Summarised by Centrist
The energy needs of the artificial intelligence (AI) revolution are immense. For instance, a single large data centre can use as much electricity as a small city.
The entire output of a nuclear reactor is not enough to feed the insatiable appetite of companies like ChatGPT, whose computing needs far outstrips basic web searches of yesteryear.
The explosion in AI’s electricity consumption is straining the grid and prompting a resurgence in fossil fuel use, delaying the retirement of coal plants in several jurisdictions.
This level of consumption contradicts tech giants’ promises of transitioning to green energy.
In response, companies like Microsoft, Google, Amazon, and Meta are exploring experimental energy solutions such as small nuclear reactors and deep geothermal energy, but these projects face significant hurdles.
Ambitious clean-energy projects, including atomic fusion, aim to replicate the sun’s energy production process here on Earth, but they face scepticism that this can be done anytime soon.
Yet, the urgency of developing practical, clean energy solutions becomes ever more apparent. It is estimated that by 2030, massive data centres will account for nearly one tenth of the power consumed in the US.
