Larry Fink said BlackRock will announce a hyperscaler partnership to build data centers, as investors brace for power and compute shortages.
AI infrastructure is moving from “tech buildout” to a defined investment theme, and BlackRock’s Larry Fink just signaled the next chapter.
Speaking at the Milken conference in Los Angeles. the BlackRock CEO said the company will announce a partnership with a hyperscaler later this week to build data centers.. The message was direct: the AI future is already here. and the bottlenecks that follow are increasingly financial questions as much as engineering ones.
Fink’s comments also framed BlackRock’s role beyond stock-picking or traditional asset management.. Misryoum reports that the firm has positioned itself to finance parts of the AI buildout. pointing toward infrastructure such as data centers and energy investments—areas that often determine how quickly new AI systems can be deployed.
This matters because data-center capacity is tightly linked to electricity availability, hardware supply, and long construction timelines. When those constraints collide with a global rush to scale AI, capital allocation becomes a competitive advantage rather than a background activity.
BlackRock’s push sits within a broader pattern across private markets.. Misryoum notes that the firm’s later-2024 acquisition of Global Infrastructure Partners (GIP) added weight to its infrastructure ambitions. and BlackRock has also been involved in data-center efforts with companies and investors across the ecosystem.. In parallel. other alternative asset managers have increasingly focused on AI infrastructure. sometimes coordinating through deals instead of competing only for separate targets.
The direction of travel is also visible in how deal strategies are evolving.. Misryoum reports that some managers are pursuing not just ownership of assets. but frameworks aimed at helping companies adopt AI technology through their portfolio businesses.. Meanwhile, large cloud players continue planning large-scale spending that underlines how central compute infrastructure has become.
Even with that momentum, Fink warned that the industry may still be underfunded relative to what’s required. He suggested there may be a shortage of capital for the buildout, even as headlines repeatedly highlight enormous AI budgets.
The key implication is that the AI race is not only about models—it’s about the physical supply chain that powers them. If capital, energy, and compute capacity lag behind demand, financing structures and new risk-management tools could become as important as the AI products themselves.
At the same time. Fink raised the idea of a new kind of asset class tied to compute futures. arguing that companies may need ways to hedge rising compute costs.. Misryoum adds that the broader theme from the conference was clear: the market is treating infrastructure constraints as the next major economic variable for AI. not a temporary hurdle.