Amazon approaches its October 29 earnings report with a measurable cash hurdle. Its second-quarter filing shows $71.419 billion of operating cash generated from January through June 2026 and $98.411 billion of gross property-and-equipment purchases. The difference is a $26.992 billion shortfall. This is TECHi’s cash-after-equipment calculation, not Amazon’s reported free-cash-flow figure. The same filing says Amazon invested $28.7 billion in OpenAI preferred shares during those six months, outside the equipment measure.
That distinction is the question for this week’s AI-stock earnings cycle. Microsoft reports October 28, and Amazon’s call is scheduled for October 29. The numbers to watch are operating cash, equipment spending, cloud profit and further financing. Amazon’s cloud growth can be strong while its consolidated cash demand remains larger still. TECHi’s Amazon stock page tracks the company’s market and earnings context.
Key takeaways
- Amazon’s January–June operating cash fell $26.992 billion short of gross equipment purchases. Microsoft retained $35.442 billion and Alphabet $4.261 billion under the same measure.
- Amazon’s $28.7 billion OpenAI and $10.0 billion Anthropic preferred-stock investments were separate investing cash outlays, not equipment spending.
- Amazon needs a July–December surplus greater than $26.992 billion on this measure to finish calendar 2026 positive. That threshold is TECHi arithmetic, not company guidance.
A common cash-flow test for three AI stocks
A comparison should hold both the calendar and the accounting definition steady. We used January 1 through June 30, 2026 for all three companies, subtracted gross purchases of property and equipment from operating cash flow, and rounded only after the subtraction. For Microsoft, whose fiscal year ends in June, the first-half total adds its March-quarter and June-quarter cash-flow lines. Amazon and Alphabet provide six-month totals in their respective Amazon and Alphabet filings.
| Company | Operating cash | Gross equipment purchases | Cash after purchases | Coverage |
|---|---|---|---|---|
| Microsoft | $102.120B | $66.678B | +$35.442B | 1.53× |
| Alphabet | $84.859B | $80.598B | +$4.261B | 1.05× |
| Amazon | $71.419B | $98.411B | −$26.992B | 0.73× |
The resulting cash left after gross equipment purchases was positive $35.442 billion for Microsoft, positive $4.261 billion for Alphabet, and negative $26.992 billion for Amazon. Operating-cash coverage of equipment purchases was 1.53 times, 1.05 times and 0.73 times, respectively. These ratios describe the group’s cash burden; they do not isolate AI returns. Amazon also buys fulfillment equipment, while Microsoft’s and Alphabet’s facilities support many products beyond generative AI.
The year-earlier comparison reveals the direction of travel. For January–June 2025, the same calculation yields $45.867 billion left at Microsoft, $24.254 billion at Alphabet and a $7.672 billion shortfall at Amazon. Microsoft’s cushion narrowed by $10.425 billion; Alphabet’s narrowed by $19.993 billion; Amazon’s deficit widened by $19.320 billion. Each is spending more aggressively relative to operating cash, even though all three continue to generate substantial cash from customers.
This matched-period test extends TECHi’s earlier Microsoft-versus-Amazon funding comparison by adding Alphabet and a common calendar window. It is more useful than dividing one company’s annual cloud-segment income by another’s six-month capital investment. It also avoids presenting a cloud segment’s profit as the return on an entire corporation’s data-center estate. Investors can reproduce every row from the linked cash-flow statements.
What the Amazon gap leaves out
The $26.992 billion figure is operating cash minus gross equipment purchases. It is deliberately narrower than total investing cash flow and differs from Amazon’s own free-cash-flow definition, which nets proceeds from equipment sales and incentives against purchases. On a trailing-twelve-month basis through June, Amazon reported negative $7.604 billion of free cash flow. The two figures refer to different periods and methods; they cannot be substituted for one another.
The filing adds investment lines that a simple “capex versus cash” chart would miss. Amazon says it paid $28.7 billion for OpenAI Series C preferred shares during the first half, including $13.7 billion in the June quarter. It also says it invested $10.0 billion in Anthropic nonvoting preferred stock during that quarter. These are equity investments, not purchases of servers. They sit outside the $98.411 billion equipment line and the $26.992 billion calculation. They should not be added to the deficit and relabeled as free cash flow, because investment assets may have future value and their accounting and financing differ. They do show that the AI strategy demands capital on more than one line of the cash-flow statement.
Amazon says it subsequently funded the remaining $21.3 billion of its OpenAI commitment after June 30. That payment falls in a later period. The October report should help readers see whether cloud cash generation is beginning to catch up while those later commitments are being funded.
The strongest counterargument is that Amazon is deliberately buying scarce capacity and strategic positions ahead of contracted demand, so a negative six-month cash balance can precede strong long-term returns. The current filing cannot settle that claim. Rising AWS operating income, improved utilization, a slower rate of equipment growth and a narrowing consolidated cash deficit across several reports would strengthen it.
AWS profit is real, but cash conversion is the test
Amazon’s six-month segment results show AWS revenue of $79.819 billion and operating income of $30.782 billion. In the June quarter alone, AWS operating income was $16.621 billion. That is a substantial profit engine and the strongest evidence for treating the buildout as an investment in future capacity rather than a cost with no customer demand.
