AI’s $650 Billion Question: Where’s the Return?

Five technology giants are chasing their AI dreams harder than ever this year, dedicating more than $650 billion for capital expenditures (CapEx) on data centers, specialized hardware and infrastructure projects.

That’s well over three times what these hyperscalers spent in 2022 but far less than the $720 billion analysts expect them to spend next year.

This spending spree by Amazon, Microsoft, Alphabet (Google), Meta (Facebook), and Oracle is changing the economic and stock-market landscape. Expectations for these five companies and their CapEx beneficiaries are sky high.

But what is their actual return on investment (ROI)?

Prior to the increase in AI spending, the five companies generated significant free cash flow—a metric that indicates the amount of cash generated by the business after CapEx. This indicator was central to an investment case that included steady long-term growth in revenues, attractive profit margins, high ROI, low capital expenditure and cash returned to shareholders via dividends and buybacks. These characteristics justified above-market valuations.

But this year, those five companies (in aggregate) will likely fail to generate any free cash flow, and the ROI on the hundreds of billions in CapEx is still an open question. The market expectation is that the ROI eventually will justify the investment—generating significantly higher revenue, profit margins and free cash flow beginning by this decade’s end.

This is a massive bet on the success of these five companies, given their outsized impact on AI beneficiaries (such as semiconductor stocks) and the broader stock market itself. If analysts are correct, the companies’ free cash flow will return to 2025 levels by 2029.

That’s a long time to wait for a return on investment.

Bottom line: Should investors put all their eggs in the AI basket? This AI story could be an exception to history, which tells us that CapEx sprees like this often end poorly. However, investors would be well served diversifying across global equity markets—in case this time is not different. Also, look for venture-capital opportunities in AI-related startups as a better way to work this gold rush.