Quartz reports:

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 [9/24/26] 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, Amazon, Meta Platforms, Microsoft, and Oracle — 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%. [Continue reading…]

MIT economist, Daron Acemoğlu, writes:

What is missing from our current debate is any discussion of a fundamental dilemma these numbers pose: can the AI boom avoid both an economically costly crash and a huge increase in inequality?

If the industry reaches these revenues, inequality surges. If the industry does not become profitable, a crash, with substantial costs in terms of lost output and jobs, becomes likely.

My assessment would be that the industry is unlikely to reach levels of revenue Van Nieuwerburgh calculates. First, diffusion has been and will likely continue to be slow. Second, competition from open-weight models, which are getting better, will limit how much proprietary models can charge. Third, despite important advances, I still believe that AI models will not be able to automate entire occupations anytime soon, thus limiting their value to businesses as cost-saving devices.

Whether this leads to a crash or not is more complicated and will depend on whether various AI companies are bailed out and what kind of support they receive.

Nevertheless, even if revenues fall short of these gargantuan amounts and we avoid a dramatic surge in inequality, I expect that the diffusion of AI will push up inequality between capital and labor and within labor.

If inequality does surge, a further question becomes central: can our democracy survive such astronomical levels of inequality?