Microsoft Earnings Preview: The Only Software Name With Nothing to Prove, and One Line That Can Break It

---

Read the full Microsoft Q4 FY2026 breakdown

UPDATE (July 29, 2026): Microsoft has reported. Revenue $90.01 billion, adjusted EPS $4.74, and Azure accelerating to 43% against 40.2% modelled, passing $100 billion of annual revenue. The number that moved the stock was commercial RPO of $678 billion, up 84% and up 25% excluding OpenAI. Shares rose about 8% after hours and roughly 9.5% in Thursday's session. The call below that the market was underpricing the tails at ±4% was right, and the move was more than double it.

---

TL;DR

• Microsoft reports fiscal Q4 results expected Wednesday, July 29, after the close.
• Context is everything: IBM just had its worst day since 1987 on a software miss, and the whole sector is guilty until proven innocent. MSFT is the industry's designated proof.
• The market prices only a ±4% move, the calmest big straddle of the fortnight, on a week when software prints have been producing 20% craters.
• The bull case runs through Azure. The kill risk runs through capex. Both below, with the trade.

The Witness for the Defense

July 29: the day enterprise software finds out whether IBM was an outlier or a preview.

The AI-eats-SaaS thesis got its loudest data point this week when IBM blamed its miss on software weakness and lost a fifth of its value, dragging ServiceNow, Salesforce and Oracle with it. Microsoft is the one enterprise software company the thesis has never laid a glove on, because MSFT sits on both sides of the trade: it sells the AI (Azure, Copilot) that supposedly eats everyone else's seats. If AI commoditizes software workflows, Microsoft is the commoditizer.

Which makes July 29 bigger than one company's quarter. A clean Azure acceleration says the AI-software economy has a functioning business model at its center, and the IBM crash stays contained as one legacy company's stumble. A soft MSFT print, days after IBM's, turns "software is guilty" from a scare into a sector-wide repricing that the Nasdaq cannot shrug off, because MSFT is too big to rotate around.

The Two Numbers That Decide It

• Azure growth. The single most-watched line in enterprise tech. It has carried every MSFT print for three years, and the whisper economy now demands acceleration; steady strength no longer clears the bar. Flat-to-decelerating Azure with a big capex bill is the nightmare combination this month's market would punish without mercy.
• Capex guidance against the gigawatt backdrop. Huang's $100B-per-gigawatt math made every hyperscaler's capex line radioactive. Microsoft spends more on AI infrastructure than almost anyone. The exact same dollar figure reads as "investing in the moat" if Azure accelerates and "burning cash into a cost curve going vertical" if it doesn't. Same number, two multiples.

Copilot monetization commentary is the swing factor between them: real per-seat AI revenue is the receipt that justifies the capex, and Microsoft is the only company positioned to show it at scale this quarter.

What the ±4% Straddle Gets Wrong

MSFT's implied move is priced off its own placid history: the steadiest reporter in mega-cap tech, rarely moving more than a few percent. That history was built in quarters when software wasn't on trial and capex wasn't radioactive. This print lands into both, with the sector's verdict riding on it. The distribution has fatter tails than a normal MSFT quarter, in both directions, and the market is charging normal-quarter prices for it.

The Options Angle

• Own the underpriced tails: MSFT strangle, six to eight weeks out. At ±4% implied, you're buying the cheapest event vol of the fortnight on the print with the most sector-wide consequence. Either resolution (Azure acceleration relief or capex-shock contagion) travels further than 4%.
• The directional lean, if forced: calls over puts. Microsoft has receipts (actual AI revenue) where IBM had excuses, and the earnings paradox regime pays uncrowded quality. Post-IBM, MSFT is as uncrowded as it gets for a $4 trillion company.
• The pairs expression: long MSFT, short IGV (or the SaaS name of your choice) through the print. If MSFT proves the AI-software model works, it wins the rotation within software; if the sector craters, the short leg pays. The spread wins in three of four scenarios.

The One-Line Read

IBM put software on trial and Microsoft testifies July 29 with the only evidence that matters: whether Azure turns the AI capex everyone fears into the AI revenue everyone wants. The options market is charging boring-quarter prices for a verdict-day print. Take the other side of the calm.