AI sector needs 9% of US GDP in revenue by 2032, analysis finds

AI tools lift code output but not completed projects, study finds

The investment now flowing into artificial intelligence implies that American businesses and consumers would need to spend roughly 9% of US gross domestic product a year on AI services to justify it, according to an analysis by Columbia University finance professor Stijn Van Nieuwerburgh. His calculation, first presented at the Brookings Institution, puts the revenue AI companies would need to reach by 2032 at $3.5 trillion, or 8.8% of GDP under assumptions that include 4% annual GDP growth, unadjusted for inflation. Van Nieuwerburgh’s paper concludes that the AI buildout is now larger than any investment boom in American history.

Writing about the analysis in The Wall Street Journal, chief economics commentator Greg Ip put the implied spending in everyday terms: roughly what Americans spend on food, about twice what the nation pays for all forms of energy or for all computers and software, and seven times what consumers spend on phone, streaming, and internet services combined. His verdict on whether Americans will actually spend that much: “You should be skeptical.”

Van Nieuwerburgh built the estimate from the published plans of today’s data-center builders and the average cost of a data center. He assumed some data centers would be canceled, that each dollar of revenue produces 50 cents of cash flow, and that builders require an unleveraged return on invested capital of 10%.

A central assumption is that AI companies can continue charging today’s prices for computing power. Van Nieuwerburgh said the current scarcity of computing capacity is holding up prices and profit margins, and that the companies are “basically saying we’re going to keep charging scarcity pricing in 2032, when presumably there will be tons of competition and we’ve quadrupled our capacity.”

Falling prices have undone earlier investment booms. Between 1997 and 2001, the price of bandwidth on fiber between London and New York plunged 96% as new fiber strands came online and new technology expanded each fiber’s capacity, a collapse that accelerated the bankruptcy of long-haul fiber companies. The same dynamic is visible in AI: models double in capability roughly every four to five months, and the effective price for a given level of capability has fallen 47% per quarter since 2023, according to Epoch AI — about six times faster than the price of computing power.

The optimistic case is that falling prices and rising capability spur enough new demand to offset the price decline — a relationship dubbed the “Jevons Paradox” after British economist William Stanley Jevons, who identified the phenomenon in 19th-century English coal consumption. Recent demand trends support that view. An expert panel surveyed by a team led by Ezra Karger at the Federal Reserve Bank of Chicago projected that OpenAI and Anthropic would post combined annual revenue of $300 billion by 2030; Ip called that projection too conservative, noting the two companies’ combined revenue run rate is already around $180 billion and is growing at double-digit rates each quarter.

Some AI advocates argue the technology could become a third factor of production alongside capital and labor, which would make 9% of GDP look modest — labor currently receives about 51% of GDP. In that scenario, the analysis notes, employment would plummet and governments would need to tax AI to replace lost income and payroll taxes. The more prosaic risk, the analysis suggests, is that current growth rates will not be sustained once adoption is widespread and the most productive applications have been deployed.

The history of the personal computer offers a caution about where a ceiling might lie. From the 1960s through the early 2000s, the density of microchips doubled roughly every two years, a phenomenon dubbed “Moore’s Law,” and the real price of computers fell about 15% per year. The personal computer and killer apps such as spreadsheets fueled an explosion of demand that more than offset that price decline: between 1974 and 1984, business investment in computers and peripherals tripled to 0.8% of GDP. Thereafter, better and cheaper computers no longer drove the same new demand — the first versions of VisiCalc and Lotus 1-2-3 revolutionized business in ways that ever more powerful versions of Microsoft Excel could not — and business spending on computers plateaued, apart from secondary booms tied to the internet in the late 1990s and to AI today.

Yet Ip argued that the current AI cycle is following a different trajectory. “AI is improving and adoption is growing much more quickly than with computers, and productivity improvements have been impressive,” he wrote, framing the technology as advancing faster and reaching further than the computer plateau that followed the early personal-computer era. But the evidence on whether those gains translate into completed work is mixed. A review of software projects by Harvard economics doctoral students Fiona Chen and James Stratton, using the platform Jellyfish, which firms use to analyze their engineering teams, found that AI assistants boosted lines of code by 12% and “pull requests” — instances when new code is merged into an existing code base — by 5%. AI agents, which can operate autonomously, boosted lines of code by 30% and pull requests by 23%. Even so, the authors found that AI’s effect on completed projects was small and “statistically insignificant,” because users were spending much more time reviewing, commenting on, and changing pull requests.

The takeaway, Ip wrote, is that AI will almost certainly raise productivity and economic growth, but not necessarily in ways that translate directly into revenue, or into the returns investors are now counting on.