Brookfield Asset Management, led by CEO Connor Teskey, estimates the AI buildout will require about $7 trillion over the next ten years. To capture a share of that market, the firm has created a joint venture with Nvidia that could acquire up to $100 billion of AI infrastructure assets. In addition, Brookfield has joined a Nvidia-organized consortium that aims to mobilise roughly $500 billion together with private-equity peers Apollo, Blackstone and KKR.
Teskey likens data centres, the servers they house and the power plants that feed them to traditional civil-engineer projects such as bridges and ports,assets that Brookfield has long built, owned and operated. He notes, however, that AI-related facilities are being constructed on a “much larger scale and happening much faster” than conventional infrastructure, creating both opportunities and the risk that some financiers will fall behind as the market expands.
Power generation emerges as the dominant bottleneck for digital infrastructure, according to Teskey, and the capital required for new energy capacity will persist through the end of the decade. He favors technologies that can be deployed quickly,fuel cells, on-shore renewables, battery storage and nuclear,while warning that new gas turbines have lead times of a decade or more. Consequently, Brookfield is steering investments toward renewables and storage, viewing gas as a necessary but slower-to-scale component.
The consortium assembled by Nvidia brings together five of the industry’s largest investors, a rare sight on Wall Street where rivalry typically discourages cooperation. Teskey explains that the sheer scale of the AI infrastructure market forces competitors such as Apollo, Blackstone and KKR to sit at the same table, because the combined capital needed far exceeds what any single firm can raise. He likens the financing evolution to a traditional path: independent owners, followed by bank syndication, then private credit and insurance participation.
Brookfield stresses that it will only pursue a tiny fraction of the opportunities it evaluates, yet even that narrow filter allows the firm to deploy “a tremendous amount of capital.” Projects are typically secured under 20- to 25-year contracts with top-tier counterparties, guaranteeing returns regardless of future demand fluctuations. Teskey dismisses narratives that suggest AI infrastructure demand is fragile, arguing that even a modest slowdown would leave a persistent supply-demand imbalance that sustains the need for new capacity.