Home Latest Insights | News Big Tech Returns Just 9.3% of Cash to Shareholders as AI Spending Pushes Free Cash Flow Into Deficit

Big Tech Returns Just 9.3% of Cash to Shareholders as AI Spending Pushes Free Cash Flow Into Deficit

Big Tech Returns Just 9.3% of Cash to Shareholders as AI Spending Pushes Free Cash Flow Into Deficit

America’s largest technology companies generated a record $542.9 billion in operating cash flow during the first nine months of 2026, but returned a sharply smaller proportion of that money to shareholders as the cost of expanding artificial intelligence infrastructure absorbed an increasing share of their resources.

Microsoft, Alphabet, Amazon, Meta and Oracle returned just 9.3% of the cash generated by their businesses through dividends and share buybacks during the period, down from 27.4% a year earlier, according to a Reuters analysis of LSEG data and company financial reports.

The decline marks a significant change in the financial profile of companies that have spent years using large cash balances to reward shareholders through dividends and share repurchases. Their ability to sustain those distributions is now being tested by an investment cycle in which data centers, specialized processors and power infrastructure require large capital commitments.

Operating cash flow across the five companies increased 29% year-on-year to $542.9 billion. However, capital expenditure doubled to $556.9 billion, exceeding the cash generated from operations for the first time in data going back to 2011.

The result was a combined free cash flow deficit of $14 billion, compared with a surplus of $143.5 billion in the same period last year. The reversal shows how rapidly the AI investment cycle is changing the cash economics of even the industry’s most profitable businesses.

The pressure has fallen most heavily on share buybacks. The five companies reduced repurchases by 83% to $13.6 billion in the first nine months of 2026, from $81.8 billion a year earlier. Dividends, which companies are generally more reluctant to cut because investors treat them as recurring commitments, moved in the opposite direction, rising to $37 billion from $33.9 billion.

The figures reveal different approaches among the five companies. Alphabet and Meta have made no share repurchases so far in 2026, after buying back a combined $66.5 billion during the first nine months of 2025. Amazon, which does not pay a dividend, has not repurchased shares since 2022.

Microsoft accounted for nearly all the remaining buybacks among the five companies, although its repurchases also declined. The company spent $4.4 billion buying back shares in the September quarter, compared with $5.7 billion a year earlier.

The distinction between dividends and buybacks is remarkable. Dividends create an expectation of regular cash payments, making reductions potentially damaging to investor confidence. Buybacks are more flexible and can be scaled back when companies need to preserve capital for acquisitions, infrastructure, or other investment priorities.

That flexibility has made repurchases the first major casualty of the current spending cycle. However, their retreat also removes a source of demand for technology shares and a mechanism that has supported earnings-per-share growth.

The cash-flow figures raise a further concern: some companies are no longer generating enough free cash flow to cover their shareholder distributions after capital expenditure. With combined free cash flow negative, the $50.6 billion returned through dividends and buybacks was effectively financed through existing cash reserves, borrowing or equity issuance.

Amazon illustrates the scale of the funding pressure. Its capital expenditure reached $75.9 billion in the third quarter, almost 80% above the $42.4 billion generated by its operations. The company also raised $48.1 billion in long-term debt during the quarter.

Alphabet faced a similar cash-flow squeeze, reporting negative free cash flow while raising approximately $12.4 billion through share sales and $29 billion through debt.

These funding choices do not necessarily indicate financial distress. The companies retain substantial businesses, assets, and access to capital markets. However, they show that the expansion of AI infrastructure is increasingly being financed through a combination of operating cash, borrowing, and equity rather than internally generated cash alone.

The Earnings-Per-Share Effect

The reduction in buybacks carries implications beyond the immediate decline in shareholder distributions. Repurchases have helped technology companies increase earnings per share by reducing the number of shares outstanding.

When a company buys back and retires its shares, its profits are divided among fewer shares. Earnings per share can therefore rise even when total net income grows more slowly or remains unchanged.

Alphabet reduced its share count by 9.5% between the end of 2020 and 2025 through buybacks, according to ING. The reduction provided an estimated annual boost of roughly 2% to earnings per share, even without additional profit growth.

With repurchases slowing or stopping, companies will have less scope to generate that mechanical improvement.

The accounting cost of new infrastructure is also expected to increase. Data centers, servers, and other equipment are depreciated over their useful lives, creating expenses that can weigh on reported earnings even after the initial construction costs have been paid.

Investors therefore face two related questions: will AI spending generate enough additional revenue and operating profit, and will those returns arrive quickly enough to offset rising depreciation and financing costs?

Mark Schilsky, a technology sector specialist at JPMorgan, said in a client note last week that investors care “so much more about EPS than FCF.” That underscores a potential tension in how markets assess the companies. Earnings per share remain central to many valuation measures, while free cash flow indicates how much cash is left after operating costs and capital investment.

The gap between the two measures could become more important as infrastructure spending accelerates and buybacks provide less support to per-share figures.

The pressure is also visible in broader shareholder returns. Technology and communication services companies, which account for approximately 43.5% of US market value, delivered a shareholder yield of just 0.28% in July, compared with 1.99% for the rest of the market, according to MSCI.

Investors Face a Different Capital Bargain

There is little indication that the investment cycle will ease soon. Amazon expects capital expenditure of approximately $220 billion in 2026, while the other companies have also signaled further spending increases.

This means that the financial proposition is changing for shareholders. Instead of receiving a substantial share of excess cash through dividends and buybacks, investors are being asked to accept lower near-term distributions while companies commit more resources to AI infrastructure.

The expectation is that these investments will eventually generate enough revenue and profit to compensate for the reduced cash returns. However, the scale of capital expenditure means that the financial burden is already visible, while the full commercial returns remain uncertain.

Ehren Stanhope, co-head of investment management at Franklin Templeton, described the shift as a departure from the shareholder-return model that characterized much of the previous 15 years.

“For the last 15 years, companies increasingly returned excess capital to shareholders through buybacks and dividends,” Stanhope said. “The current regime may increasingly be characterized by greater equity and debt issuance, a request of capital from investors.”

That change could alter how investors value the largest technology companies. Their financial strength has traditionally rested on high margins, substantial cash generation, and the ability to fund expansion while returning capital to shareholders. The current spending cycle is testing whether those characteristics can be maintained alongside the enormous investment required to compete in AI.

No posts to display

Post Comment

Please enter your comment!
Please enter your name here