The Government can’t have AI dominance and net zero, it needs to choose - Chris Hocknell
Ofgem recently released a figure that should have stopped the country in its tracks. Around 140 proposed data centre projects in Britain are currently seeking grid connections. Their combined power demand: roughly 50 gigawatts. Britain's entire peak electricity consumption, every home, every factory, every hospital, on the coldest evening of the year, sits at around 45 gigawatts.
Look at those numbers again. We are planning to build data centre infrastructure that would, at full operation, consume more electricity than the entire country currently does. At the same time, we have committed to a clean power grid by 2030 and net zero by 2050.
When MPs asked Ed Miliband how data centres featured in the government's climate plans, he replied that future demand "remains inherently uncertain”. The government's own Seventh Carbon Budget contained no specific projections for data centres.
Data centres do not run on good intentions. They need continuous, reliable, always-on electricity, 24 hours a day, every single day. When the wind stops blowing or the sun goes down, a data centre cannot simply pause. It needs what engineers call firm baseload power: electricity that is there regardless of the weather. Renewables, however rapidly we deploy them, cannot provide this on their own.
And here is where the strategy falls apart. Britain's most recent nuclear project, Hinkley Point C, was approved in 2016. It will not generate electricity until 2030 at the earliest, at a cost now approaching £50bn because of government incompetence. Sizewell C, approved last year, will not be operational until the mid-2030s. The Rolls-Royce small modular reactors, heralded as a breakthrough solution, will not connect to the grid until the mid-2030s either.
The data centres the government is courting are being planned now, for this decade. The clean power to run them will not exist for another ten years. So what fills the gap? Gas. That is the only honest answer, which is precisely what blows the carbon budget the Energy Secretary has staked his political identity on.
The government announces AI investment strategies and net zero targets in the same breath, as though they are two legs of a coherent plan rather than two destinations with no road between them. Remember the plans for Britcoin? The promises of a fintech revolution? A shiny announcement, and very little thought about what happens next.
The reality is starker than the government wants to admit. With gas plants ageing off the grid and no replacements being built, supply is already shrinking. Add 50 gigawatts of new data centre demand on top, and the arithmetic does not work. The choice is not between two different visions of growth. It is between managed stagnation and genuine reckoning.
Britain's real advantage has nothing to do with cheap land or planning loopholes. It is Oxford, Cambridge, Imperial, and UCL; institutions that have produced a disproportionate share of the world's scientific breakthroughs and raised £2.6bn in spinout investment in 2024 alone. The government is not leveraging any of this; it is waxing lyrical about AI and data centres and calling it a strategy.
The problem is action; we need real, hard infrastructure made of concrete, steel and copper. Instead, we get investment summits, glossy action plans, and a Secretary of State who cannot tell MPs how much electrical power we will commission to ensure the UK actually has an AI strategy. The government has not lost the AI race because it hasn't even entered it.
Chris Hocknell is the founder and CEO of Eight Versa.