BlackRock chief executive Larry Fink delivered a blunt warning on Wednesday: a broad U.S. moratorium on data center construction could weaken an economy that has become increasingly dependent on artificial intelligence investment.
"If we have a moratorium of some sort in this country, we will be in a recession," Fink told CNBC after BlackRock reported third-quarter results. He argued that the buildout is supporting demand for construction labor, computing hardware, electrical equipment and grid infrastructure.
The statement matters because it links a local political dispute to the national investment cycle. Communities near proposed data centers have raised concerns about electricity bills, water use, environmental effects and limited transparency around development incentives. Those objections have become more visible ahead of the November 3 midterm elections.
Fink acknowledged those fears but said much of the debate rests on misinformation. His recession forecast is an opinion, not an established economic result. A slowdown in data center approvals could reduce capital spending, but the effect would depend on the size, duration and geographic reach of any restrictions, as well as whether projects moved to other states.
The scale of the spending explains why investors are paying attention. CNBC cited Wall Street estimates that annual AI capital expenditure could exceed $1 trillion in 2027 and total more than $5 trillion by 2030. Those forecasts are uncertain, but they illustrate how much expected demand has already been built into suppliers, utilities and infrastructure developers.
BlackRock's own results provided the backdrop. A company filing showed $205 billion of quarterly net inflows and assets under management of roughly $15.5 trillion. The firm also said demand for capital is growing across the economy, placing its comments about AI infrastructure within a wider financing boom.
Data centers are creating jobs and equipment demand, but they also impose real costs on host communities and power systems. Investors should therefore treat the policy debate as a constraint on project timing and returns, not as a simple choice between growth and no growth. The most important evidence will be whether permitting delays begin to change construction schedules, utility capital plans or orders for AI hardware.