Scientists tell us that the heat accumulating across the Earth as a result of human emissions is roughly equivalent to the energy that would be released by 800,000 Hiroshima-sized atomic bombs exploding across the face of the planet every day. That number has doubled since it was first calculated in 2012, due to the continuing increase of greenhouse gases in the atmosphere.¹

The predicted consequences include the intense heat waves, the remarkably heavy rains and the ongoing sea-level rise that have been evident in recent years, among many others. The predictions emerging from computerised models of the climate include an intensification of the El Niño and La Niña cycles that can alter weather patterns worldwide. We may be about to live through such a cycle, with an El Niño having already begun — and predicted to intensify through the coming winter into an exceedingly powerful and disruptive force.²

During an El Niño, the ocean pours an immense amount of heat into the atmosphere, altering the global circulation. It is a natural cycle that starts with shifts in wind patterns in the tropical Pacific, but nowadays it always rides on top of human-caused global warming, which appears to intensify the damage. This El Niño can be expected to dump heavy rains in California, the southern United States, Peru and East Africa; to increase cyclones in the Pacific Ocean but suppress hurricanes in the Atlantic; to increase the spread of tropical diseases like malaria and dengue fever as heavier rains encourage mosquitoes; and possibly to cause a mass bleaching of coral reefs around the world, for what would be the fifth time in recorded history.³ Both 2026 and 2027 could set all-time global temperature records, though this is less certain. By hitting crops with drought or floods, El Niño has the potential to further destabilise a food supply already under strain.

This will be, of course, just one more grand disruption in a world that feels less stable and less predictable by the year.

Economic inequality has fed a growing public anger that, while it may be justified in principle, has been harnessed to the benefit of populist political parties offering false solutions. The populists have destabilised some Western democracies to the point that they are becoming ungovernable. Britain just inaugurated its seventh prime minister in a decade. The US continues its drift toward authoritarianism. Only in a few countries has the populist wave been decisively pushed back in recent elections.

Now comes a new disruption, potentially the most profound of the 21^st century: the advent of artificial intelligence programs that may be able to do much of the work people are accustomed to doing. How disruptive the advent of AI will be for society at large is still anybody’s guess. Economic history would seem to suggest that if AI removes barriers and eliminates some types of scarcity, the long-term result will be to leave society better off; but that hardly means no job losses or other painful consequences in the near term. Indeed, the potential of AI to roil entire industries seems obvious, though it is also true that many of the imagined uses of the technology have yet to be proven at scale.

In the long run, as we said in this report last year, AI could be a boon for the sustainability transition. It may help us improve efficiency, eliminate waste and solve many longstanding problems. However, it has big downsides for sustainability: the AI models are already being used to extract more fossil fuels,⁴ and in the near term, AI is placing sudden and profound demands on a power grid that is not ready for them.

The scale of the investments that companies are making to build their AI services is staggering. McKinsey & Company, the consultancy, has estimated that spending on data centres to meet AI workloads is already well into the hundreds of billions of dollars per year, and may exceed $5 trillion cumulatively by 2030.⁵ That sum is hard to believe: it would be five times the current annual budget of the Pentagon. Historically, the major Silicon Valley technology companies could finance themselves from their cash flow, but now they are borrowing immense sums to build data centres. At least half the construction is located in the US, although that country is also seeing a massive political backlash against data centres that cuts across political ideologies.

Data centres are not the only driver of rising electricity demand, and in most countries, not the most important. Electric cars and heat pumps are both contributing, a good sign for the energy transition. But in many countries the biggest factor is a rise in air conditioning use — partly a consequence of rising affluence, but also a response to rising temperatures.⁶ The intense heat waves we are now seeing can cause premature deaths by the tens of thousands, and access to air conditioning is becoming a matter of life and death in some parts of the world. That rise will, of course, lead to higher emissions on grids still powered by fossil fuels, one of the many troublesome feedback loops that is worsening the climate crisis.

Power prices are rising in many countries, for complex reasons that go beyond the rise of AI, but the apparent correlation between AI and power bills has fed public distrust in the technology. However, it is not axiomatic that rising power demand needs to lead to higher prices. In fact, for the power grid, the data-centre expansion may well be a massive opportunity disguised as a problem.

The tech companies are so desperate for connections to the electricity grid that they are flashing their wallets, willing to finance upgrades that could benefit all electricity ratepayers. The truth is that generators, transformers and power lines are often underutilised; if data centres can be operated in a way that uses these facilities when they would be idle, that could actually improve the overall economics of the grid and lead to downward pressure on rates. To make it work, the data centres will have to agree to cut their power demand at times when the grid is stressed. Some tech companies, including Google, are already signing deals along these lines. Governments and grid operators are starting to establish rules to capture these potentially broad benefits.

Still, there is no doubt that AI programs are going to demand electricity that would not otherwise have to be generated. The big question for us is: how clean will it be?

The tech companies are buying clean power whenever they can, but they are also willing to buy dirty power, and are now at risk of missing their own climate goals as a result. One of Elon Musk’s companies, xAI, faces a lawsuit claiming that it illegally operated large, portable gas generators to power a data centre in Memphis, Tennessee, polluting nearby Black neighbourhoods in the process.⁷ The breakneck pace of the buildout has even led Meta, the corporate owner of Facebook, to install huge computer clusters in tents located in Ohio and Tennessee.

It would, of course, be a public service for the tech companies to throw their political muscle into getting more clean power built. They may get a chance soon. In Washington, the Trump administration is turning every bureaucratic knob to try to block new solar and wind farms. That it has largely failed thus far does not necessarily mean it will continue to fail. But, with at least one house of Congress likely to be captured by the Democratic Party in November, the Trump administration will face a new political reality starting in January.  Restoring tax breaks and other policies to favour clean energy will be high on the Democrats’ agenda, and Mr Trump may have to accept such a plan to get a budget through Congress. The odds will improve if the tech companies throw their weight behind a clean-power bill.

So far, the populist wave roiling politics in the democracies has managed to defy physics, in the sense that the populists get away with disinforming people about the climate crisis. In the end, though, we think physics will win. The signs of climate breakdown are evident all around us. Polls suggest that the public understands it and wants action, even if concerns like the cost of housing and food are higher near-term priorities for voters.

We await a visionary new politics that sees cheap, clean energy as the key to solving many of these problems at once.