3.2 GW. One site. Roughly a tenth of Britain’s average electricity demand.
National Grid’s proposed Teesside substation could become the UK’s largest, supporting what may be Europe’s biggest data centre complex.
It’s welcome news, and a reminder that Britain’s AI ambitions will only go as far as its grid can take them. Data centres can be built in a few years. Major transmission upgrades can take a decade or more. Access to power, not just to computing hardware, is becoming the binding constraint.
Yet building more infrastructure cannot be the only answer, for three reasons:
Time. Every gigawatt waiting for conventional reinforcement is investment, jobs and capability delayed.
Cost. Without smarter integration, very large loads could push up network costs for households and businesses. The UK’s new data centre policy statement rightly puts billpayer protection among its priorities.
Stability. Data centres are not passive loads. In 2024, a single transmission fault in the US triggered the simultaneous disconnection of approximately 1,500 MW of data centre load. As these facilities grow, their collective response to grid disturbances is becoming a significant reliability concern
So what if we changed the relationship between data centres and the grid?
A data centre that is observable, controllable and responsive to network conditions can become an active participant in grid operation. With the right intelligent coordination, it can:
Support voltage and reactive power
Improve power quality and ride through disturbances
Provide flexibility when the system needs it
Give network operators the predictability to manage capacity better
That is the shift from grid-dependent consumer to grid-positive contributor, with benefits across the system:
Consumers: better use of existing infrastructure and less pressure on future network costs
Network operators: a load they can see and rely on, rather than one they must over-build around
The economy: faster access to power for the AI investment Britain is competing to attract
Regulators internationally are already recognising that large loads must become more predictable and responsive. The UK has the opportunity to lead.
At ENODA, we’re building the technology that control how electricity flows at the point of connection, and the grid intelligence that gives operators the visibility, forecasting and coordination to rely on it.
We will still need more substations, cables and transmission capacity. But the future of AI infrastructure cannot depend solely on a bigger grid. It must also depend on a smarter one.
The question is no longer just how much power a data centre needs. It’s what that data centre can do for the grid in return.