A.I. Needs Power, Land and Data Centers. Bruce Flatt’s $1 Trillion Brookfield Already Owns Much of It.

“It took us 25 years to be in the right spot,” Bruce Flatt, chairman of Brookfield Asset Management, told investors at the firm’s New York investor day on Sept. 17. He was talking about artificial intelligence. Long before ChatGPT triggered a mad dash for raw compute, Brookfield spent decades building up what has become the physical backbone of today’s A.I. race, from power plants and renewable-energy projects to utilities, industrial businesses and even the nuclear reactor maker Westinghouse.

The A.I. boom isn’t running on algorithms alone, but needs land, data centers and staggering amounts of electricity. That puts Brookfield, the Wall Street investment giant managing more than $1 trillion, in an unusually advantageous position. Now the firm is pushing into nearly every layer of the A.I. buildout, financing computing equipment, developing sprawling data center campuses and striking multibillion-dollar power deals from France to South Korea.

Connor Teskey, who succeeded Flatt as CEO of the asset management arm in February, was even more direct: “A.I. infrastructure and A.I. is undoubtedly the biggest theme at Brookfield today,” he said. At 38, Teskey is one of the youngest executives to run a firm of Brookfield’s size. Flatt stayed on as chair of the asset manager, a seat previously held by Mark Carney before he entered politics and became Canada’s prime minister.

By Teskey’s estimate, between 25 percent and 40 percent of Brookfield’s global activity now touches digital infrastructure growth in some fashion. The opportunity is also reshaping assets the firm already controls. Land once earmarked for solar or battery projects is, in some cases, being reevaluated as a data center site, where surging demand for computing capacity can produce more attractive economics.

Brookfield’s pitch rests on a simple premise: Much of what A.I. needs looks remarkably similar to the long-lived infrastructure and cash-generating businesses it has been buying up for decades.

Flatt, a Winnipeg-born accountant who joined Brookfield’s predecessor, Brascan, in 1990, built his career around infrastructure, real estate and businesses capable of generating steady cash flows over long periods. That approach helped transform Brookfield into one of the world’s largest alternative asset managers. Flatt also serves as CEO of Toronto-based Brookfield Corporation, the broader investment company from which Brookfield Asset Management was spun out in 2022. The New York-headquartered asset manager trades on both the New York and Toronto stock exchanges.

The philosophy helps explain the firm’s enthusiasm for what private equity chief Anuj Ranjan described at investor day as “boring businesses”: companies that are easily overlooked while investors chase flashier technology bets.

“People are chasing the toy, the A.I. toy, and they are forgetting about industrial businesses,” Flatt said at investor day. Brookfield is not the only investment giant angling for a piece of A.I.’s physical backbone. Rivals are assembling war chests of their own.

In May, Blackstone committed an initial $5 billion to a venture with Google that will offer customers computing capacity built around Google’s custom A.I. chips. The following month, KKR and partners unveiled Helix Digital Infrastructure with more than $10 billion in committed capital to finance and build data centers, power infrastructure and networks. KKR tapped former AWS chief Adam Selipsky to lead the effort.

BlackRock, meanwhile, teamed up with Microsoft and Abu Dhabi investment firm MGX in 2024 to pursue $30 billion in equity for A.I. data centers and supporting energy infrastructure. With debt financing added to the mix, the group says the initiative could eventually mobilize up to $100 billion in total investment.

“The industry has signed up and committed to a build-out at a scale and speed that we’ve never seen before,” Gordon Bell, a principal at consulting firm EY-Parthenon, told S&P Global Market Intelligence in May.

Increasingly, Brookfield has also been moving closer to the technology itself. In August, Nvidia announced partnerships with some of the world’s investment firms (Brookfield, Apollo, BlackRock, Blackstone, Goldman Sachs and KKR) to create financing platforms designed to mobilize more than $500 billion in third-party capital for A.I. infrastructure. The idea is to make it easier for A.I. companies and enterprises to finance expensive Nvidia computing infrastructure rather than buying all of it outright.

Nvidia CEO Jensen Huang has described the shift as a new investable asset class. “The computer is now part of the infrastructure, like electricity, like the internet,” Huang told CNBC in August after the Nvidia announcement. “And so, you have to think about it like it’s infrastructure and build it out accordingly.”

Flatt argues that money itself is not the biggest constraint. “This is not about, ‘Is there too much financing being made?’” he also told CNBC, speaking alongside Huang and other Wall Street executives. “It’s that we can’t build it fast enough.”

Brookfield is putting plenty of its own capital behind that idea. Last November, it launched a global A.I. infrastructure program designed to deploy up to $100 billion across the full A.I. value chain. At its core is the Brookfield Artificial Intelligence Infrastructure Fund, backed by Brookfield, Nvidia and the Kuwait Investment Authority, targeting $10 billion in equity commitments. Its mandate stretches beyond traditional data center investments, spanning energy, land and computing infrastructure.

However, “the biggest mistake that could be made is assuming the only way we’re playing it is through the A.I. Infrastructure Fund,” Teskey told investors.

Power may actually be Brookfield’s biggest advantage. Its landmark 2024 agreement with Microsoft calls for more than 10.5 gigawatts of new renewable-energy capacity, including wind and solar, in the U.S. and Europe between 2026 and 2030. Brookfield has estimated the deal at more than $10 billion, while Microsoft described it as nearly eight times larger than any previous corporate power-purchase agreement.

When the electric grid cannot move fast enough, Brookfield is also exploring alternatives. Last October, it committed up to $5 billion to finance projects using Bloom Energy’s fuel cells, which can generate electricity directly at data center sites. By June, as demand surged, the companies expanded that framework fivefold to $25 billion.

The strategy also extends beyond the U.S. In France, Brookfield expanded an A.I. infrastructure framework with the government, increasing the total to €30 billion. In Sweden, it has committed up to $10 billion to build A.I. and cloud infrastructure centered in Strängnäs. And in South Korea, Brookfield signed a nonbinding agreement to provide up to $9 billion toward expanding Naver’s A.I. infrastructure at its hyperscale data center in Sejong, while Nvidia separately plans to invest $1 billion in Naver.

And then there is nuclear power. Brookfield acquired Westinghouse Electric Company out of bankruptcy for roughly $4 billion in 2018, at a time when new U.S. nuclear construction had largely stalled. Five years later, Brookfield brought uranium producer Cameco into the business, with Brookfield-led investors retaining a 51 percent stake and Cameco taking the remainder.

Today, Westinghouse sits at the center of Brookfield’s A.I.-era energy ambitions. Last October, Brookfield and Cameco announced a partnership with the U.S. government to construct at least $80 billion worth of new Westinghouse reactors across the country. The proposed buildout is tied, in part, to the enormous power demands of data centers and A.I. computing.

“Nobody was building a nuclear plant,” Flatt recalled on CNBC. Now, he added, “there’s a renaissance going on today in the United States.”

Brookfield did not respond to Observer’s request for comment.