Microsoft free cash flow quality faces scrutiny as AI spending rises
Investors are paying closer attention to how large technology groups generate free cash flow as AI-related capital spending accelerates. Microsoft appears stronger than its hyperscaler peers on this measure, but some analysts argue its cash generation relies heavily on financing-related mechanics rather than only operating profitability.
Highlights
- Microsoft's $67 billion free cash flow for the last fiscal year is heavily supported by stock-based compensation ($12.4 billion), leasing ($25.6 billion), accrued spending ($19.8 billion), and investing adjustments ($19.6 billion).
- Total non-core adjustments to Microsoft’s free cash flow reached $75.3 billion—more than 100% of reported free cash flow—raising questions about its sustainability and comparability.
- Microsoft trades at a premium valuation of 25 times forward earnings, with markets re-examining whether elevated AI and capital spending justify its cash flow multiple versus other hyperscalers.
What drives Microsoft's cash flow
As reported by the Financial Times, investors are reassessing whether Microsoft’s free cash flow reflects durable operating strength or accounting and financing effects that may deserve a lower valuation multiple. The debate comes as markets focus more sharply on capital intensity across hyperscalers including Alphabet, Amazon, Meta and Oracle.Free cash flow is commonly calculated as operating cash flow minus capital expenditures, and investors favour it because it captures cash effects that earnings measures such as operating profit and net income can smooth over or miss. Working capital movements, including changes in receivables, payables and inventory, can materially alter a company’s economic position and make one company’s free cash flow more resilient than another’s.
Dave Zion of Zion Research Group argues that Microsoft’s free cash flow is heavily supported by four main factors. He points to stock-based compensation, lease-based capital spending, accrued capital expenditures that are not yet paid in cash, and investment items recorded outside standard capital expenditure lines, including receivables tied to server component purchases.
Zion estimates that stock-based compensation reached $12.4 billion in the fiscal year ended in June, equal to nearly 19% of Microsoft’s $67 billion in free cash flow. He also calculates that leasing rather than outright asset purchases added $25.6 billion, while accrued capital spending contributed $19.8 billion and other investing adjustments another $19.6 billion.
Valuation implications for hyperscalers
Taken together with smaller adjustments, Zion says these items total $75.3 billion, more than 100% of Microsoft’s reported free cash flow for the last fiscal year. His argument is not that Microsoft is misleading investors, but that the market should distinguish between cash generated by core business profitability and cash supported by favourable financing terms and balance-sheet structure.Such financial engineering can still create real liquidity that pays bills, funds investment or earns interest, and strong companies often secure advantageous terms from suppliers, customers and investors. But cash flow linked to those mechanisms can be more volatile, less sustainable and harder to scale than cash produced directly from selling products and services at attractive margins.
That distinction matters because Microsoft remains the most richly valued among the hyperscalers cited in the analysis, trading at 25 times forward earnings estimates. Whether the company deserves a valuation premium based on its free cash flow remains an open question as investors weigh AI spending, capital discipline and the quality of cash generation across the sector.
Our earlier report on the US AI data center boom highlighted how soaring financing costs, construction delays and power constraints are making many expansion plans less credible. We noted that only a minority of planned projects were seen as realistic, as developers face tighter capital markets, cost overruns and growing local and grid-related limitations—raising questions about whether massive AI capex can translate into sustainable returns.