Financebotresearch desk研究台

MSFT › analyze

MSFT · Analyze

Technology

Date: 2026-07-15 · Price: ~$384.93 (7/14 close $390.99) · Sector: Technology (Software — Infrastructure) · Mkt cap: ~$2.86T Verdict: 🟢 BUY / Add — Conviction 8/10 · Fair value $420–500 (bear floor ~$360)

Data note: Yahoo Finance MCP was not connected this session. Figures sourced from roic.ai (FY2025 ratios/statements) + Alpha Vantage (multi-year history, real-time quote) + WebSearch (sentiment). FCF CAGR pre-FY2024 is approximate — flagged below.


The one thing that matters

MSFT is ~30% below its $551 52-wk high and 12% down YTD, its worst stretch since 2008 — not because the business broke, but because AI capex is temporarily eating free cash flow. FCF fell FY2024→FY2025 ($74.1B → $71.6B) even as net income grew +15.5%, because capex jumped from $44.5B to $64.6B. Last quarter (Q3 FY2026) FCF collapsed to $15.8B against $46.7B operating cash flow — capex is now 66% of OCF, heading toward a ~$190B calendar-2026 guide.

This is the whole debate: the market is pricing the capex it can see and discounting the AI monetization it can't yet model. The business quality is not in question; the return-on-capex timing is.


1. Fundamentals — Health Scorecard

Metric Value Read
Revenue (FY25) $281.7B +14.9% YoY
Revenue TTM $318.3B Q3 FY26 rev +18.3% YoY
Net income (FY25) $101.8B +15.5% YoY
FCF (FY25) $71.6B ↓ from $74.1B FY24 — capex drag
Operating cash flow (FY25) $136.2B +14.9%; healthy
Capex (FY25) $64.6B ↑ from $44.5B (+45%)
Gross margin 68.8% Elite, but guided toward ~63% FY27
Operating margin 45.6% Best-in-class
Net margin 36.1%
Debt/assets 9.8% Fortress balance sheet
Net debt/EBITDA 0.08x Essentially net-cash
Interest coverage (EBITDA/int) 68x Immaterial leverage
Shares out (diluted) 7,465M Slowly falling (~$18.4B buybacks FY25)
Book value/share $55.78

Growth trajectory (CAGR): | | 5yr | 8yr | |---|---|---| | Revenue | 14.5% | 14.3% | | Net income | 18.1% | — | | FCF | ~9.6%* | — | | Dividend/share | 9.7% | ~10% |

*FCF CAGR lags revenue/NI badly — this is the capex story in one number. Pre-FY2024 base (~$45B FY2020) is approximate.

Capital allocation of FY25 cash: growth capex $64.6B (dominant, AI datacenters) → buybacks ~$18.4B → dividends $24.1B → M&A $6.0B → debt paydown $3.2B. Reinvestment-first, appropriate for a company earning 28% on capital. Dividend + buyback are comfortably covered even with capex surging (payout ratio 23.6%).

Verdict: Balance sheet is a fortress; top/bottom line compounding at mid-teens+; the single blemish is FCF suppression, which is a choice (growth capex), not distress. Clean bill of health.


2. Moat — Wide & Widening

Metric Value
ROIC 28.2%
ROE 32.8%
ROA 18.0%
Gross margin trend Stable ~69% (near-term dip to ~63% guided as AI infra scales)

Moat sources, all present and reinforcing: switching costs (Windows/Office/Azure/Entra lock-in across enterprise IT), network effects (Teams, M365 ecosystem, GitHub), cost advantage / efficient scale (hyperscale infra), intangibles (brand, OpenAI relationship, distribution).

Adversarial stress-test: "Well-funded rival attacks Microsoft." Azure faces genuine pressure from Google Cloud (grew faster last quarter) and AWS — this is the one live front. But displacing the M365 + Windows + Azure enterprise bundle requires ripping out entrenched identity, productivity, and compliance rails — no rival can do that at scale. The AI layer strengthens the moat: $37B AI run-rate (+123% YoY), Copilot embedded across the installed base gives Microsoft a distribution advantage OpenAI-native startups can't match.

Evergreen rating: Yes. Enterprise software + cloud + AI infrastructure is a forever business, and MSFT sits at the center of it. Near-term margin compression from AI capex is the cost of defending and extending the moat into the AI era — reasonable risk/reward.


3. Valuation — Modestly Undervalued (with real risk)

At ~$385:

Model Output Weight Note
P/E (TTM 16.51) 23.3x High Below its own 5yr average; forward P/E ~20x
Bogle expected return ~14%/yr High 0.9% yield + ~13% earnings growth, flat multiple
FCF yield 2.5% trailing / 4.8% on OCF Med Trailing FCF artificially depressed by capex
PEG ~1.2 Med Fair for quality mega-cap
Graham √(22.5·EPS·BVPS) $144 Low Useless for asset-light 7x-book compounder — ignore
DYT Yield 0.91% vs ~0.8–1.1% band Low At high end of a tiny range — mild positive, too small to lean on

Fair value range: $420–500 (forward P/E 22–28x on ~$18–19 FY27 EPS), midpoint ~$465. Current $385 sits below the range → attractive entry for a business of this quality. Bear floor ~$360: if FY27 gross margins compress to 63%, FCF stays suppressed for 2–3 years, and Azure growth decelerates toward the mid-30s%, a 20x multiple on flatter earnings gets you here.

Consensus is lopsidedly bullish: 13 Strong Buy / 41 Buy / 3 Hold / 0 Sell, avg target $559.86 (~45% upside). The Street sees the selloff as an overshoot.


4. Sentiment & Intelligence

  • Selloff drivers: (1) ~$190B CY2026 capex guide spooking FCF-focused investors; (2) CFO/analyst commentary flagging FY27 gross margin down ~4.5pts to ~63%; (3) Azure growth optics vs. Google Cloud.
  • Offsetting positives: Azure grew 40% constant-currency (beat ~39% expected); AI run-rate $37B, +123% YoY; Nadella framing capex as demand-constrained, not speculative.
  • Positioning: 75.7% institutional. Consensus bullish; the disagreement is timing of capex payoff, not the thesis.

5. Synthesis — Weighted Verdict

Tension: Valuation/Fundamentals (FCF suppressed, margins guided lower, near-term) vs. Moat/Sentiment (Azure+AI compounding, moat widening, long-term). For a wide-moat mega-cap compounder, the framework weights Moat + long-term Fundamentals over a single depressed FCF year. The capex is discretionary growth spend by a company earning 28% ROIC with a net-cash balance sheet — the opposite of distress.

The market has handed a ~30% discount on a business still compounding revenue mid-teens and net income high-teens, for a reason (capex timing) that is real but self-inflicted and reversible. That is the "great company, temporarily cheaper" quadrant of the framework — the closest MSFT has traded to value in years.

🟢 BUY / Add — Conviction 8/10.

Key risks (named): 1. Capex-to-return lag — if AI monetization underdelivers, FCF stays depressed and the multiple de-rates further. 2. Gross margin compression — the guided drop to ~63% is real; watch whether it stabilizes or keeps sliding. 3. Azure deceleration — the one live competitive front (Google Cloud). Growth dropping below ~35% would dent the thesis. 4. Multiple risk — still 20x forward; not a deep-value margin of safety, so a broad AI-sentiment unwind could overshoot to the bear floor.


A portfolio-specific passage was removed from the public build.