Free cash flow is the better first test of whether AI spending is financially paying off, but it cannot prove returns alone. Earnings show profitability; free cash flow shows how much cash remains after the capital spending needed to build AI capacity. Neither measure works in isolation. Investors should compare earnings, free cash flow, capital spending, and related revenue growth while checking how each company defines and reports those figures.
Table of Contents
- Why earnings can look strong during an AI spending surge
- What free cash flow reveals
- The revenue test for AI investment
- Why company comparisons can mislead
- A practical scorecard for investors
- Frequently Asked Questions
Why earnings can look strong during an AI spending surge
Net income measures accounting profit after expenses, but large infrastructure investments often appear gradually through depreciation. A company can therefore report rising earnings while committing substantial cash to data centers, servers, and other equipment. Microsoft's FY2026 net income increased 31% to $133.7 billion, while additions to property and equipment rose 80% to $115.9 billion, according to Microsoft's FY2026 earnings release.
The combination shows why earnings alone cannot reveal the full near-term cost of expanding AI and cloud capacity. Earnings still matter. They show whether the broader business generates profit and whether demand may support future investment. But they do not necessarily show how much cash the company must spend today to pursue that growth.
What free cash flow reveals
Free cash flow commonly means operating cash flow minus capital expenditures. It indicates how much cash remains after a company funds purchases such as servers, networking equipment, and data-center infrastructure. Using that definition, Microsoft generated about $67.0 billion of FY2026 free cash flow: $182.9 billion of operating cash flow less $115.9 billion of equipment additions.
That figure was materially below its $133.7 billion net income, according to Microsoft's reported FY2026 results. That gap does not automatically mean the spending is failing. It means the investment is consuming cash before the associated capacity fully produces revenue or profit. A falling or weak free-cash-flow figure deserves attention when capital spending rises sharply.
The revenue test for AI investment
Cash spending becomes more convincing when revenue tied to the investment grows quickly and sustainably. investors should look for evidence that customers are using the new capacity, rather than treating a large capital budget as proof of future returns. Microsoft reported 43% growth in Azure and other cloud-services revenue in the June 2026 quarter. Management also said capacity remained constrained through 2026 and that spending could precede revenue readiness, according to Microsoft's FY2026 results and April 2026 earnings call.
Alphabet offers a similar mixed picture. Its 2025 results included $132.2 billion of net income, $73.3 billion of trailing-12-month free cash flow, and $91.4 billion of capital expenditures. Google Cloud revenue rose 48% in the fourth quarter amid demand for enterprise AI infrastructure and solutions, according to Alphabet's 2025 earnings release. The practical question is whether revenue growth, pricing, utilization, and margins eventually catch up with the cash investment. A single strong quarter cannot settle that question.
Why company comparisons can mislead
Free cash flow is not perfectly standardized. Meta, for example, subtracts both equipment purchases and finance-lease principal from operating cash flow. It reported $10.62 billion of free cash flow in the third quarter of 2025 and warned that non-GAAP measures have important analytical limitations, according to Meta's third-quarter 2025 results.
Before comparing companies, investors should check: Segment reporting also limits precision. Alphabet records shared AI research and technical-infrastructure costs at the corporate level rather than allocating them to operating segments. That makes it difficult to assign a precise AI return using either earnings or free cash flow alone.
- Which capital expenditures the company subtracts.
- Whether finance leases or other cash commitments are included.
- Whether the figure covers one quarter or the trailing 12 months.
- Whether spending supports AI specifically or the broader business.
A practical scorecard for investors
Use both measures, but give each a different job: Also separate existing performance from future expectations. Alphabet projected $175 billion to $185 billion of 2026 capital expenditures. Its documented 2025 free-cash-flow result therefore shows current cash generation, not that the next investment wave has already paid off.
Amazon illustrates the same warning. Its second-quarter 2025 net income rose to $18.2 billion and AWS operating income reached $10.2 billion, while trailing-12-month free cash flow fell to $18.2 billion from $53.0 billion. Strong earnings can coexist with substantial cash pressure when infrastructure investment accelerates.
- Earnings: Is the overall business profitable?
- Free cash flow: How much cash remains after the investment?
- Capital expenditures: How aggressively is the company expanding capacity?
- Revenue growth: Are customers buying the resulting services?
- Margins and utilization: Is the new revenue becoming economically attractive?
Frequently Asked Questions
Is negative or declining free cash flow proof that AI spending is wasteful?
No. It may reflect capacity being built ahead of demand. The warning becomes stronger when spending rises without corresponding revenue, margin, or utilization improvement.
Which metric should investors prioritize?
Start with free cash flow to measure the cash burden, then use earnings and revenue growth to assess profitability and demand.
Can investors calculate an exact return on a company's AI spending?
Often not. Shared research, infrastructure, leases, and corporate costs may not be allocated cleanly to AI-related segments.