Labour’s eagerness to welcome investment from US technology firms may bring billions of pounds to the UK economy, but experts warn it risks ignoring the environmental, social and long-term consequences of handing critical infrastructure to a handful of powerful American companies.
Nvidia’s blunt advice
Last week, as ministers signed a flagship tech pact with Washington, Jensen Huang, chief executive of chipmaker Nvidia, offered an unusually candid piece of advice: burn more gas.
“I’ve every confidence the UK will realise it takes energy to grow new industries,” he said, pointing to nuclear, wind and solar as part of the mix – but urging that gas turbines remain in play.
That suggestion runs counter to the government’s pledge to phase out fossil fuels in its drive to net zero. Yet the deal Labour struck with the US is expected to unleash a wave of new energy-hungry data centres across Britain. The official announcement mentioned “datacentre” no fewer than 14 times, underlining their centrality to the agreement.
A global AI gold rush
Demand for these vast digital warehouses has soared with the rise of generative artificial intelligence, which requires enormous computing power to train and operate.
OpenAI’s Sam Altman recently joked on a podcast that “a lot of the world gets covered in datacentres over time” before half-seriously suggesting: “Maybe we put them in space.”
Labour’s willingness to roll out the red carpet for American firms is understandable: the government urgently needs growth and is pinning hopes on AI to boost productivity. But critics warn there has been little public debate about the potential costs.
Nick Clegg, the former deputy prime minister turned Meta executive, described the deal as “just another version of the United Kingdom holding on to Uncle Sam’s coat-tails”.
Mounting environmental concerns
Researchers are already sounding the alarm. A Massachusetts Institute of Technology (MIT) study warned last year that the “unfettered growth” of generative AI could trigger “significant costs” to the environment unless a more cautious, managed approach replaces the current “Klondike gold rush”.
The International Energy Agency estimates global data centre electricity consumption will more than double between 2024 and 2030, reaching 945 terawatt hours – equal to Japan’s total current demand.
As AI models grow more sophisticated, their energy requirements rise dramatically. Some consume thousands of times more power than smaller predecessors, with academics warning the sector’s electricity use could increase 24-fold.
Labour’s agreement with the US also included a nuclear cooperation plan, with hopes of building small modular reactors to supply datacentres’ voracious needs. But these will take years to develop, leaving the UK grid to absorb the immediate strain.
Water use is another pressing issue. Datacentres require vast quantities for cooling. The Environment Agency, already forecasting water shortages for homes and farming, has admitted the AI boom makes future demand difficult to predict.
Google research has shown that fulfilling a single prompt in its Gemini chatbot consumes the equivalent of five drops of water and energy comparable to watching nine seconds of television. Multiply that by billions of prompts each day, and the costs mount quickly.
Social trade-offs
Beyond energy and environment, there are social questions. A Social Market Foundation report last week warned that excessive reliance on chatbots in schools could erode pupils’ cognitive skills, urging ministers to update classroom guidelines.
Even more troubling, US parents recently alleged their son took his life after “months of encouragement” from ChatGPT, forcing OpenAI to promise tighter safeguards. The case has amplified concerns about AI’s role in mental health and the risk of a flood of low-quality machine-generated content distorting public debate.
Balancing growth and risk
None of this negates AI’s transformative potential. Its applications in drug discovery, medical diagnostics, financial modelling and software development could deliver enormous economic and social benefits.
But Labour’s strategy appears to rest on a largely unquestioning embrace of Silicon Valley’s biggest players. Critics argue the UK risks becoming dependent on firms with little incentive to stay if profits or politics shift.
Research for the Department for Energy Security and Net Zero attempted to reassure by comparing datacentre power use to the energy that would otherwise be consumed by human labour – likening machine translation to human translators, for instance. Yet it remains unclear whether AI is replacing demand elsewhere or simply adding to it.
With the electrification of transport and heating also ramping up demand, the competition for scarce energy is intensifying. How Britain balances these priorities – and who ultimately bears the costs – remains unresolved.
A call for clear-eyed debate
Labour hopes the tech pact will underpin a new era of growth and innovation. But voices across academia, industry and politics are urging a more open conversation about trade-offs.
As MIT researchers cautioned, AI’s rapid advance “requires genuine reasoning” not just about technical capability but about long-term sustainability.
Without a more deliberate approach, the government risks tying Britain’s future to technologies that are not only dominated by a small circle of US giants but also carry hidden costs for the environment, public services and society at large.
For now, Labour’s bet on AI remains high stakes – promising growth on one hand, but demanding urgent answers to questions of energy, resilience and responsibility on the other.
