We value MSFT at $549 a share. The report's $529.78 price (Oct 26, 2026) was 4% below fair value: Hold.
Restated Oct 26, 2026 (method case-spread-2026-10-24); the report's original value was $549, Buy.
- Five-year return
- 11.0%
- a year, at the report's price
What would change the view
- FY2029 capex guided or tracking at FY2028's level or above with no lift to the revenue outlook: the FY2028 plant per dollar is lasting, not gliding back; keeping it for good is $489 (Hold)
- Capex per dollar of next year's revenue about 4.2% above our path in every year (10-K and 10-Q PP&E notes): the value reaches the price; 10% heavier is $503
- Azure growth of 40% or more in any FY2028 quarter: our FY2028 +29% is too low; Azure +3 points a year is $572
- Azure growth 2.9 points a year below ours over FY2028-32: the value reaches the price
- Commercial RPO growth excluding OpenAI below 15% in two consecutive quarters, or OpenAI's receivable above $10B: weight the pessimistic case ($298)
- Depreciation as a share of average net PP&E staying near 13.5-14% through FY2028 (10-Q notes) as server lives hold at six years: our 18.3% steady state is too high; at 14.2% the central is $609
- Microsoft Cloud gross margin below 62% in any FY2027 quarter: depreciation is arriving faster than we model
- A filed change to the OpenAI revenue share, the Azure commitment or the equity stake: revalue the after-tax stake and the revenue share
Catalysts
- Q1 FY2027 results after the close (company-announced 2026-10-07): capex including finance leases against our $53B and any FY2027 or FY2028 direction (capital 4.2% heavier than our path every year takes the value to the price; each further 10% is about -$46 a share); Azure against our 46%, FundaAI's 47% and the 44-45% constant-currency guide (one point carried as a level is about +$1.91); the Q2 Azure guide; commercial RPO
- Q2 FY2027 results (expected, not confirmed): whether Azure accelerates in the first half as guided, and the first word on FY2028 capex
- Q4 FY2027 results and the FY2028 outlook (expected, not confirmed): FY2028 capex against our $205B and FY2028 revenue growth against our 16.5%
This text argues the original call. The current figures are at the top.
Microsoft is a Buy: fair value $549 against $529.78, 3.7% above the
price. Nothing about the business or the forecast changed since msft-4
($541) this morning. One input changed: the long-run return on new capital
that the long fade uses after FY2031. Our four hyperscaler reports set it
four ways; the owner approved one rule for all of them (valuation.md,
study 2026-10-26-hyperscaler-return-rule).
The rule. Each case earns, on growth after FY2031, its own year-five return on its net capital: year-five after-tax operating profit over FY2026 net property and intangibles ($331.7B, 10-K) plus the case's FY2027-31 capital spending less depreciation. Growth then needs the same capital per revenue dollar the case reaches in FY2031, 1.1 times revenue. The returns are 27.5%, 32.2% and 32.7% for the pessimistic, central and optimistic cases (output.txt), floored at the 9.92% WACC and capped at Microsoft's median five-year return on the same measure, 144%; neither binds. msft-4 used 15%, 22% and 25%, set by hand below each case's own 2027-31 build return (24-29%), which assumed capital per revenue dollar kept rising after FY2031. Microsoft's history (62-317% over FY2013-25) was earned on 0.1-0.8 times revenue of capital and does not stand in.
| Step | Central | Change |
|---|---|---|
| 2026-10-22 report, valued 2026-10-22 | $544.90 | |
| Same model, valued 2026-10-26 | $545.48 | +$0.58 |
| FY2028 capex held at FY2027's $205B, and the plant per dollar it reaches kept | $543.45 | −$2.03 |
| Optimistic case the same at $220B | $543.45 | 0 |
| Mature return on new capital at each case's volume-only return (central 22%, optimistic 25%): msft-3 | $481.96 | −$61.49 |
| The held FY2028 level glides back onto each case's path by FY2030: msft-4 | $540.98 | +$59.02 |
| Each case's mature return holds its own FY2031 capital per revenue dollar: this report | $549.26 | +$8.28 |
(output.txt, each step a full rerun.) The capital rule from msft-4 stands:
FY2028 capex is held at FY2027's $205B and glides back onto each case's own
path by FY2030. Central capex including finance leases: $205B, $205B,
$135B, $126B, $186B, $203B for FY2027-32.
