW of the projected pipeline still awaiting planning consent, the window for starting construction by 2027 is closing. What strategies can developers use to secure local approvals for 500MW+ sites, and how do these mega-projects differ from traditional builds regarding resource and supply chain management?
Securing approval for a 500MW+ site like East Havering or Ravenscraig requires a narrative shift from “industrial warehouse” to “national economic engine.” Developers are increasingly aligning these projects with the government’s “AI growth zone” framework to provide a sense of urgency and prestige that local councils find harder to dismiss. Unlike traditional builds, these mega-projects require a specialized supply chain capable of delivering advanced GPUs and liquid cooling infrastructure at an unprecedented scale. The resource management is also far more complex; for instance, the Elsham project is being built in stages to reach its 1GW potential by 2037, requiring a long-term commitment to local labor and power infrastructure that smaller 10MW builds never had to worry about. Developers must now prove they have the energy agreements in place years before a single server is installed, as the sheer power draw of these sites can be three to five times higher than a standard industrial average.
Large-scale projects in Scotland, such as the 550MW Ravenscraig site, represent a significant portion of the pipeline but currently lack final planning approval. How can regional authorities better support these massive developments, and what are the specific economic risks if these northern hubs fail to materialize?
Scotland is in a precarious position because while it has projects like the 550MW Ravenscraig site in the pipeline, the vast majority of its planned megawatts currently lack official planning consent. Regional authorities can support these by moving them quickly through the public consultation phase and recognizing them as vital components of the UK’s broader AI strategy. If these northern hubs fail to materialize, the economic risk is a total loss of digital sovereignty and a failure to capture the high-value jobs associated with AI processing. Currently, Scotland and the North East only have about 50MW of installed capacity each; failing to approve the 550MW Ravenscraig site would mean missing out on a project that represents nearly 10% of the entire national target. Without these sites, the UK will remain tethered to an aging, underpowered southern infrastructure that cannot compete on the global stage.
What is your forecast for UK datacentre capacity?
My forecast is that the UK will narrowly miss the 6GW target by 2030, likely landing closer to 5.2GW or 5.5GW, as the bureaucratic lag for the 6.2GW of unapproved projects proves too difficult to overcome in a single three-year construction window. However, by the mid-2030s, we will see a massive surge as the M62 corridor and the Elsham 1GW project fully come online, potentially pushing us past 8GW of total capacity by 2037. The real story won’t just be the total megawatts, but the total replacement of our legacy 10MW sites with a handful of high-density, liquid-cooled “titans” that define the AI era. We are moving toward a bifurcated market where old sites either reinvent themselves as edge facilities or fade into obsolescence while the north becomes the new powerhouse of British computing.