This week’s selloff in technology stocks underscores a gnawing anxiety among investors: What happens if you throw a big party and few people show up?

The Nasdaq Composite Index has slipped nearly 3% this week as Wall Street frets over whether the trillions of dollars going into artificial intelligence will deliver the revenue and profit growth needed to justify that exorbitant cost.

Goldman Sachs estimates tech companies will spend $7.6 trillion through 2031 to build thousands of new data centers to power the rise of AI. But fresh data is raising questions about whether enough consumers and businesses are willing to pay up for these services, even as the tech giants leading the AI charge borrow heavily to build the required infrastructure.

“There’s concern around how much hyperscalers are turning to debt markets in order to finance the infrastructure buildout,” Kate Brennan, associate director of independent research institute AI Now, told CBS News, referring to the tech companies driving the torrent in AI capital spending — Alphabet, Amazon, Meta, Microsoft and Oracle. 

She added, “The returns are not coming in, and the claims that are being made, in terms of efficiency or productivity numbers, are not netting out.”

Brennan also pointed to rising skepticism among some consumers and workers about the utility of AI. To be sure, Americans are increasingly using AI, but for now few appear willing to pay for it. That reluctance is coupled with what polls show are major public concerns with AI: 40% of adults think the technology will be a negative societal force over the next two decades, versus 16% who believe it will be positive, according to Pew Research.

Meanwhile, more companies are laying off workers and investing in AI instead, heightening concerns about the technology’s impact on jobs. For employers, the payoff is uncertain. A May study from tech research firm Gartner found that businesses that replace workers with AI agents often fail to generate a return on investment.

One takeaway is that many consumers are using AI less out of a desire to chat with a bot than because there’s simply no escaping the technology, Brennan said. Enter a search query on Google, and you’ll get an AI response at the top of the page. Call a company’s helpline, and chances are that you’ll get an AI agent with a soothing voice accompanied by fake typing in the background.

“The current push for AI adoption that we’re seeing is directly coming from the financial incentives of AI firms,” she added. Because of the massive capital expenditures, the hyperscalers and other AI firms are making a “deliberate push for AI everywhere — no matter whether the demand is there or if customers want it or not.”

Bubble or bust?

Wall Street has long worried about an AI bubble as companies like Alphabet and chipmaker Nvidia have repeatedly propelled the U.S. stock market to new records. To some investors, the current moment is analogous to the dotcom bubble of the late 1990s. While many of those early Internet high-flyers flamed out, the ones that survived — think Amazon and Google — eventually became profitable businesses or even household names.

As with that earlier boom-and-bust cycle, the AI landscape is likely to yield uneven outcomes, according to Qian Wang, global head of capital market research at Vanguard, and senior global economist Kevin Khang.

“Some firms may emerge as more profitable and with significant competitive advantages, while others could find their core businesses obsolete in a new AI economy,” they said this week in a report. “As we continue to learn what the economics of AI look like in practice — the trajectory of AI capital expenditure, how effectively hyperscalers can monetize AI investment, and the size and shape of AI’s addressable market — the market’s sensitivity to the ups and downs is likely to be significant.”

They added, “Investors should expect a bumpy ride.”

The payback test

A key question underlying the lofty valuations of the hyperscalers and other AI companies is whether their capital spending plans reflect realistic revenue forecasts, according to economist Ed Yardeni of Yardeni Research. 

Companies including Alphabet, Amazon, Meta and Microsoft are spending heavily on data centers and chips in expectation of strong demand for AI services, while large language model developers like OpenAI and Anthropic pay to use their data centers. Yet it remains to be seen whether consumers and businesses will ultimately generate enough revenue to justify those investments.

“The AI ecosystem falls apart if the expected end-user demand for the AI/LLM products does not materialize or if prices for their offerings fall sharply below expectations,” Yardeni said in a note to investors. 

Yardeni’s team examined annualized revenue estimates for OpenAI and Anthropic to assess whether they’re adding users fast enough to cover their spending commitments with the hyperscalers — what he calls a “capex payback test” to check whether these companies can support the industry’s capital expenditures.

Their conclusion: Not right now, but the picture will improve in several years if current growth forecasts hold.

“We find that the AI ecosystem is not fully end-user revenue-backed yet, but it is not entirely speculative either,” Yardeni said. “Expected 2030 revenues make the math look much better. But those forecasts depend on a big assumption: AI revenues must continue to scale, and compute efficiency must improve, or both.”