Operators Who Can Secure Power Will Capture the AI Buildout

What's New

The scarce input in the AI infrastructure buildout is the ability to deliver projects. Connor Teskey, CEO at Brookfield Asset Management, argues this in a podcast interview on Goldman Sachs Exchanges: Great Investors. Capital and demand are both plentiful. The shortage is in teams that can develop land, secure grid connections, and build power supply. Large power buyers now favor a few scale providers that deliver on time. Investors should weigh operating track record ahead of check size when they assess AI infrastructure exposure.

Why It Matters

Markets treat AI data center demand as fragile. They react hard when a more efficient model appears or a hyperscaler walks away from a lease. Teskey argues supply trails demand by such a wide margin that even falling demand would leave providers short. His view puts pricing power with established operators and against smaller developers competing on cost. Teskey still leads Brookfield's energy business, so the thesis supports his firm's positioning.

Zoom In

Ten years ago, Brookfield sold power mainly to utilities and governments. Today it sells most of its power directly to corporates. Teskey says buyers now specify the shape, location, source, and duration of the power they want. A hyperscaler that is ready to turn on a data center without power in place takes a real hit. Buyers respond by concentrating on a small number of providers who can work through problems and meet deadlines. Brookfield's relationships with these buyers have grown from single contracts into global framework agreements. Energy storage, which offsets the intermittency of renewables, is now the fastest growing segment of its energy business.

Memorable Quotes

• "Demand is up here and supply is down here." Teskey reduces the AI power market to one gap and builds his whole case on its size.

• "Even if demand growth was to slow, even if it was to plateau, even if it was to decline, there would still be a huge supply-demand imbalance to the providers of those assets." He dismisses the market's sensitivity to individual model or lease headlines.

• "It's a large sum of capital that's going to go into the ground for multiple decades going forward." The admission sets the stakes: assets this long-lived punish any error in the demand forecast.

• "There is no shortage of capital and there is no shortage of demand. The bottleneck is who can actually pull these projects out of the ground." He names execution as the source of returns in the buildout.

The Wrap

Teskey's case depends on power staying scarce relative to demand through the build cycle. It also depends on large buyers continuing to pay for delivery certainty over the lowest power price. Sovereign offtakers in France, Sweden, and Qatar must keep backing projects where corporate demand is thin. The case weakens if grid access opens up and many smaller developers start delivering on schedule. The end of this decade, the horizon Teskey sets for the theme, will settle it.