The AI boom is predicted to disrupt developed economies, but to make that possible companies will have to build new data centers and power plants to operate them. That’s something climate tech investors have taken note of: Companies that provide that kind of infrastructure secured some of the biggest deals of 2024, according to a recent review of the year’s climate tech investing trends.
The review, published this month by Sightline Climate, a leading provider of intelligence on climate tech investing, reveals that total spending continued to be hit by high interest rates and other factors, causing 2024 to see the third successive year of declining investment. But sectors impacted by data center growth bucked the trend: Energy attracted $9.4 billion, a 12 percent increase, and buildings $2.7 billion, up 10 percent.
The increases were driven in part by anticipation of the energy-hungry computations that future AI products will carry out. A typical generative ChatGPT query, for example, requires close to 3 watt-hours of electricity, roughly 10 times more than a Google search consumes, according to the International Energy Agency. This additional load is arriving just as efficiency gains in the data centers that perform these computations are slowing down, analysts at Goldman Sachs noted last year. These and other factors will cause power demand from data centers, which has been flat over the past decade, to grow at 2.4 percent annually until 2030, the analysts forecast.
The money flows into climate tech serve as a barometer into larger trends, both positive and sometimes negative.
Billions for cleaner data centers
Some of the tech companies behind AI products, including Google, Amazon and Microsoft, also have relatively ambitious emissions targets. That’s helped spur interest in technologies that can underpin a cleaner growth trajectory for AI. “The rapid rise in AI demand creates challenge and opportunity for climate tech investment,” said Kim Zou, co-founder and CEO of Sightline Climate. “On one hand, we’re seeing unprecedented load growth coming onto an already constrained grid. On the other hand, AI-led demand is driving momentum for emerging clean firm power solutions like nuclear and geothermal as well as more sustainable data center tech, such as energy-efficient chips and liquid cooling systems.”
One example is Amazon’s $500 million investment in X-Energy, a startup that’s developing new nuclear reactor designs. Amazon and X-Energy are targeting deployment of more than 5 gigawatts of new power projects across the U.S. by 2039, around 10 percent of the total additional capacity that Goldman Sachs estimates will be required to support the growth of data centers in the U.S. through 2030.
Another example from the top 10 list is Form Energy, which has developed an iron-based battery that can feed electricity into the grid for 100 hours, more than 20 times longer than most of today’s equivalent systems. The battery allows utilities to smooth out mismatches between supply and demand caused by surges in emergy use in the mornings and evenings, as well as the intermittent nature of solar and wind. The company raised a $405 million round last year.
Data Centers dominate deals landscape
The biggest beneficiary in dollar terms was Crusoe Energy, which announced a $600 million raise in December. The company was founded in 2018 with a relatively niche product: data centers designed to host cryptocurrency mining, powered by natural gas from oil wells that would otherwise have been flared. It has since expanded to offer a vertically integrated suite of AI services, including construction of data centers optimized for AI, supply of clean power and cloud services. Scala, a data center provider based in Brazil that also has a focus on clean energy, raised $500 million last September.
“We’ve had very, very flat power demand and growth and what that means is that there hasn’t really been an incentive to try and change,” said Anku Madan, a principal at venture capital firm Obvious Ventures. “Now we have a perfect storm. We have EV demand coming online. We have industrial electrification. We have AI data centers. And all of that means that now there is a true incentive for utilities and for developers to quickly iterate and get the best technologies up and running.”
It’s unlikely that these will be the only startups to benefit from the AI boom. Madan cited a Microsoft recent announcement that it expected to spend $80 billion on AI-enabled data centers this year. “When you hear one company say we’re going to spend $80 billion dollars on this particular trend, that means that we’re talking about multiple hundreds of billions of dollars of spend across all the different hyperscalers,” he said. “There’s an opportunity there, Crusoe and others are examples of that, to help power those data centers.”
