Power demand from UK datacentres set to surge six-fold over the next decade

National Grid CEO's warning comes as OpenAI chief Sam Altman says nuclear fusion is the answer

John Pettigrew, the chief executive of UK's National Grid, has forewarned that power demand from British datacentres is poised to skyrocket by an astonishing six-fold over the next decade due to exponential growth in AI usage and advancement in quantum computing.

Pettigrew delivered the ominous forecast at a conference in Oxford on Tuesday.

"Future growth in foundational technologies like artificial intelligence and quantum computing will mean larger scale, energy-intensive computing infrastructure," Pettigrew said.

This trajectory mirrors global trends, with many nations confronting the daunting task of financing the monumental investments required to bolster capacity in the face of surging demand.

According to projections by the International Energy Agency, global electricity demand stemming from datacentres, AI, and cryptocurrencies is slated to more than double over the next three years, amplifying the urgency of grid reinforcement initiatives.

The AI industry alone could rival the energy consumption of an entire country like the Netherlands by 2027.

In the Republic of Ireland, which hosts the European headquarters of tech giants such as Google and Meta, datacentres have emerged as significant consumers of electricity. Accounting for nearly a fifth of the country's total electricity consumption in 2022, the proliferation of these facilities has reignited debates over infrastructure capacity and sustainability.

The nascent field of quantum computing, heralded for its potential to revolutionise computation and cryptography, adds another dimension to the energy conundrum.

While still at an experimental stage, quantum computers can perform complex calculations at unprecedented speeds, promising scientific breakthroughs and advancements in secure communication protocols.

According to Pettigrew, the surge in energy usage could strain the already overburdened grid in the UK, necessitating urgent and innovative interventions to accommodate the escalating demand.

Pettigrew reckons this juncture demands bold action, comparing the current grid constraints to those faced in the 1950s.

"Demand on the grid is growing dramatically, and forecast to double by 2050 as heat, transport and industry continue to electrify."

The UK's aging supergrid, a network of high-voltage transmission cables and pylons dating back over 70 years, is ill-equipped to cope with such exponential growth. To align with global targets for net-zero emissions by 2050, significant upgrades to the grid infrastructure are imperative, ensuring sustainable energy provision by integrating renewable sources.

In response to this impending crisis, National Grid is exploring the feasibility of introducing an ultra-high-voltage network with a capacity of up to 800 kilovolts – a twofold increase from current levels.

This ambitious proposal aims to facilitate seamless bulk power transfers across the country, effectively interlinking major energy sources with burgeoning demand centres. However, the proposed grid enhancement scheme comes with a staggering price tag, estimated at approximately £112 billion.

Can nuclear fusion be the solution?

As AI continues its meteoric rise, Sam Altman, the head of OpenAI, has ignited a spirited debate by heralding nuclear fusion as a potential saviour.

In a series of interviews, Altman underscored the indispensability of fusion in providing the massive amounts of power required for next-generation AI systems.

"There's no way to get there without a breakthrough, we need fusion," Altman said in a January interview.

Then earlier this month, during a conversation with podcaster and computer scientist Lex Fridman, Altman once again highlighted fusion as the solution to AI's "energy puzzle."

Altman has himself invested hundreds of millions of dollars in fusion research.

But, scepticism abounds among experts, who view Altman's emphasis on a future energy breakthrough as emblematic of the wishful thinking permeating the AI landscape. Fusion could be decades away.

Alex de Vries, a data scientist at Vrije Universiteit Amsterdam, said that prioritising immediate solutions is paramount, given the urgency of addressing AI's escalating energy consumption. Efficiency savings through optimisation won't be enough, he added.

"Efficiency gains have never reduced the energy consumption of cryptocurrency mining," de Vries told CNN. "When we make certain goods and services more efficient, we see increases in demand."