The same filing allocates $90.120 billion of first-half net additions to property and equipment to AWS. “Net additions” can include noncash assets acquired but not yet paid for, so it is not the same as the consolidated $98.411 billion cash-purchases line in our comparison. The figures should be read together, not divided into a claimed return rate. They indicate where Amazon is building and why current operating profit alone does not describe its funding needs.
For the stock, the October 29 decision is a sequence: Did AWS sales and income rise? Did the company commit still more equipment? Did operating cash rise enough to cover that spending? If only the first answer is yes, a strong earnings headline may coexist with a weak cash verdict. If all three improve, the gap can narrow faster than a backward-looking chart implies.
Microsoft and Alphabet set different benchmarks
Microsoft offers the clearest current cash cushion. Its March and June 2026 quarters together produced $102.120 billion of operating cash against $66.678 billion of property-and-equipment additions, leaving $35.442 billion. Its June-quarter release also reported 43% Azure and other cloud-services growth. Strong cash coverage does not prove the marginal return on new AI servers; it shows the existing business funded reported equipment spending over this period.
The caution lies in the trend. A year earlier, Microsoft’s first-half equipment purchases were $33.824 billion, against $79.691 billion in operating cash. In one year, purchases nearly doubled while the residual cushion fell. The October 28 report should show whether Azure’s growth and broader cash generation can keep pace with the next step in capacity spending. Management’s guidance on depreciation, lease commitments and monetization matters alongside the capital-spending number.
Alphabet is closer to the line on this measure. Its first-half filing reports $84.859 billion of operating cash and $80.598 billion of equipment purchases, leaving $4.261 billion. A year earlier, the same cushion was $24.254 billion. Google Cloud’s June-quarter revenue rose 82% to $24.768 billion, and its operating income reached $8.814 billion. The company’s official remarks put Cloud backlog at $514 billion, a demand signal rather than cash already collected.
There is another difference for equity holders. Alphabet’s filing says it raised $49.6 billion net in common and mandatory-convertible preferred stock in June for general corporate purposes, including scaling AI infrastructure. It authorized a separate up-to-$40 billion at-the-market program, primarily intended for taxes tied to employee equity grants; no shares had been sold under that program by June 30. Treating the full authorization as completed AI financing would be wrong. Issued securities and an unused authorization have different effects on dilution and flexibility.
A low trailing price-to-earnings ratio also does not answer this cash question. Alphabet’s June-quarter filing records about $99.0 billion of net gains on equity securities, primarily unrealized, in other income. Operating cash, cloud operating income and new financing offer a cleaner view of this investment decision than an unadjusted comparison of GAAP earnings multiples.
A conditional 2027 forecast with a 2026 threshold
To finish calendar 2026 with positive operating cash after gross equipment purchases under the same measure, Amazon’s July–December surplus must exceed its first-half $26.992 billion shortfall. That is arithmetic, not company guidance. A second-half surplus of exactly $26.992 billion would bring the full year only to zero. Equipment proceeds, finance leases, acquisitions, equity investments and financing flows remain outside this measure.
In a bear case for 2027, AWS keeps growing but cash purchases and depreciation rise faster than profit and collections. Amazon’s gap persists; Alphabet’s narrow cushion slips below zero; Microsoft retains a positive group cash balance but loses more headroom. The risk is a lower valuation for capital intensity despite impressive cloud revenue.
In a base case, AWS and Google Cloud convert backlog into recognized revenue, operating cash grows, and the pace of new equipment purchases moderates. Amazon narrows its shortfall before turning positive; Alphabet protects or rebuilds its cushion; Microsoft continues to fund its program internally. This case requires evidence in successive cash-flow statements, not a single quarter of bookings.
In a bull case, utilization and software revenue rise materially on infrastructure already ordered. Amazon’s AWS income and consolidated cash receipts accelerate enough to cross the second-half hurdle; Alphabet’s Cloud growth expands its cash headroom; Microsoft’s Azure and AI products keep operating cash well above equipment purchases. The forecast would weaken if management raises capital commitments faster than demand converts into collected cash.
The ranking is conditional. Microsoft has the strongest observed first-half cash coverage. Amazon has the largest visible improvement opportunity and the highest near-term cash hurdle. Alphabet has rapid Cloud growth but a thin cash cushion and a distinct dilution question. None of those observations is a precise share-price target or a promise of stock outperformance. They identify what the next reports need to prove.
What could reverse the conclusion
Read each cash-flow statement alongside the earnings release. Compare operating cash with gross equipment purchases over the same period as the prior year; check whether cloud operating income expands; then examine new leases, debt or equity. Amazon should separate AWS performance from consolidated funding and investment cash flows. Alphabet’s backlog needs a path to revenue and cash. Microsoft’s Azure growth needs to preserve cash headroom as equipment and leases rise.
The third-quarter releases may reverse the first-half ranking. Strong AWS cash conversion would weaken the Amazon-risk case; a larger Microsoft commitment without matching collections would weaken the Microsoft-cushion case. The scheduled results deserve a fresh, sourced follow-up when the companies publish their actual numbers.