Microsoft has abandoned plans for a major UK artificial intelligence data center after electricity grid officials said power wouldn't be available until the 2030s, according to a report from Channel Insider. The tech giant ended its contract with Nscale for a planned facility in Loughton, Essex, that was slated to begin operations in early 2027. The cancellation highlights a growing infrastructure bottleneck across the UK and broader EMEA region, where data center ambitions are colliding with the reality of grid capacity constraints.
The Loughton campus was designed to house 23,040 Nvidia GB300 GPUs across a 50MW site supporting Microsoft Azure services in Britain, the report notes. Nscale is still pursuing the £2 billion development and exploring alternative power sources while its Microsoft projects in Norway, Portugal, and the United States continue unaffected. The Essex facility is expected to generate 250 permanent positions if it ultimately launches. Grid connection bottlenecks are pushing some UK data center projects back by as long as 13 years, creating a major roadblock to the country's AI infrastructure expansion efforts. The site could eventually grow from 50MW to 90MW in total capacity.
Microsoft said the decision affects only one location and that its broader partnership with Nscale remains in place, according to the report. Nscale reported $103.4 billion in active and contracted total contract value as of August 31, though just $2.6 billion was classified as active. The company's wider Microsoft agreements could produce up to $43.8 billion through 2033, subject to delivery and service-availability requirements. "Capacity that isn't energized generates no revenue, but the costs keep running," Alex Cordovil, research director at Dell'Oro Group, told Data Center Knowledge.
The Loughton setback reveals how grid availability is shaping where AI infrastructure actually gets built, alongside factors like connectivity, customer demand, and hardware access. Neocloud providers building GPU-intensive facilities around a handful of major customers face particular exposure to delays because they typically operate with concentrated customer bases, elevated capital costs, and equipment schedules tied to specific rollouts. A site that opens years behind schedule has less time to generate returns on costly accelerators before newer chip generations alter pricing and performance benchmarks. For partners selling cloud, data center, and AI infrastructure, the central question is whether new capacity can actually be powered on the timelines customers anticipate.
The report recommends UK MSPs, resellers, cloud partners, and infrastructure providers should verify when facilities are expected to receive grid capacity and whether alternative generation is realistic, as multi-year connection delays can push customer deployments back even when hardware and contracts are already secured. Partners should also determine whether providers can shift workloads or future builds to other UK or EMEA locations when one market can't supply power on schedule, and review delivery milestones, termination rights, and when accelerators will actually be installed, particularly when financing depends on specific customers or deployment dates. For partners serving customers across the UK and Europe, the key issue is where usable AI capacity will actually come online—GPU counts and signed contracts matter, but grid access, site readiness, and backup locations may ultimately determine which providers can deliver when customers need the compute. The infrastructure race is no longer just about securing chips and signing customers; it's about whether the grid can keep pace with ambition, and whether partners can distinguish between capacity that's promised and capacity that's powered.