Operations | Monitoring | ITSM | DevOps | Cloud

London still hosts the biggest concentration of UK data centre capacity, but the centre of gravity is starting to move. AI workloads are changing the infrastructure maths, pushing power, space and planning considerations up the decision list. That is exactly where regional locations start to look like the sensible option. Government data shows how concentrated the market remains: as of autumn 2024, London is estimated at 1,048MW of colocation IT load. Compare that with 44MW in the East of England, 17MW in the North East and 30MW in Scotland. The gap is huge, yet it is not a permanent advantage.

In recent days, the headlines of most technology titles have been dominated by the perfect storm that has led to a global shortage of Random Access Memory (RAM). As the short-term, temporary memory that handles data for processing and applications, RAM - and specifically Dynamic Random Access Memory (DRAM) - is a foundational business technology.

Cloud residency has moved from a technical preference to a board-level control question, as organisations are being asked to evidence who can access data, under which jurisdictions, and what happens when something goes wrong across borders. A Gartner survey of CIOs and IT leaders in Western Europe found that 61% expect geopolitical factors to increase their reliance on local or regional cloud providers, while also predicting that by 2030, more than 75% of enterprises outside the US will have a digital sovereignty strategy.

More UK organisations are treating cloud location as a governance risk decision, because incidents and audits expose questions around jurisdiction, access and evidence. Recent research found that 87% of respondents plan to partially or fully move workloads away from the public cloud over the next two years, with 54% considering private cloud, 38% exploring greater reliance on their own data centres, and 36% assessing colocation.

Imagine sending a letter to your neighbour across the street, only for it to be routed through London or even Amsterdam before landing in their letterbox. This is effectively what happens to much of Scotland's internet traffic. Despite physical proximity between users, businesses and services, digital data is frequently sent on needlessly long journeys, often leaving the country before reaching its destination. This approach is inefficient, costly and poses questions about privacy, resilience and digital sovereignty.

Organisations running AI workloads, like banks training fraud detection models, hospitals testing diagnostic tools, or manufacturers using predictive analytics, all face the same problem: hosting them is costly and resource-intensive. They require dedicated GPUs running non-stop, vast amounts of data moving in and out, and far more power and cooling than a typical IT system.