Skip to main content

Archived Article — The Daily Perspective is no longer active. This article was published on 12 March 2026 and is preserved as part of the archive. Read the farewell | Browse archive

Politics

UK puts AI datacenters ahead of housing in power grid queue

Fast-track grid connections for tech infrastructure risks stalling home building amid soaring electricity demand

UK puts AI datacenters ahead of housing in power grid queue
Image: The Register
Key Points 3 min read
  • UK government proposes fast-tracking AI datacenter grid connections while housing and other projects wait
  • 140 datacenters seeking grid access want 50 GW of power—more than current peak British electricity demand
  • West London's grid already at full capacity, contributing to delays on residential projects since 2022
  • Home Builders Federation warns priority for datacenters could worsen UK's housing crisis

The British government's latest push to dominate artificial intelligence has created a problem that raises hard questions about where limited resources should go. AI datacentres could soon move to the front of the queue for electricity connections as the connection queue has expanded by 460 per cent in the first half of 2025, largely driven by applications from energy-hungry data centres.

The tension is stark. A single AI facility can consume as much energy as 100,000 homes, and a London Assembly report found that West London's local grid hit full capacity in 2022, stalling planned housing developments. This is not abstract. London Assembly members warned in December that housing developments in west London had already been temporarily delayed after the local grid reached capacity.

people queueing
Electricity grid connections are increasingly scarce as AI infrastructure competes with housing for limited capacity.

According to The Register's reporting, about 140 datacenters are in the queue to be connected to Britain's power grid, and their combined energy requirements are estimated to be more than the current peak electricity use for the entire country. Put another way, if every one of those schemes were built and operated at full load, data centres alone could roughly double Britain's peak electricity requirement.

The government's solution involves two elements. First, the government wants to tackle speculative applications by increasing the financial requirements for developers in the queue, including deposits or fees payable if key milestones are not met, a measure first raised in a report by Uptime Institute. This part addresses a real problem. Many datacenter operators secure multiple land parcel rights and request a grid load connection for each, then see what gets approved, which the capital investment to take seriously is clearly untenable and a huge financial risk for the energy sector if the demand doesn't fully materialize.

But the second element troubles developers and housing advocates. The government aims to publish a list of "strategically important projects" including AI Growth Zones, which will be "at the front of the queue" as capacity is freed up or created. In addition to priority grid access, the AI Growth Zone reforms mean datacenters in some locations may get discounts on their electricity bills.

The Home Builders Federation sees this as a failure of basic policy balance. A spokesperson for the Home Builders Federation told The Register that the energy industry should be encouraged to invest in infrastructure capable of supporting both new housing and economic growth, saying "As we continue to face a housing crisis with the social and economic issues that it brings, it is frustrating that regulatory, planning, and policy arrangements effectively prioritize energy-intensive datacenters over energy-efficient homes for families."

The government's argument carries weight. AI will be the most significant driver of datacentre electricity demand, with electricity demand from AI-optimised data centres projected to more than quadruple by 2030. Britain's strategy to lead the global AI race depends on infrastructure. The proposed capacity is around five times higher than what had previously been assumed in Government planning tied to the UK's 2030 clean power target.

Yet the UK's situation offers a cautionary lesson that other nations, including Australia, are watching closely. Australia's data centre energy capacity could grow at least four times current levels by 2035, according to recent research by Baringa, and analysis shows the country will need to bring additional renewable and storage capacity online to meet this rising demand. Australia has more time to plan. Britain does not.

The broader issue cuts to institutional accountability. The document claims these issues are exacerbated because there are no mechanisms to prioritize strategically important demand projects, which in practice means datacenters. This is less about whether datacenters deserve priority and more about whether the prioritisation process itself is sound. Was the connection queue system designed with planning flexibility built in, or has it been forced to adapt to unprecedented demand?

The government's consultation document, "Accelerating electricity network connections for strategic demand," is available online now with views welcomed from the wider public as well as industry, with the deadline for submitting responses at midnight on April 15.

This reflects a genuine policy trade-off. Economic growth and housing supply are both legitimate public goods. Grid capacity is finite in the near term. The question is not which matters more in principle, but how the constraint should be managed fairly and transparently. Whether fast-tracking datacenters while housing waits achieves that balance is what the public consultation should determine.

Sources (7)
Grace Okonkwo
Grace Okonkwo

Grace Okonkwo is an AI editorial persona created by The Daily Perspective. Covering the Australian education system with a community-focused perspective, championing evidence-based policy. As an AI persona, articles are generated using artificial intelligence with editorial quality controls.