Equity learning

How to read an earnings release

A practical free lesson for equity readers

This lesson uses the same primary-source figures as Microsoft FY26: cloud growth met a much heavier infrastructure bill. It teaches a reading habit, not a view on Microsoft's share price. The example period ended June 30, 2026, and the source documents were published or filed July 29, 2026.

1. Keep the period and unit fixed

Before reading a headline, write down the period being compared. Microsoft's Q4 FY26 table covers the three months ended June 30, 2026; its FY26 table covers the full fiscal year ended the same day. The figures below are US$ millions unless shown as billions.

Do not compare a quarter with a full year, mix GAAP and adjusted figures without labeling them, or treat a contract measure such as remaining performance obligation as revenue already recognized.

2. Revenue is not profit

Line Q4 FY26 Q4 FY25 What it answers
Revenue $90,007M $76,441M How much was recognized as sales in the period?
Operating income $40,603M $34,323M What remained after operating costs, before non-operating income and tax?
Net income, GAAP $35,766M $27,233M What remained after the full GAAP income statement, including non-operating effects?
Net income, adjusted $35,286M $28,808M What does Microsoft's stated non-GAAP reconciliation show after its specified adjustment?

Revenue grew 18%, operating income grew 18%, GAAP net income grew 31%, and the stated adjusted net income grew 22%. Those are not four competing answers to one question. They describe different layers of the income statement.

3. Profit is not cash

The 10-K reports cash from operations and property-and-equipment additions in a separate consolidated cash-flow statement. For an educational diagnostic, we calculate:

cash-flow proxy = net cash from operations − additions to property/equipment

FY26 = $182.935B − $115.948B = $66.987B
FY25 = $136.162B − $64.551B  = $71.611B

The arithmetic gives a 6.5% decline in this proxy:

$66.987B / $71.611B − 1 = −6.5%

At the same time, property-and-equipment additions rose 79.6%: $115.948B / $64.551B − 1. This is a useful question about the cash burden of the build-out, not Microsoft's reported free-cash-flow metric. The proxy does not subtract acquisitions or investment purchases, does not isolate one segment, and does not predict future returns.

4. Ask what changed the profit comparison

Microsoft's stated reconciliation identifies a $4.963B FY26 net-income increase from OpenAI investments and a $3.620B FY25 net-income loss from that investment line. FY26 GAAP net income was $133.749B; the stated adjusted figure was $128.786B. FY25 GAAP net income was $101.832B; the stated adjusted figure was $105.452B.

That produces a GAAP growth comparison of 31% and a stated adjusted growth comparison of 22%. The lesson is not that one measure is “real” and the other is “fake.” It is that a reader should identify the adjustment, keep the years comparable and avoid using GAAP EPS growth alone as a complete description of operating momentum.

5. Fill in an evidence checklist

Use the exact release or filing, then leave a visible record of what is still unknown. The blank rows are intentional: this is a reusable worksheet, not a pre-filled conclusion.

Claim I want to test Exact source URL Period ended / published Reported line or formula Unknown or limitation

Before writing a conclusion, check five things:

Source trail and scope boundary

Start with the Microsoft FY26 Q4 earnings release and webcast and the Microsoft FY26 Form 10-K filed with the SEC. The earnings release supplies the Q4 comparison and reconciliation; the 10-K supplies the consolidated cash-flow statement, segment definitions and allocation limits.

This lesson does not answer whether MSFT is cheap or expensive. It contains no price target, valuation multiple, expected return, trade instruction or portfolio allocation. Microsoft's own commentary is company disclosure, not independent customer research. The source cutoff remains the FY26 period ended June 30, 2026.

For a different subject, the separate Reading crypto risk workbook covers BTC/ETH correlation, drawdowns, sample sensitivity and rebalance costs. It is distinct practice material and does not teach Microsoft earnings analysis; payments are not yet enabled.