Blackstone President Jon Gray says the surge in AI spending is the opposite of a classic bubble because demand for computing power already exists while the infrastructure needed to meet it is in short supply.
In an interview published on Friday by the trade weekly Capital Finance, Gray said a traditional bubble is driven by speculative construction in anticipation of demand, citing property and telecommunications in the 1990s as examples.
With artificial intelligence, he said, real demand for computing power is already here and no one is building a one-gigawatt data centre speculatively.
Instead, he argued, the supply of infrastructure needed to support AI demand is falling short, held back by political and regulatory constraints, which is exactly the opposite of what occurs in a classic bubble.
Gray pointed to recent restrictions in the United States. In July, New York State signed a one-year moratorium halting the issuance of permits for new large-scale data centres.
In Texas, Governor Greg Abbott has also paused two new data centre projects pending a review of their impact on water and energy supplies.
Gray added that a shortage of turbines for power plants needed to supply electricity to data centres is also slowing the expansion of capacity.
He cautioned that not every company in the race will succeed and acknowledged that not all valuations will prove justified. Even so, he said, demand for AI and its overall impact will exceed expectations.
In early August, Blackstone joined five other major financial groups — Apollo, BlackRock, Brookfield, Goldman Sachs and KKR — in a partnership with US chipmaker Nvidia aimed at raising $500 billion in capital.
