NEW YORK CITY -- There鈥檚 been one big question on the minds of Wall Streeters this tech earnings season: When will anyone start making actual money from artificial intelligence?

In the 18 months since ChatGPT kicked off an AI arms race, tech giants have promised that the technology is poised to revolutionize every industry and used it as justification for spending tens of billions of dollars on data centers and semiconductors needed to run large AI models. Compared to that vision, the products they鈥檝e rolled out so far feel somewhat trivial 鈥 chatbots with no clear path to monetization, cost saving measures like AI coding and customer service, and AI-enabled search that sometimes makes things up.

But Big Tech still has relatively little to show for all their billions spent in terms of significant revenue gains from AI or profitable new products, and investors are starting to get antsy.

Amazon鈥檚 (AMZN) less-than-impressive earnings and outlook Thursday could be mostly chalked up to concerns that it is spending a ton on AI without much to show for it, at a time when its core business also faces hurdles. That dragged the stock down nearly nine per cent Friday. Intel鈥檚 (INTC) stock plunged 25 per cent on Friday after the company said Thursday night that after big spending to adapt to the AI wave, it鈥檚 now trying to rein things in by cutting US$10 billion in costs and laying off tens of thousands of workers.

In short, investors鈥 fears can be boiled down to: is all of this actually worth anything? Or is it just another shiny object the industry is chasing to bring back its dreams of endless growth, before it abandons it and moves onto the next big thing?

As Morgan Stanley analyst Keith Weiss put it on Microsoft鈥檚 earnings call: 鈥淩ight now, there鈥檚 an industry debate raging around the (capital expenditure) requirements around generative AI and whether the monetization is actually going to match with that.鈥

UBS analyst Steven Ju asked Google CEO Sundar Pichai锘 how long it would take for AI to 鈥渉elp revenue generation 鈥 (and) create greater value over time, versus just cutting costs?鈥

And a Goldman Sachs report last week asked if there was 鈥渢oo much spend, too little benefit鈥 on generative AI.

Shares of both Google and Microsoft dipped following their earnings reports, a sign of investors鈥 discontent that their huge AI investments hadn鈥檛 led to far-better-than-expected results. Meta 鈥 which experienced similar shareholder frustration last quarter 鈥 avoided the same fate this time around by showing how its AI investments were at least contributing to its core business, including by enabling companies to easily make compelling ads with its AI tools.

Some investors had even anticipated that this would be the quarter that tech giants would start to signal that they were backing off their AI infrastructure investments since 鈥淎I is not delivering the returns that they were expecting,鈥 D.A. Davidson analyst Gil Luria told CNN.

The opposite happened 鈥 Google, Microsoft and Meta all signaled that they plan to spend even more as they lay the groundwork for what they hope is an AI future. Meta said it now expects full-year capital expenditures to be between US$37 and US$40 billion, raising the low end of the guidance by US$2 billion. Microsoft said it expects to spend more in fiscal 2025 than its US$56 billion in capital expenditures from 2024. Google projected capital expenditure spending 鈥渁t or above鈥 US$12 billion for each quarter this year. (Even for extremely rich companies, those are big numbers 鈥 for Google, its second quarter capital expenditures amounted to about 17 per cent of its total sales).

And tech leaders have said that what they need is more time 鈥 a lot more time.

Microsoft CFO Amy Hood said on the company鈥檚 earnings call that its data center investments are expected to support monetization of its AI technology 鈥渙ver the next 15 years and beyond.鈥

Meta, similarly, anticipates 鈥渞eturns from generative AI to come in over a longer period of time,鈥 CFO Susan Li told analysts. She added: 鈥淕en AI is where we鈥檙e much earlier 鈥 We don鈥檛 expect our gen AI products to be a meaningful driver of revenue in 鈥24. But we do expect that they鈥檙e going to open up new revenue opportunities over time that will enable us to generate a solid return off of our investment.鈥

That time horizon is uncomfortable for many investors, who have grown accustomed to mostly reliable, quarter-after-quarter sales and profit growth from Silicon Valley.

鈥淚f you鈥檙e going to invest now and get returns in 10 to 15 years, that鈥檚 a venture investment, that鈥檚 not a public company investment,鈥 Luria said. 鈥淔or public companies, we expect to get return on investment in much shorter time frames. So that鈥檚 causing discomfort, because we鈥檙e not seeing the types of applications and revenue from applications that we would need to justify anywhere near these investments right now.鈥

And some investors question whether AI investments will ever pay off. Goldman Sachs analyst Jim Covello argued that 鈥渢he technology isn鈥檛 designed to solve the complex problems that would justify the costs鈥 in last week鈥檚 report.

As an example of just how long it can take AI products to come to fruition, take Tesla鈥檚 AI-based 鈥渇ull self-driving鈥 technology. Tesla has sold the driver-assist technology as key to the company鈥檚 business plan since 2015, and consistently promised that it would be fully capable within a short timeframe. But FSD still requires an attentive human driver capable of taking the wheel in case something goes wrong, and is regularly plagued by safety concerns, nearly four years after it was first released to Tesla customers.

For now, tech CEOs appear to agree that 鈥渢he risk of underinvesting is dramatically greater than the risk of overinvesting,鈥 as Google鈥檚 Pichai said in last week鈥檚 earnings call (a similar line was repeated by Meta CEO Mark Zuckerberg during his company鈥檚 call). Data centers take time to build and if someone is going to come out the winner in the AI race, no company wants to miss their shot at the top simply because they didn鈥檛 have enough computing capacity. And they鈥檙e earning enough from their core businesses that investors will put up with the spending for now.

But at some point soon 鈥 Luria predicts it will be either later this year or early next 鈥 the pressure from investors to back off on infrastructure investments and let revenue growth play catch-up will be strong enough to get tech leaders to pull back.

鈥淩ight now, the game is, 鈥榳e all have to signal that we鈥檙e willing to invest as much as we need because we want to keep this leadership position,鈥 but at some point the investment is going to be so onerous that one of them 鈥 will say, 鈥榤aybe next quarter, we won鈥檛 invest so much,鈥 and then you鈥檒l see that happening for the rest of them,鈥 Luria said. 鈥淏ig picture, this level of investment is not sustainable.鈥