BlackRock chief executive Larry Fink warned on 21 October that delays in building AI data centres could raise computing costs and threaten economic growth, Bloomberg reported. He made the comments to Bloomberg Television at the Berlin Global Dialogue, where he argued that adding capacity would bring down the cost of compute.
Key points
- Fink said slow data-centre construction could keep compute costs high and add to inflation.
- On 14 October, he warned that a US construction moratorium would cause a recession.
- Analysts cited by CNBC estimated AI capital expenditure could exceed $1 trillion in 2027.
Fink links Berlin warning to compute costs
“If we don’t build out these data centers, it is actually more inflationary because the cost of compute is so high right now,” Fink said in the television interview. His warning concerned the price of the computing capacity used for AI as demand for it accelerates.
Fink argued that building more data centres faster could lower compute costs: “And the faster we can build out more data centers, the more compute we have, the faster we can bring down the cost.” The speed of construction was central to both parts of his argument — the inflationary pressure from constrained supply and the prospect of cheaper compute as more capacity becomes available.
The Berlin remarks followed a warning about the economic reliance on construction spending. On 14 October, Fink had addressed what a halt to US data-centre projects could do to growth. His later comments concerned the cost of the capacity those projects would supply.
Fink’s 14 October moratorium warning
“If we have a moratorium of some sort in this country, we will be in a recession,” Fink told CNBC on 14 October. The interview took place after BlackRock reported third-quarter earnings that exceeded expectations. BlackRock shares were up about 1% in late-morning trading, CNBC reported.
“So much of our economy is based on the buildout of this capex,” Fink said, referring to investment in AI infrastructure. He described demand for construction workers, physical digital infrastructure, AI computing hardware and energy infrastructure to secure power for data centres. That spending formed the basis of his recession warning in the CNBC interview.
CNBC reported local opposition to data-centre construction involving concerns about utility costs, water use, environmental effects and transparency over investment. Fink acknowledged concerns about the buildout but said “most of it is on misinformation”. He also argued that data centres were creating jobs. His recession warning addressed a moratorium on construction, rather than a delay at an individual project.
AI spending estimates extend to 2030
Wall Street analysts estimated that total AI capital expenditure could exceed $1 trillion in 2027 and amount to more than $5 trillion through 2030. The projections cover AI investment, while Fink’s warnings focused on data-centre construction and the computing capacity it would bring into service.
Moody’s estimated US data-centre capital spending would reach $700 billion in 2026 and as much as $2.48 trillion through 2030, CoinDesk reported. That estimate concerns data centres in the United States, rather than all AI capital expenditure.
