The largest US technology companies are starting to make money from artificial intelligence, but the cost of building the necessary infrastructure is rising even faster. For investors, therefore, it is becoming increasingly important not only to consider the rate of growth in AI revenue, but also how much cash remains after funding data centres, servers and networks.
According to a Reuters analysis based on the LSEG consensus, Microsoft, Alphabet, Amazon, Meta and Oracle could collectively spend more on capital expenditure in 2027 than they generate in free cash flow. Between 2025 and 2027, their annual operating cash flows are set to rise by around $340 billion, whilst capital expenditure is set to increase by around $534 billion. This equates to $1.57 of additional investment for every additional dollar of cash from operating activities.
Free cash flow shows how much cash a company has left after covering operating costs and capital expenditure. It is used to finance, amongst other things, dividends, share buy-backs and acquisitions. When infrastructure spending grows faster than the cash generated by the business, companies’ room for manoeuvre shrinks.
The first effects of these investments are already visible. Microsoft reported that its AI business exceeded an annual revenue scale of $37 billion, growing by 123 per cent year-on-year. Meanwhile, Amazon Web Services’ revenue increased by 28 per cent in the first quarter of 2026, to $37.6 billion.
The pressure is most evident at Oracle. The company ended the 2026 financial year with $32 billion in operating cash flow, $55.7 billion in capital expenditure and a negative free cash flow of $23.7 billion. It also announced plans to raise $45–50 billion through debt and share issues to further expand its cloud business.
The key question is no longer whether Big Tech will invest in AI, but when this spending will translate into sustainable growth in revenue, margins and cash flow. If monetisation lags behind infrastructure expansion, investors may increasingly question the scale of the current boom.
