Big Tech burns cash on AI — and markets keep buying in

Global markets this week were defined by a head-on collision between AI investment skepticism and AI growth conviction. As Microsoft, Meta and then Amazon reported earnings in succession, one question dominated market attention: does the astronomical spending on AI actually pay off?

Capital expenditure at major tech companies has surged to record levels as they race to build AI data centers and secure semiconductors. As a result, some have seen their free cash flow turn negative. Yet as earnings reports have rolled in, the market has increasingly shifted toward the view that AI investment is already translating into revenue.

Amazon is the clearest example. Amazon Web Services posted second-quarter sales growth of 36.7 percent year-on-year, beating market expectations on growth rate, profitability and order backlog alike. AWS accounted for 21 percent of Amazon's total sales but contributed 61 percent of its operating profit, while its order backlog grew 36 percent from the previous quarter to $496 billion.

"AWS has demonstrated across growth rate, profitability and order backlog that expanding AI investment is necessary," said Kim Seung-hyeok, an analyst at Kiwoom Securities.

The financial strain is real, however. Amazon's free cash flow over the trailing 12 months swung to a deficit of $7.6 billion, driven by net capital expenditure climbing to $169 billion. The company also raised its full-year cash-based capital expenditure plan to $220 billion.

But markets focused on the prospect of recouping those costs rather than the deteriorating cash flow. The company said AI server investments are typically recovered within three years on average, that most recent AI contracts run five years or longer, and that a significant portion of current spending is already backed by secured demand. Amazon's share price strengthened after the earnings release.

Analysts at South Korean brokerages have also cautioned against excessive concern over AI spending. Kwon Young-bae, an analyst at Mirae Asset Securities, said worries about Big Tech's free cash flow are overblown. "Don't doubt Big Tech's financial firepower," he said.

He added that these companies are already monetizing AI in earnest — not only through cloud services but also in core businesses such as advertising, search and e-commerce — and that they have ample means to fund investment through share buyback suspensions, external borrowing or dividend cuts.

Mirae Asset Securities identified the most significant development confirmed during this earnings season as AI demand feeding directly into cloud growth.

Combined cloud sales at Alphabet, Amazon and Microsoft jumped 42.9 percent year-on-year, and new orders also surged. Particularly notable, the firm said, is that AI demand is spreading beyond specific industries into finance, the public sector and the broader economy — a sign that the AI investment cycle remains in its early stages.

The race to build AI infrastructure shows no sign of letting up for now. Amazon raised its capital expenditure plan to $220 billion this year, pledging to direct most of it toward AWS and AI infrastructure. The company also plans to double AWS's power supply capacity by the end of 2027. It projected that data center investment will continue for years, given that AI adoption remains in its early stages.