The next great constraint on artificial intelligence (AI) isn’t silicon — it’s copper, transformers, and transmission lines.
That was the stark message delivered by BlackRock Chairman and CEO Larry Fink in a recent interview with CNN, where he argued that America’s ageing electricity grid has become the single largest obstacle standing between the country and AI supremacy.
“We don’t have enough power in the United States,” Fink said bluntly, framing the problem not as one of raw generation but of distribution. The country sits on ample energy resources — particularly natural gas — but lacks the transmission infrastructure to move that power where it’s needed most.
“We have plenty of power through natural gas, but we can’t distribute it in a proper way,” he explained, estimating that hundreds of billions of dollars in grid investment will be required to close the gap.
The stakes, in Fink’s telling, could hardly be higher. “If we don’t do that, we are not going to succeed in AI,” he warned. His logic is disarmingly simple: “AI is just a bunch of electrons. So you need the power to create the electrons.”
Fink’s diagnosis goes beyond electricity alone. He described a broader imbalance where demand for AI computing capacity outstrips supply at multiple levels — advanced chips, power, cooling, and the physical real estate to house data centres. “At this moment, there’s more demand than supply,” he said.
“We have shortages of compute right now which to me is the biggest problem we have in this country today.”
This bottleneck has a winner-takes-all dimension that troubles Fink. While he expressed little concern about whether BlackRock or J.P. Morgan can afford to invest in next-generation AI models, he drew a sharp line at the implications for everyone else.
“I am very worried about municipalities or hospitals,” he said. “Are they going to invest in this?”
The solution, Fink argued, lies in systematically lowering the cost of entry. Computing costs must fall far enough that local governments, transport systems, hospitals, and small businesses can deploy advanced AI tools without requiring the capital reserves of a Wall Street giant.
He called this project “democratising AI,” and cautioned that failing to achieve it would produce “some real structural issues” — a polite way of describing a two-tier economy where only the largest institutions harness the productivity gains that AI promises.
No bubble, just scarcity
On the question that has rattled markets for two years — whether the torrent of AI investment constitutes a speculative bubble — Fink offered a contrarian view. He dismissed bubble concerns outright, pointing instead to the same supply-demand dynamics reshaping the energy landscape.
Unusually intense demand has created genuine shortages, and those shortages have allowed certain companies to command premium pricing. In Fink’s framework, that’s not froth; it’s a straightforward signal that capacity needs to expand.
Fink also stepped back to survey the broader macroeconomic picture, and his assessment was notably sanguine. Recent geopolitical shocks — including the conflict involving Iran — had, in his view, demonstrated the adaptive strength of the global economic system rather than its fragility.
“The global economy actually mitigated much of the stresses,” he observed, crediting a combination of increased energy production, supply-chain diversification, and rapid technological adaptation. His shorthand for this phenomenon was characteristically direct: “We solve problems.”
The growth imperative
Turning to America’s fiscal trajectory, Fink was equally clear-eyed. Rising government debt, in his analysis, cannot be taxed away — it must be outgrown. He argued that the United States needs to sustain roughly three per cent annual economic growth to keep its debt burden manageable. “If we cannot grow the economy by three per cent a year, we’re in trouble,” he said.
The prescription that followed was classically Fink: encourage private-sector investment, streamline the regulatory approvals that bog down infrastructure projects, and resist the temptation to lean on higher taxes as a primary remedy.
In his view, the path from here runs through construction cranes and transmission towers — not just through server racks and semiconductor fabs.
