The Weekly Pick · Report Nº 001 · MONDAY 06 JULY 2026
MicrosoftNASDAQ · MSFT
SOFTWARE · CLOUD & AI
“The Street sold a 40%-growth cloud business because the capex number had three more digits than the model expected. We think that's the opportunity.”
Microsoft is the most complete platform in enterprise software — cloud, productivity, developer tooling and, increasingly, the distribution layer for applied AI. The stock is down 28% from its 52-week high of $555.45, not because the business slowed, but because the Street flinched at a capex number: $190B of planned calendar-2026 spend, well above the ~$152B analysts had modelled. We think that reaction has it backwards.
The FY2026 Q3 print, released in the same window as this de-rate, showed Azure growing 40% in constant currency, the AI business crossing a $37B annualised run-rate (+123% year over year), and Copilot paid seats up a third sequentially to 20 million. A business accelerating this fast, priced at 23.5× trailing earnings — its lowest multiple since 2023 — is not the setup the sell-off implies.
Three segments drive the P&L, and — contrary to the popular framing of Microsoft as a 'cloud stock' — the largest by revenue is still Productivity and Business Processes (Microsoft 365, LinkedIn, Dynamics), not Intelligent Cloud. FY2025 (ended June 30, 2025) revenue split $120.8B / $106.3B / $54.6B across the three, per the 10-K. The AI thesis threads through all three, but it is Intelligent Cloud carrying the growth and, increasingly, the market's attention.
The segment split matters for quality as much as size: Productivity carries the highest operating margin of the three (58%), funded by a subscription base that renews without Microsoft having to re-win the account every quarter. That cash engine is what lets Microsoft fund a $190B capex year without touching the balance sheet's investment-grade footing.
How it makes money
Switching costs, distribution and scale compound into one of the widest moats in software. Enterprises don't rip out the stack that runs their identity, email, documents and cloud — they add to it, and every AI feature ships through a door that is already open.
Our conviction rests on three legs. Each is independently underwritten against the actual Q3 FY26 print; together they explain why we think the capex-driven de-rate is a mispricing, not a warning.
“The Street sold a 40%-growth cloud business because the capex number had three more digits than the model expected.”
FY2025 segment operating income confirms the margin story isn't just top-line: Productivity ($69.8B op. income, 58% margin) and Intelligent Cloud ($44.6B, 42% margin) both expanded, even as More Personal Computing stayed the low-margin ballast (26%). The capex debate is about the shape of FY2026-27, not the quality of what's already been delivered.
Revenue · $B
| Metric | FY25 (actual) | FY26E | YoY |
|---|---|---|---|
| Revenue | $281.7B | ~$320B | +14% |
| Segment operating income | $128.5B | ~$148B | +15% |
| Azure growth (cc) | +34% | ~39–40% | guided |
| Buybacks (FY, cash) | $13.0B | n/a | vs. $18.4B FY23 |
At 23.5× trailing earnings — its lowest since 2023 — Microsoft trades below both its own five-year average and its large-cap software and cloud peers, for a business now guiding to its fastest Azure growth in years. Even hyperscaler peers with less durable moats aren't priced this conservatively.
Forward P/E vs. reference
We publish the case against as plainly as the case for. Here is what would prove us wrong, sized against the actual FY26 numbers and current regulatory/market backdrop.
A rare setup: the fastest Azure growth in years, arriving in the same quarter the stock fell 28% off its highs on a capex number. We're buyers in the $390–$420 range with a $558 target over 12–18 months — a target that, notably, is exactly where Street consensus already sits.
THE MOAT
Enterprise IT does not switch identity providers, productivity suites and cloud tenancy on a whim. Microsoft owns all three, and AI is being delivered through the same front door — Copilot inside Office, Azure OpenAI inside the tenant, security inside Entra. Each new AI product deepens a relationship that was already the stickiest in the industry.
That is why Azure can grow 40% at its scale while segment margins expand: it is not winning greenfield price wars against Google Cloud or AWS, it is expanding wallet share inside accounts it already controls.
The one thing to remember
Microsoft is not selling AI as a new product to new customers — it is selling AI as an upgrade to the customers it already owns, with a 400M-seat head start. That is the cheapest distribution in software.
THE NUMBERS
It's a common mistake to treat Microsoft as primarily a cloud stock. By FY2025 revenue, Productivity and Business Processes ($120.8B) is still the largest segment — ahead of Intelligent Cloud ($106.3B) — and it carries the highest margin of the three at 58%. Intelligent Cloud is the growth leg and the one re-rating the multiple; Productivity is the cash engine funding the AI build without touching the balance sheet.
FY2025 revenue by segment
THE COMPETITION
Azure's 40% constant-currency growth doesn't happen in a vacuum. Google Cloud is growing even faster off a smaller base (Q1 2026 revenue +63% YoY), and Oracle's cloud infrastructure line is up 77%. AWS, the largest by revenue, is the laggard of the group at roughly 18%. The honest read: the AI infrastructure race is genuinely competitive, and Microsoft is winning share from AWS more than it is running away from the field.
| Cloud provider | Growth (recent) | Backlog signal |
|---|---|---|
| Microsoft Azure | +40% (cc) | $627B commercial backlog |
| Google Cloud | +63% YoY (Q1'26) | $460B, doubled sequentially |
| Oracle Cloud | +39% (total) | Infra +77%, apps +11% |
| AWS | +18% YoY | $244B, +40% YoY |
Read this honestly
Google Cloud and Oracle are growing faster off smaller bases — normal for scale. What matters is Azure is accelerating at Microsoft's size, which is much harder to do, and its backlog is real, contracted revenue, not a growth story yet to be booked.
Edge publishes educational research, not personalised financial advice. Reasoning is shared as opinion, never an instruction to buy or sell. Buy ranges and targets are scenarios; figures are real and dated. © 2026 The Financial View.