Opinion

Every AI conversation eventually becomes an energy conversation

There is something rather familiar about the way we are talking about artificial intelligence. The conversation is dominated by models, chips, valuations and astonishing demonstrations of what machines can now achieve. It reminds me a little of the early days of the internet when we obsessed over websites rather than fibre optic cables, or the railway age when we marvelled at locomotives without paying much attention to the steel beneath them.

Technology has a habit of distracting us from the infrastructure that makes it possible. Yet every discussion about AI eventually arrives at the same place, namely energy and somewhere, vast quantities of electricity were consumed to make it happen.

That has implications far beyond technology. For much of the last fifty years, developed economies have quietly assumed that electricity would simply be there. It was reliable, relatively inexpensive and largely invisible. Businesses worried about wages, interest rates and taxation. Even now. few board meetings devote much time to the availability of power itself.

But we are entering an era where economic growth increasingly depends upon digital infrastructure, and digital infrastructure depends upon energy. As AI becomes embedded into financial services, healthcare, manufacturing, logistics, defence and scientific research, electricity stops being an operational cost and becomes a strategic asset.

We have been here before not least in The Industrial Revolution which was as much about access to coal as it was steam engines per se. The twentieth century was not simply defined by the motor car but by oil. Entire alliances, conflicts and economies were and continue to be shaped by whoever controlled the fuel that powered progress.

It is entirely possible that the defining strategic resource of the twenty first century will not be data but the energy required to process it and this will change how we think about investment and public policy.

Countries that can generate abundant, dependable and increasingly clean electricity will enjoy an advantage that extends far beyond lower household bills. They will attract data centres, advanced manufacturing, pharmaceutical research, semiconductor fabrication and the thousands of businesses that increasingly depend upon computational power. Those that cannot may discover that capital is global and, unfortunately for them at least, remarkably mobile.

This is why debates around nuclear power, grid investment, battery storage, renewable generation and transmission networks are no longer separate conversations about environmental policy or infrastructure spending. They are discussions about economic competitiveness.

Artificial intelligence promises extraordinary gains in productivity, medical research, education and countless other fields. It may help us solve problems that have frustrated generations. But every technological leap creates second and third order consequences that deserve just as much attention as the breakthrough itself.  One of those consequences is that energy becomes valuable in new ways.

Ironically,  AI itself may become one of the tools that helps solve the challenge it creates. Smarter energy grids, more efficient generation, predictive maintenance, better storage management and optimised consumption all become more achievable through the very technology that is increasing demand. Throughout history, innovation has often found ways to ease the constraints created by earlier innovation and that should offer some cause for optimism rather than alarm.

AI is not simply another software cycle. It is the beginning of a profound reorganisation of the economy. The winners will not only be those writing better algorithms or building faster processors. They will also be those producing, transporting and managing the energy that underpins the entire system.

30th July 2026