Stay informed with free updates
Simply register at Technology sector myFT Digest – Delivered straight to your inbox.
One of the happiest promises made about artificial intelligence is that it will allow us to tackle the world’s biggest challenges, like climate change. AI can help manage smarter power grids, design electric vehicles, and perform more efficient monitoring. Plastic pollution in our oceans.. But data centers hosting the latest AI models consume shocking amounts of energy and water. Is AI more of a problem than a solution when it comes to the climate emergency?
The ways in which technology is moving in opposite directions are highlighted by the experience of Microsoft, which in 2020 made one of the bolder environmental commitments in corporate history. By 2030, the technology company promised, it would be carbon negative and by 2050 it would have offset all the emissions it had generated since its birth in 1975. But Microsoft still has a long way to go. Last month, it reported that its emissions had increased. 29 percent from 2020 while continuing to invest massively in data infrastructure.
This week, the company announced it would invest $3.2 billion over the next two years to expand its cloud computing infrastructure in Sweden. In total, Microsoft intends to spend more than $50 billion this year on data centers in what an analyst has called “The largest infrastructure construction humanity has ever seen.” Data consultancy Gartner forecasts that global data center spending will rise 10 percent this year to $260 billion.
Additionally, Microsoft and OpenAI, the AI startup it heavily backs, also plan to invest up to $100 billion in building a US-based supercomputer and data center, according to Information.
The large data infrastructure required to run energy-intensive generative AI models is already causing environmental damage. As Microsoft noted in its latest sustainability report, the infrastructure and electricity needed to power the latest technologies create “new challenges to meeting sustainability commitments across the technology sector.” Although the company remained optimistic that it would achieve its long-term goals of becoming carbon negative, water positive and zero waste, it acknowledged that it was not yet on track to reduce indirect emissions and replenish more water than is used in its centers of data.
Researchers are working hard to make AI models do more with less energy, which clearly has great financial and environmental appeal. As part of its green commitment, Microsoft is also spending heavily on renewable energy. Last month, the corporate giant pledged to bring 10.5 gigawatts of renewable energy online in the US and Europe in partnership with Brookfield Asset Management. The additional capacity, equivalent to powering 1.8 million homes, will cost approximately $10 billion.
The company is making some wilder bets small modular nuclear reactors to produce carbon-free energy. It also agreed to buy electricity generated by nuclear fusion from Helion Energy by 2028 (assuming it can be produced by then). Helion is a startup generously funded and backed by Sam Altman, CEO of OpenAI.
Big AI companies argue that, in the grand scheme of things, the energy demands of the digital economy remain comparatively small. According the International Energy Agency, data centers account for less than 1.5 percent of global electricity use. But what is disconcerting is how quickly they are growing. The IEA predicts that by 2026, data centers could consume 1,000 terawatt hours of electricity, more than double that of 2022 and about the same usage as Japan.
Since the environmental disadvantages of AI are so evident, it is even more imperative to grasp the advantages. One of the most intriguing fields is weather prediction, which can help us adapt to climate change. In November, Google DeepMind researchers GraphCast presentedan AI model that is more accurate (and more energy efficient) at making 10-day weather forecasts than conventional methods.
As more countries consider geoengineering responses to climate change, such as cloud seeding and carbon capture, AI researchers are also focused on modeling potential impacts. “Weather issues related to national security are a huge issue right now,” the CEO of an artificial intelligence company tells me. “People assume geoengineering is going to happen and want to know what the effects will be.”
For now, the environmental costs of AI are real, while the benefits remain more unclear. It is time for the industry to deliver more on its ambitious promises.