(NEW YORK) — Fears of an artificial intelligence bubble have rattled the stock market in recent weeks and set off concern among critics about a wider risk to the U.S. economy.
A surge of AI spending accounted for roughly two-thirds of gross domestic product growth over the first half of 2025, JPMorgan Asset Management found, outpacing the contribution made by hundreds of millions of U.S. consumers. Many of the nation’s largest companies have poured funds into the chips and data centers necessary to operate AI.
A central question looms over the fate of the technology and the trillions of dollars being spent to develop it: Will AI deliver the type of profits that could turn the product into a moneymaker?
Proponents say a lag between the buildout of AI infrastructure and an onrush of gains is to be expected, pointing to a similar lull after the introduction of other watershed technologies, such as the internet. The widespread adoption of products like OpenAI’s ChatGPT has revealed a massive potential customer base, they add, noting AI firms have prioritized product development over profits.
Critics, however, say the considerable costs have put pressure on AI to deliver stratospheric profits, but little evidence suggests businesses or everyday users will get enough value to warrant forking over a mountain of cash. The technology must deliver within years rather than decades, they add, since the current level of spending cannot be sustained.
“It’s not particularly unusual for a market at this early stage to not be making much profit,” Paul Kedrosky, a venture capitalist and research fellow at MIT’s Institute for the Digital Economy, told ABC News. “Of course, the difference is most markets at this stage aren’t also spending a trillion dollars.”
AI boosters and skeptics alike have raised alarm about the economic stakes. “A reversal would risk recession. We can’t afford to go backwards,” David Sacks, a venture capitalist and White House czar for crypto and AI, said in a post on X on Monday.
Gary Marcus, a professor emeritus at New York University and author, who often criticizes hype surrounding AI, said in a Substack post in September: “It’s not going to be pretty when the music stops.”
A “bubble” is a term used to describe a market in which an asset’s price far outpaces its value on the market. Questions centering on the productivity gains and profitability of AI take up the task of assessing the economic value of the new technology.
Chip giant Nvidia has delivered major profits selling the semiconductors behind AI, becoming the most valuable company in the world by market capitalization. Such success indicates appetite for the building blocks of AI rather than its end uses, however.
For now, AI has failed to achieve gains on a scale near its immense costs, some analysts said. A product like AI would typically generate revenue in the form of sales either direct to consumers or to third-party businesses using the technology to enhance their offerings. AI has faced challenges on both fronts, some analysts said.
Roughly 95% of businesses invested in AI have failed to make money off of the technology, an MIT study in July found, estimating the combined amount spent by the firms is around $40 billion.
“Despite high-profile investment, industry-level transformation remains limited,” the study said.
Consumer-driven profits have also proven elusive. OpenAI’s ChatGPT, for example, boasts about 800 million weekly active users, making it one of the fastest-growing apps ever. That user base makes up about a quarter of the 3 billion monthly active users combined on the array of apps offered by Meta, a company that generated more than $50 billion over a recent three-month period. But OpenAI’s sales do not come close.
OpenAI CFO Sarah Prior told CNBC in September the company is on pace to earn about $13 billion in revenue over the course of 2025, which amounts to $3.25 billion per quarter. On the BG² podcast earlier this month, OpenAI CEO Sam Altman said the company is generating “well more revenue than that.”
Revenue is “growing steeply,” Altman added. “We are taking a forward bet that it will continue to grow, and that not only will ChatGPT keep growing, but we will be able to become one of the important AI clouds, that our consumer device business will be a significant and important thing, that AI that can automate science will create huge value.”
Some analysts said the rapid adoption of chatbots underscores the usefulness of the technology, noting that it paves the way for a potentially significant revenue stream if firms were to populate the AI assistants with advertisements or charge for access.
“It’s the fastest adoption of basically any consumer technology that we know about,” Ethan Mollick, a professor of management at the University of Pennsylvania who studies AI, told ABC News. “There is a path to making money.”
Arun Sundararajan, a professor of entrepreneurship at New York University, said a delay in uptake from businesses is to be expected for a potentially paradigm-shifting technology like AI.
“It’s true that we haven’t yet seen evidence of significant productivity gains from AI investments, but I’m not surprised,” Sundararajan said. “At the early stages of the rollout of a technology like this, there’s a lot of experimentation and learning.”
“As businesses start to understand how to fundamentally change the way that they work using this technology, that’s when you start to see the big productivity gains,” Sundararajan added.
Other analysts disagreed about the likelihood of profits, pointing in part to the challenge posed by infrastructure costs associated with AI.
For many digital products such as software or smartphone apps, the profitability owes to the relatively low cost of providing the service on a massive scale, Kedrosky said. For instance, the initial cost burden of developing a website is significant, but once completed, a website can reach millions of users with little extra cost.
For AI, however, the energy and computational costs increase in proportion to a given number of chat prompts or users, meaning the technology lacks such low-cost scalability.
“Every time you prompt an AI model, it eats up costs to maintain and cool servers. Those costs rise with the number of users. That’s a problem,” Kedrosky said.
The scale of investment also places pressure on AI companies to deliver major profits within a limited timeframe, since the current level of financing cannot continue into perpetuity, Andrew Odlyzko, an emeritus University of Minnesota mathematics professor who focuses on financial bubbles, told ABC News.
“The problem is when you talk about investments in data centers in the trillions of dollars and do the basic financial arithmetic of how much revenue you have to bring in to justify that, it gets into figures larger than total revenues of Google,” Odlyzuko said.
To be sure, some analysts said the technology remains in an early stage of its development, making the outcome uncertain.
“We’re in the early innings,” Vasant Dhar, a professor of data science at New York University who believes AI will ultimately deliver significant profit, told ABC News. “It remains to be seen what form it will take.”
Copyright © 2025, ABC Audio. All rights reserved.
