Some research has claimed that waste heat from data centres could warm more than 3.5 million UK homes. It couldn’t. And it won’t.
Throughout my career, the idea that waste heat is a vast untapped opportunity has resurfaced again and again: mine water, nuclear plants, factories, the Tube, sewers — and now data centres. The story is always the same: we just need to capture the heat and pipe it to homes cheaply. It almost never happens, and for good reasons.
Waste heat isn’t something people want to produce
No-one sets out to generate waste heat. Any organisation producing significant amounts of it is already trying to reduce it, because it represents a cost. Many have succeeded: factories and hotels reclaim heat, data centres can operate at higher temperatures reducing cooling, and so on. But the bigger issue isn’t just efficiency — it’s uncertainty.
Projects like these require 20–25 year contracts to be viable. Yet very few organisations can commit to anything like that. I’ve worked on schemes involving factories and oil refineries; neither could offer long-term guarantees because they simply don’t know what they’ll be doing in a decade, let alone a quarter-century. Even sewer-heat schemes — one of the most stable sources — have gone bankrupt.
Data centres are even more uncertain. They sit in the fast-moving world of AI and IT, where change is constant and unpredictable.
Data centres will evolve
Over the next decade, data centres may:
- operate at higher temperatures, reducing the need for heat rejection
- use far more efficient chips, cutting heat output
- go bankrupt
- relocate
- fragment into highly distributed micro data centres
Across my career, I’ve watched technologies and markets shift in ways no-one foresaw. Large data centres are a major issue today, but the sector will evolve — and we cannot predict how. This is precisely why securing a 20–25 year heat-supply contract is so difficult.
Even if you could get a contract…
The heat supply would be rigid. Waste heat is only available when the source is producing it. For data centres, that means the supply is highest in summer — exactly when heat demand is lowest. You’d still need a full backup system, meaning just as much plant (heat pumps) as you would without the waste heat. Plus all the pipes, pumps, heat exchangers and infrastructure on top.
And it’s not just the supplier who must commit. The heat taker must also sign long-term guarantees. What happens if the development is delayed? Or the property mix changes? What if residents use less heat than expected? What if they install solar and want to use it to generate hot water?
A bigger priority: flexible energy use and microgrids
My own view is that the real opportunity with data centres lies in flexible energy demand. One analysis estimated 40% of data-centre processing could be shifted in time or moved to another site. Combine that with a microgrid linking buildings, generators and loads on-site, and you could significantly cut bills, reduce grid-connection requirements, and integrate more renewable energy cost-effectively.
This is where we should be working with the data-centre sector — not chasing marginal, unreliable heat sources.
Waste heat has its place — but not a big one
I’m all for using waste heat where the case stands on its own merits. We must be realistic, however: it will not be a major component of the UK’s heat system. And we absolutely should not put taxpayers on the hook for infrastructure that may never deliver the promised benefits. Let the market determine what’s viable.
Leave a Reply