Summary
- DESNZ says data centres behind existing transmission constraints can make use of renewable generation that would otherwise be curtailed.
- The government has consulted on aligning large data-centre connections with regional infrastructure targets.
- Future data-centre geography is becoming more closely tied to grid conditions, rather than connection availability alone.
The Department for Energy Security and Net Zero has said data centres could reduce electricity-system constraint costs by locating in areas where renewable generation is being limited by transmission bottlenecks.
Responding to a parliamentary question on 7 September, the government said data centres can lower constraint costs when they locate behind existing transmission constraints and consume renewable electricity that would otherwise be curtailed.
DESNZ linked that position directly to recent proposals to align data-centre connections with regional infrastructure targets.
The answer makes location an increasingly explicit part of UK data-centre power policy. The question is no longer only whether a facility can obtain a grid connection, but whether placing that load in a particular part of the network helps or worsens the way the wider electricity system operates.
Constraint costs arise when the transmission system cannot carry all the electricity available in one region to consumers elsewhere.
Grid operators may have to reduce output from generators on one side of the bottleneck and increase generation elsewhere to maintain balance. The consumer ultimately bears the cost of managing those constraints through the electricity system.
A large new load positioned on the generation side of the bottleneck can, in principle, consume some of that electricity nearer to where it is produced.
Data centres are unusual in that calculation because individual campuses can create loads measured in hundreds of megawatts and may consume electricity continuously.
That does not mean any facility built close to renewable generation automatically removes network constraints. The value depends on the precise location, the timing and profile of demand, future transmission reinforcement, regional generation patterns, and whether the data centre itself triggers additional network work.
A new constraint can also emerge elsewhere as the electricity system changes.
Nevertheless, the government is increasingly treating large digital loads as a system-planning variable rather than simply another connection request.
The policy could influence where future UK data-centre capacity is built. Established clusters offer fibre, customers, contractors, and mature supplier ecosystems, but some also face intense competition for grid capacity.
Regions with large volumes of renewable generation may have a stronger electricity-system case for new demand, while developers still have to solve fibre, planning, construction supply chains, workforce availability, cooling, and customer requirements.
The balance may differ according to workload. Compute that is relatively insensitive to latency can be easier to locate according to electricity availability than services requiring close physical proximity to end users or other networks.
Connection policy therefore has the potential to influence not only which sites are developed but which forms of compute are economically suited to different UK regions.
There is also a fairness problem for government and grid planners. Prioritising data centres in strategically useful locations would require a way to distinguish credible developments from speculative capacity requests while balancing demand from industry, housing, transport electrification, and other large users.
Regional targets may help direct loads towards parts of the network where they impose lower system costs, but they do not create electricity or transmission capacity by themselves.
Developers will still need buildable connections, and the power system will still require reinforcement as overall demand increases.
DESNZ’s answer nevertheless establishes the policy direction clearly. The government sees at least some data-centre demand as potentially useful when it is positioned behind the right network constraint.
If that principle is carried into connection reform, data-centre site selection will become still more closely tied to transmission geography, renewable curtailment, and regional network planning alongside the traditional requirements for land, fibre, cooling, and planning consent.
