Cloud Week Energy Day
A year of change in the energy sector
Every time we use a cloud service, whether checking a bank balance, accessing NHS records, streaming a film or asking an AI chatbot a question, the work happens in a data centre. Data centres are the physical home of the cloud, and they need a large, reliable and affordable supply of electricity to run. As cloud adoption and AI drive demand for computing power, the relationship between data centres and energy is becoming more important, and more complex. Put simply, the future of cloud in the UK depends on how we power the data centres that host it.
That was the question at the heart of our Powering Digital Infrastructure report, published in February 2026. It identified two main barriers to powering data centres, and therefore to growing the cloud: high energy prices and grid constraints. Nine months on, both remain, but much around them has changed. Here is where things stand.
Energy prices
For cloud providers, electricity is one of the largest operating costs, and it feeds directly into what businesses and public services pay for cloud. Our report warned that UK industrial electricity prices, the highest in the IEA, were deterring data centre investment and slowing wider electrification. Prices have still not fallen, and the energy crisis following the closure of the Strait of Hormuz has made things worse.
Our report argued that moving policy costs (levies) off electricity bills and into general taxation is the quickest way to cut prices, faster than decoupling gas and electricity prices, and entirely within the Government’s gift. Over the summer, techUK campaigned on this with other trade bodies and called on the Treasury Committee to hold an inquiry into levy reform. Yet, as last year, the 2026 Autumn Budget largely left businesses out. The one bright spot is a Cost Control Framework, due at the 2027 Spending Review, which could set a path for moving levies off bills.
The debate has also widened beyond what data centres pay for electricity, to what they might mean for everyone else’s bills. There are legitimate questions about whether their demand will require new generation and grid capacity, which could raise network charges for consumers. But techUK has continued to point out that demand is good for the system. Data centres channel funding into new renewables through corporate power purchase agreements, and they spread network costs and levies across more users, lowering the average unit price for all. Boston Consulting Group estimates that connecting 14GW of the data centre pipeline would cut £35 from every household’s annual bill by 2035.
Grid connections
The cloud can only grow as fast as new data centres can connect to the grid, and this is where the most change has happened. As we began writing our report in December 2025, grid applications were moving from “first come, first served” to a system based on whether projects are both ready and needed. techUK has long highlighted the problems of an over-subscribed queue, with some projects facing energisation dates 8–10 years away. NESO went on to show the scale of the issue, identifying 73GW of data centre demand in the queue. That is troubling not only because it far exceeds the UK’s peak demand of 47.4GW, but because industry estimates only 8-15GW of it is viable.
To tackle this, Ofgem’s reforms have focused on data centres and grown in substance and complexity. They fall under three pillars: Curate (removing speculation from the queue), Plan (deciding which projects are needed) and Connect (new ways of connecting to the grid).
On Curate, Ofgem consulted this summer on a refundable data centre commitment fee to enter the queue, alongside strict progress milestones. We were glad to see our report reflected: we argued a fee was the best way to remove speculation and set out what robust entry requirements should look like. However, we were extremely surprised by the level proposed, £237,500 to £712,500 per MW for projects above 40MW, far above our suggestion of £2,500 to £20,000. We have argued that a fee that high risks deterring viable projects and investment.
On Connect, self-build has gained momentum. Under self-build, developers construct their own high-voltage connection and either operate it or transfer it to the network owner. This takes pressure off the network, speeds up time to power, and means consumer bills are not footing the cost of the build. Flexibility has also moved up the agenda, with the Government setting up a Flexible Tech Taskforce, due to growing interest in whether data centres can adjust their energy use to support the grid. As our report noted, this is difficult for most UK data centres, which run the live, time-critical services our society and economy rely on. We have instead explored alternatives, such as battery storage at congested parts of the network and virtual power portfolios. We now need NESO to help unlock these at scale, and we await Ofgem’s consultation on flexibility, including alternative connection agreements.
Finally, on Plan, our report called on Government to set out how its Connections Accelerator Service (CAS) would work, and warned against a two-tier system in which only some data centres get support. Post the Autumn Budget, we now know the CAS will be the main route for prioritising strategic projects, giving faster access to the grid to those that meet a new Data Centre Covenant on sustainability, transparency and investment. techUK will follow this closely and push for a definition of “strategic” that reflects the whole sector.
Taken together, these changes will shape how quickly, and at what cost, the UK’s cloud capacity can grow.
