AI Build-Out Is the Biggest Economic Bet - and Risk - In American History

When your economic boom is being compared, not to railroads, automobiles or the internet, but to the 2008 subprime mortgage crisis, it might be time to reflect.

A new study by the respected Brookings Institution reveals that the AI boom is on track to exceed every other technology-driven expansion in American history. It may also be the riskiest because the combination of uncertain revenue and profit forecasts along with increasingly complex as well as legally and financially murky debt placements is, to be polite, Panglossian. The projections underpinning these funding arrangements are based on assumptions that require 80% annual growth rates, which implies that all the residents of Mars, Venus and Saturn in addition to everyone on earth, will be buying nothing but AI for the foreseeable future. But other than that, what could go wrong? JL

Cris Tolomia reports in Quartz:

The U.S. AI buildout is on pace to consume a larger share of economic output than any prior technology in US history, including railroads, interstate highways and the internet — while the financing underpinning it carry systemic financial risks. Total spending on data centers and related AI infrastructure will be $10.3 trillion between 2025 and 2032, 3.6% of gross domestic product each year. That tops the 2.2% the railroad boom  - the previous record - three times higher than GDP devoted to building the interstate highways or the telecom/internet. AI' scale, combined with still-unproven revenue streams and complex debt arrangements, is also driving inflation. Spending has outpaced what major tech companies can fund from their own cash flows. A growing share of that is financed through debt, often routed off-balance-sheet with little public disclosure. Achieving return targets implies annual growth of 80%. "This is reminiscent of the 2008 subprime mortgage crisis." 

A new Brookings Institution study finds that the U.S. artificial intelligence infrastructure buildout is on pace to consume a larger share of economic output than any prior technology rollout in American history, including railroads, the interstate highway system, and the internet — while the financing structures underpinning it carry systemic financial risks. 

The study — authored by Columbia Business School finance and real estate professor Stijn van Nieuwerburgh — puts total spending on data centers and related AI infrastructure at $10.3 trillion between 2025 and 2032, equivalent to roughly 3.6% of gross domestic product each year. That figure tops the 2.2% of annual GDP that the railroad boom of the late 1800s absorbed — the previous record — and runs more than three times higher than the GDP share devoted to building the interstate highway network from the 1950s onward or to the telecom expansion that got underway in the mid-1990s, according to The Wall Street Journal.

Van Nieuwerburgh's paper, presented Thursday at a Brookings conference, warns that the scale of the buildout, combined with still-unproven revenue streams and increasingly complex debt arrangements, has created conditions for a potential downturn. "This is freaking complicated," van Nieuwerburgh said of the financial arrangements linking AI firms, major tech companies, banks, private credit lenders, and real estate firms. 

The investment has already outpaced what major technology companies can fund from their own cash flows. According to the Wall Street Journal, FactSet data show the combined capital expenditure of five major cloud and tech players — Alphabet $GOOGL +1.34%, Amazon $AMZN +0.04%, Meta $META +4.50% Platforms, Microsoft $MSFT -0.53%, and Oracle $ORCL -3.47% — is on pace to total $4.2 trillion across the four-year span through 2029. A growing share of that spending is financed through debt, often routed through off-balance-sheet entities with limited public disclosure.

Van Nieuwerburgh drew a direct comparison to the 2007-2009 financial crisis. "This opacity of all these special purpose vehicles is somewhat reminiscent of what happened in the subprime mortgage crisis," he said. Hitting that return target would require the AI industry to reach roughly $3.7 trillion in annual revenue by 2032 — a pace that, starting from an estimated combined revenue base of about $100 billion at OpenAI and Anthropic today, implies annual growth of around 80%.

The buildout is also generating broader economic effects. The pattern has been consistent: data center construction produces temporary employment spikes during building, but the facilities themselves operate with minimal permanent staff. LinkedIn data cited by the Wall Street Journal credit AI with generating upward of 750,000 U.S. jobs between 2023 and 2026, even as economists continue to debate what the technology means for the workforce overall. 

The buildout is also driving inflation. Import price data show that computers, peripherals, and semiconductors cost 20% more in August compared with the same month a year prior, according to the Wall Street Journal. Chicago Fed President Austan Goolsbee has warned that data center investment is bidding up wages in related sectors, while Fed Chair Kevin Warsh pointed to hyperscaler borrowing as a factor lifting long-term rates — a dynamic that has rippled into mortgage costs for ordinary homebuyers.

Van Nieuwerburgh said the risks do not necessarily point to imminent financial distress. "Strong growth in AI applications, high utilization, and continued improvements in model capability could support the projected infrastructure and generate stable cash flows," he wrote. "But the combination of uncertain demand, rapid technological change, execution bottlenecks, and high leverage creates meaningful downside risk if expectations are revised."