AI is rewriting the rules of global competition, but no matter who gets a head start, the race will ultimately be won by whoever can power it cheaper.
As the UK lines itself up to become a global AI hub – on the heels of the UK-US Tech Prosperity Deal and Nvidia’s £2 billion investment into the UK AI ecosystem – the conversation keeps circling the same issue: chips, talent, and data.
But the biggest constraint isn’t digital, it’s electrical.
CEO and co-founder of tem.
Global data-centre electricity demand will rise 165% by the end of the decade, according to Goldman Sachs. That’s billions of pounds of infrastructure and terawatts of power.
No matter where you look, energy is becoming the new currency of AI, and right now, most energy systems around the world are nowhere near ready for what’s coming.
The next industrial revolution is an energy revolution
Training models, running inference at scale, cooling, redundancy – it all demands vast electricity. Training one GPT-scale model can use as much power as 100,000 homes in a year, making data centers no longer a niche infrastructure but a true national interest.
Reports in the US show that communities near major data hubs have seen electricity prices rise by up to 267% in the past five years, especially in Virginia and Texas. It’s a vicious cycle where more and more subsidies will have to flow through to the 99% of customers to pay for AI hyperscalers to grow.
The UK cannot afford that cost to hit the bottom line of small business. Equally, hyperscalers shouldn’t be penalized for producing what the world demands of them. We’re at a genuine stalemate.
We either allow AI’s exceptional growth to stack on even more structural inefficiency and risk higher costs for every business and household, or we lose the global competitive edge we in the tech sector are working so hard to build.
Our next move has to be price.
I truly believe the new world order will be defined by access to low-cost electricity, and the answer doesn’t simply sit in cheaper production – we need a radical rebuild of the transaction itself that determines the price for all.
A wholesale market built in the 1980s
Even before AI arrived, Britain had some of Europe’s highest electricity prices and the highest industrial electricity prices in the world, despite us producing renewable generation at a rapid pace.
That’s down to the market within which we operate – a wholesale electricity market stacked with inefficiencies, outdated trading logic not fit for the new world of supply and hundreds of intermediaries that drive up the cost. Not to mention that the cost of UK-generated renewable energy still ties back to the global price of gas.
Today, a gas price spike in another continent can bankrupt a business in Britain, even if the physical power is coming from local sources.
That’s the flaw at the heart of our market design. It’s like trying to run AI applications on dial-up infrastructure, or watching Stripe and PayPal build modern fintech layers while Mastercard clings to their 1960s rails.
We’re still transacting electricity using last century’s logic and it’s holding the economy – and the business owners who form its backbone – hostage. The knock-on effect is generators and businesses get no incentive to act efficiently because $1 trillion in inefficient transaction fees globally are extracted from their bottom lines.
The growth in AI only accelerates a legacy fragile problem. It puts the UK between a rock and a very expensive place, footing the bill for data-center expansion instead of profiting from it.
To stay competitive, we need to create an unfair advantage in how we produce and move electrons. Because it’s not just compute on the line; it’s manufacturing, economic sovereignty, and the cost of living itself. If we get it wrong, we risk becoming a tourist economy in a digital world.
The UK’s moment of opportunity
AI will be the defining generational technology of our era. But to lead, the UK needs an energy system that’s as modern as the models it’s trying to run. Today we’re already helping hyperscalers to immediately get savings and reduce cost to compute without moving to long corporate PPA agreements.
I’ve always believed the UK to be a testbed for intelligent energy markets. The UK not only has one of the most complex markets in the world but it leads two to ten years ahead of other markets, with unfair advantage in wind, tidal and biofuel.
It’s already possible today for data centers and AI giants to buy transparent, cleaner low-cost power that doesn’t come with a small print catch that costs the little guy nearby.
The infrastructure to transact transparently and fairly exists. And very soon we’ll have a modern system where every new megawatt of AI demand drives a corresponding drop in national energy waste.
This is all entirely achievable if we fast connect the dots between compute demand, generation to power it and a modern transaction flow. And every second we spend debating those dots or squabbling over who should pay for it, we lose momentum, businesses lose profit and we lose huge progress for humanity.
We've featured the best green web hosting.
This article was produced as part of TechRadarPro's Expert Insights channel where we feature the best and brightest minds in the technology industry today. The views expressed here are those of the author and are not necessarily those of TechRadarPro or Future plc. If you are interested in contributing find out more here: https://www.techradar.com/news/submit-your-story-to-techradar-pro
You must confirm your public display name before commenting
Please logout and then login again, you will then be prompted to enter your display name.