As data centre electricity demand is forecast to rise more than fivefold by 2030, a new report argues that only a rapid expansion of nuclear power will stop AI investment drifting to countries with cheaper, cleaner and more reliable energy.

Unsurprisingly, that new report comes from the Nuclear Industry Association itself, who worked in partnership with Oxford Economics. It notes that the UK risks falling behind in the global race to scale AI unless it accelerates nuclear development.

In Powering the UK Data Boom: The Nuclear Solution to the UK’s Data Centre Energy Crunch, the authors forecast that UK data centres’ electricity use will climb to 26.2TWh by 2030 – more than five times today’s level. Put another way, it would be equivalent to almost 9% of projected national electricity demand and around 30% of commercial consumption.

The report warns that without a significant increase in firm, low-carbon generation this decade, the UK will be forced to lean more heavily on gas to keep digital infrastructure running, driving up both emissions and energy costs.

Data boom colliding with grid constraints

AI workloads and cloud services are already reshaping power demand profiles, with operators working to build new sites on 18–24-month timelines. By contrast, the report highlights decade-long waits for some grid connections and notes that the UK faces among the highest industrial electricity prices in advanced economies.

That combination of long lead times, expensive power and higher emissions is painted as a growing competitive risk. The authors argue that hyperscale data centre and AI investments could be diverted to markets that can offer faster access to clean, reliable supply at lower cost.

Clarissa Hahn, Economist and Report Author at Oxford Economics, said, “We forecast that UK data centres’ electricity consumption will rise by 2030, accounting for about 9% of the UK’s projected national electricity demand. In this report, we identify several competitiveness challenges facing the UK—such as high industrial electricity prices and delayed grid connections. Our findings show that nuclear energy is uniquely positioned to meet this growing demand: offering reliable 24/7 power with a low-carbon footprint that the tech sector is calling for.”

Nuclear pitched as backbone for AI economy

The NIA and Oxford Economics position nuclear as a critical enabler of the data economy, arguing that 24/7 low-carbon generation is needed to underpin ‘five-9s’ reliability – 99.999% uptime – for AI and cloud services.

They call for a national strategy bringing together government, regulators, the AI sector and energy developers. They also echo calls from the UK Government’s Nuclear Regulatory Taskforce, which last month found that without significant reforms, nuclear power would not be able to save the UK’s AI ambitions. The reforms include:

• Accelerate all types of new nuclear, including large, small and advanced reactors
• Reform siting policy and streamline licensing
• Enable co-location so nuclear plants can power data centres directly
• Prioritise clean power access for critical digital infrastructure

Tom Greatrex, Chief Executive of the Nuclear Industry Association, commented, “Britain’s digital future depends on secure, clean energy. We can only build the AI economy of tomorrow with rapid, expedited deployment of clean, reliable, sovereign nuclear power in the mix. The recommendations of the Government’s Nuclear Regulatory Taskforce give us a perfect opportunity to create a better framework for deploying nuclear to meet data centre needs, but we must move now.”

William D. Magwood, IV, Director General of the Nuclear Energy Agency, added, “The Oxford Economics-NIA study provides crucial insights into the potential role of nuclear energy in meeting the growing energy demand of data centres in the UK – which the study estimates will have a five-fold increase by 2030. Nuclear energy, with its reliability, scalability and low carbon emissions, make it an ideal partner for fast-growing data centres.”

Questions over timelines and alternatives

While the report makes a strong case for nuclear as the backbone of the UK’s AI ambitions, there are some obvious caveats that deserve more scrutiny. For a start, the timelines simply do not align. Even on optimistic assumptions, new large-scale nuclear projects in the UK typically take well over a decade from approval to operation, with recent experience at Hinkley Point C hardly inspiring confidence on schedule or cost control. 

Small modular reactors are often presented as a quicker alternative, but no SMR design has yet been deployed commercially in the UK, and most are still navigating lengthy regulatory and financing hurdles. Setting out nuclear as the solution to an energy crunch that is expected to materialise by 2030 risks over-promising what this technology can realistically deliver in that timeframe.

There is also a question of proportionality. Forecasts that data centres could account for around 9% of UK electricity demand by 2030 sound alarming, but need to be weighed against wider system trends. National Grid ESO and others already expect substantial growth in offshore wind, solar, interconnection and storage over the same period, alongside continued efficiency improvements in IT hardware. 

That raises the question of whether targeted grid investment, demand-side response and strategic siting of data centres near existing low-carbon generation could meet much of this demand more cost-effectively than a wholesale bet on nuclear power.

Finally, the report leans heavily on the UK’s relatively high industrial power prices and slow grid connections as arguments for nuclear, but those are largely problems of regulation, network planning and market design rather than technology choice. Faster build-out of renewables, reinforcement of transmission capacity, reforms to grid queue management and a more coherent planning regime could all improve competitiveness without relying on reactors that will not come online for many years. 

Framing nuclear as the ‘only’ route to winning a global AI race risks crowding out a more balanced discussion about the mix of tools – including renewables, storage, efficiency and smarter networks – that could support both decarbonisation and digital growth.

But then again, the Nuclear Industry Association wouldn’t be doing its job if it didn’t work to promote the industry further.