Datacentre gas plans expose grid connection strain

Ofgem has identified a large datacentre connection pipeline as electricity grid delays push some UK developers to examine gas-fired power arrangements.

More than 100 proposed UK datacentres are seeking gas connections, with requests understood to represent more than 15TWh of annual energy use. In some cases, gas is being considered beyond standby resilience, with onsite generation explored as a route around delayed electricity connections.

The scale of prospective demand is substantial. Ofgem has cited around 100GW of datacentre projects in the queue to connect to the National Grid, although only a proportion of that capacity is expected to reach operation. Even with attrition, the pipeline shows how artificial intelligence, cloud computing, and hyperscale digital infrastructure are creating concentrated electrical loads that can quickly exceed available network capacity in preferred locations.

Those loads are arriving as the electricity system is already absorbing new generation and demand from several directions. Offshore wind, solar, storage, EV charging, industrial electrification, and heat electrification all require grid capacity, while the transmission and distribution networks remain bound by long reinforcement cycles and equipment lead times.

Large datacentres create a particular planning challenge because they combine high demand, firm capacity expectations, commercial time pressure, and resilience requirements. A single site may need substantial import capacity, dedicated substations, upstream reinforcement, backup generation, power quality control, and carefully designed protection arrangements.

If grid access is delayed, gas-fired onsite generation can appear commercially attractive. It may reduce immediate dependence on constrained electricity networks, but it also risks placing major new digital loads outside the cleaner supply pathway being built through the power sector. That tension becomes sharper as the UK works toward Clean Power 2030 while also seeking to attract AI and cloud infrastructure investment.

National Grid’s £70bn network investment plan shows the scale of reinforcement now being prepared across the electricity system. Major network investment, however, remains a long-cycle engineering programme. New circuits, substations, transformers, control systems, and grid interfaces cannot be delivered on the same timetable as every commercial datacentre development.

The connection issue is not confined to generation. Demand-side connection reform is becoming just as important as renewable queue reform, particularly where projects are large, power-intensive, and geographically concentrated. Networks need to distinguish between credible demand, speculative applications, strategic load, and developments that could lock in higher-carbon power arrangements because electrical capacity is not available in time.

Operational tools can reduce some pressure on the system. The launch of MW Dispatch on the UK grid allows distribution-connected assets to help manage transmission constraints, showing how system operators are drawing more flexibility from resources below the transmission boundary. Such mechanisms improve utilisation of existing assets, but they cannot replace the firm import capacity required by large datacentres.

Planning around digital infrastructure is therefore moving into the core of power-system development. Datacentres are no longer a peripheral commercial property load; they are becoming a driver of regional reinforcement, generation strategy, gas connection activity, and grid access policy.

The engineering challenge is to connect economically important digital infrastructure without undermining electricity decarbonisation or overloading local networks. That will require clearer demand-side prioritisation, faster reinforcement where projects are strategically justified, and more disciplined treatment of speculative connection requests.