Technology companies like OpenAI, Google, and Meta are vying for leadership in the AI arms race, battling a shortage of engineers and advanced chips to prepare for the massive growth expected in the coming years.
But winners are already emerging in the AI infrastructure sector, which includes critical data centers and energy projects needed to meet AI's vast appetite for power and information. It's private equity.
Companies like Backstone, Carlyle Group and KKR are quietly spending tens of billions of dollars on energy projects and data centers aimed at supplying AI developers, hoping that AI will spark the expected surge in demand. It is expected that huge profits will be made.
“Even before AI came along, we saw an industry growing at more than 20% annually, and having AI at the top will only accelerate that growth even further,” says a managing director at Blackstone. Director Greg Blank said.
The data centers that developers use to train large language models behind AI consume enormous amounts of energy, potentially sucking up a quarter of all electricity in the United States by 2030.
While energy shortages have not yet materially hurt U.S. tech companies' profits, price hikes in other areas of the AI supply chain, such as TSMC's chips, are an example of what they may face in the years ahead. is.
Enter private equity. Building AI energy and data infrastructure from scratch is a multibillion-dollar undertaking, but companies like BlackRock at $1 trillion, his KKR at $553 billion, and Carlyle Group at $426 billion, We have the financial power to make that happen.
Blackstone was an early entrant, acquiring data center provider QTS in 2021 for about $10 billion. CEO Stephen Schwarzman said on an earnings call Thursday that Blackstone has invested $50 billion in data centers to date.
“The amount of money being invested in this area is amazing and it’s happening all over the world right now…AI and EV [are] There are huge investment opportunities being created,” Schwarzman said in a speech at the Asia-Pacific Finance and Innovation Symposium in Melbourne earlier this week. “U.S. states are starting to run out of power. U.S. power growth has been about 1% a year. AI is expected to add at least 2%, and some think 3%.”
Carlyle Group is targeting the renewable energy sector, investing $2 billion in solar power projects outside Phoenix, a chip manufacturing hub, that will bring in more factories for top AI chip suppliers such as TSMC.
“We knew there was a lot of demand from our corporate customers for energy from these projects,” Pooja Goyal, head of renewable energy at Carlyle, told Semaphore. “But we obviously didn't take into account the AI-driven demand lift that's happening now. That was a huge acceleration to our original investment thesis.”
Meanwhile, KKR and its competitors, including Bain Capital and Warburg Pincus, are focusing on Asia, with KKR investing $1 billion in data centers in the region.
“The total capacity of all operating hyperscale data centers will nearly triple over the next six years,” KKR said in a recent report. KKR partner Waldemar Schlesak added that this means some AI companies concerned about demand are already securing data center space for five years.
AI players in the tech space are also following private equity firms with investments of their own. Microsoft has announced it will invest nearly $6 billion in data centers in Japan and the UK, and OpenAI founder Sam Altman is backing a startup that provides small-scale nuclear power for AI data centers. And the two companies are reportedly discussing a massive $100 billion AI infrastructure project, dubbed Stargate, that will support future AI models.
Privately funded infrastructure projects could relieve some of the strain on the power grid, and federal regulators are rushing to meet AI's power demands. But given that only the biggest tech and finance players appear to be able to provide the large sums of money needed to fund AI infrastructure projects, observers believe they will dominate the field. and seeks intervention to prevent it from crowding out competitors.
“Policymakers should use antitrust tools to control the harms that result from the concentration of AI in the 'tech stack,'” said Ganesh Sitharaman, a law professor at Vanderbilt University, and Tejas, a law professor at the University of California, Berkeley. Narechania writes in a recent paper.