Britain’s debate about data centers is entering a new phase.

For the past few years, the central question for government has been how quickly we can build the computing infrastructure needed for the AI economy. That remains important. Britain will need considerably more computing capacity as AI becomes embedded across businesses, public services and everyday life.

The debate is now broadening. Communities are increasingly asking what large data centers mean for local power supplies, water, land and the environment. At the same time, the new Prime Minister, Andy Burnham, has made clear that he wants economic growth to deliver tangible benefits for people and places.

This gives the government an opportunity to ask a bigger question about Britain’s AI infrastructure: what kind of data centers do we want to build?

A broad contribution

Data centers can make a much broader contribution than simply providing computing capacity. The most efficient facilities can operate with very low water use, make better use of electricity and capture the heat they generate for surrounding homes and public buildings.

Our policy framework needs to recognize these differences. The previous government’s AI Growth Zone program was designed to accelerate investment by tackling barriers including grid connections and planning. There is a strong case for doing that. Britain cannot become an AI-ready economy if it takes years to connect the infrastructure on which AI depends.

Faster development can go hand in hand with higher expectations. This is where the Burnham government has an opportunity.

First, community benefit should become part of the test for major new data center developments. At Deep Green’s new facility in Urmston, Manchester, heat generated by servers is being captured and used to heat the swimming pool at the neighboring leisure center. This is expected to save the center around £80,000 ($107,000) a year in energy costs while cutting its carbon emissions by approximately 150 tonnes a year.

In Bradford, where we have secured planning permission for another facility, the intention is to provide the heat generated to 1Energy’s city-wide heat network, which will warm buildings including the Grade I listed City Hall.

This principle can be applied much more widely. Data centers generate heat continuously and predictably. Where there is viable local demand for that heat, it can become a valuable resource for the surrounding community.

The government should require major new data centers to demonstrate credible heat-recovery capability and consider proximity to potential heat users when sites are selected.

Creating wider value

Second, the planning and grid systems should recognize the wider value different data centers can create. Grid capacity is a valuable national resource and there is an opportunity to direct it towards projects that deliver the greatest overall benefit.

Projects receiving faster planning or grid connections should demonstrate what they will contribute in return. That could mean heat reuse, very low water consumption, exceptional energy efficiency or measurable benefits to the wider electricity system and local community. The principle is straightforward. We should prioritize projects that create the greatest value from the electricity and grid capacity they use.

Third, government should use its own purchasing power. The state is going to become a major customer for computing capacity as AI is adopted across public services. Procurement gives ministers a powerful opportunity to shape the market and encourage higher standards.

Government contracts should increasingly favor UK-based computing infrastructure that meets demanding standards on efficiency, resilience and heat reuse. As demand for AI and high-performance computing grows, Britain also needs to ensure it retains sovereign compute capability, with critical data and computing capacity hosted securely in the UK rather than relying disproportionately on overseas infrastructure. Public procurement can help create a market for the kind of infrastructure Britain wants to build.

There are useful precedents elsewhere. Germany already places requirements on new data centers around energy efficiency and waste heat. Denmark and Norway require qualifying facilities to assess opportunities to reuse surplus heat. Britain can learn from the principle underpinning these approaches: heat generated by data centers is a valuable resource that can be put to productive use.

Growth in every postcode

Public interest in the impact of data centers is growing. That gives our industry an opportunity to show communities what this infrastructure can contribute to the places where it is built.

AI infrastructure can help heat swimming pools, homes, hospitals and civic buildings. It can reduce heating costs for public services. Distributed computing capacity can support British businesses and institutions and strengthen the country’s digital resilience.

Burnham has talked about pursuing growth in every postcode, economic development that people can see improving their communities. Britain’s next generation of AI infrastructure is an opportunity to put that principle into practice. The government can set ambitious standards and make it easier and faster for projects that meet them to get built. It is time to build an AI economy in which digital infrastructure strengthens the communities around it. That should be our ambition for the next generation of data centers.