1. The judgment range, and which rating each choice gives
Central fair value from output.txt; the rating uses restate.view() at
$529.78 (Buy at or above $529.78, Avoid below $450.31).
| Choice | Central capex FY2027-32 ($B) | Central | Rating |
|---|---|---|---|
| msft-3's floor: the FY2028 plant per dollar kept for good | 205, 205, 171, 190, 210, 230 | $489 | Hold |
| No unannounced cut in FY2028 or FY2029, mature return set at 19% | 205, 205, 205, 67, 186, 203 | $536 | Buy |
| Gradual normalisation, a reviewer's path (not guidance) | 205, 205, 170, 170, 186, 203 | $542 | Buy, 2.3% above |
| No unannounced cut in FY2028 or FY2029 | 205, 205, 205, 67, 186, 203 | $548 | Buy |
| This report: FY2028 held, back on the path by FY2030 | 205, 205, 135, 126, 186, 203 | $549 | Buy, 3.7% above |
| FY2028 brought forward: back on the path from FY2029 | 205, 205, 99, 157, 186, 203 | $550 | Buy |
| Three-year glide: on the path only in FY2031, the exit year | 205, 205, 147, 147, 159, 203 | $595 | Buy (rejected) |
| Depreciation reaches 14.2% of net PP&E (our Alphabet report's rate) instead of 18.3% | 205, 205, 121, 104, 158, 173 | $609 | Buy |
Every capital choice we judge defensible gives a Buy, by 1.2-3.8%; only the permanent floor gives a Hold. We reject the three-year glide: it puts the lowest capex in FY2031, the one year the FCF exit capitalises, and so inflates the value by $46.
How thin the Buy is. Capital 4.2% heavier than our path in every
year, Azure growth 2.9 points a year below ours, or an exit at 32.6x
free cash flow instead of 34.8x each takes the value to the price
(market.txt). Wednesday's capex figure and its direction matter more
than Azure.
2. Forecast and value
Central case (output.txt): revenue $393.4B in FY2027 and $644.3B in
FY2031. Capex including finance leases is 25.8%, 21.6% and 28.8% of
revenue in FY2029-31. FY2031 free cash flow on the vendor's basis $209.2B, at
34.8x EV/FCF plus the $110.6B bridge for the outside view.
Sensitivities (output.txt, each alone): Azure growth ±3 points a year
−$20.2 / +$22.5; Agents and Infra margin ±1.5 points −$19.6 / +$19.6;
EV/FCF exit 31x / 38x −$34.0 / +$28.7; mature return 15% −$17.0 and 30%
−$2.1; capital 10% heavier −$46.1; OpenAI stake halved −$7.2.
msft-3's sections 2-5 and 7 still stand as written
(../2026-10-26-msft-3/report.md): Anthropic's S-1 minimum, FundaAI's
preview, the rest of the counter-case, the Q1 FY2027 preview (revenue
$92.65B, EPS $4.87 excluding OpenAI, capex about $53B) and the review
answers. Where msft-3 or msft-4 states a dollar effect, this report's
figures replace it.
3. Why the other side could be right
For Hold: AI plant may stay heavier per dollar of revenue than any earlier build, because accelerators wear out in six years or less and inference demand keeps rising; then the FY2028 level is the new normal, and the floor's $489 is right. The long-run return could also sit below the case's own: at msft-4's 22% the central is about $541. Our 18.3% depreciation is a computed steady state, not a measured rate; if Microsoft's plant wears out faster still, the Buy goes. For a wider Buy: depreciation at Alphabet's 14.2% gives $609, and Azure growth 3 points a year above our path gives $572.
4. Catalysts and falsifiers
- 2026-10-28, Q1 FY2027 results after the close: capex including finance leases against our $53B and any FY2027 or FY2028 direction; Azure against our 46%; the Q2 Azure guide; commercial RPO.
- About 2027-01-27, Q2 FY2027: the first word on FY2028 capex.
- About 2027-07-28, Q4 FY2027 and the FY2028 outlook.
- Falsifiers: FY2029 capex at or above FY2028's with no lift to the revenue outlook (the floor, $489, Hold); capex per dollar of next year's revenue 4.2% above our path every year (the price); Azure growth 2.9 points a year below ours over FY2028-32 (the price); commercial RPO growth excluding OpenAI below 15% for two quarters (toward the pessimistic $298); depreciation staying near 13.5-14% of net PP&E through FY2028 as server lives hold at six years (toward $609).
- Anthropic tripwires (added by section 6). Microsoft's Anthropic
minimum, $44.9B from November 2026 to May 2033, has no true-up and can
be cancelled tranche by tranche for undelivered capacity (S-1 p. 171),
so a lab under strain shows here as slower RPO, not as a stake loss
(Microsoft's stake is not disclosed). None moves the central by
itself; each counts toward the RPO falsifier above.
- V2. Anthropic's IPO prices, or trades for 30 days, below $58.90, the Series H price a common share.
- V3. Broadcom's notes to Anthropic step from 9% to 13% paid in kind (they do when Anthropic misses revenue conditions; S-1 p. 167), or Amazon's or Google's facility draws stall on missed milestones.
- V4. Vendor money rises above 25% of the new compute Anthropic
signs; about 20% today (
../2026-10-26-googl-5/funding.txt, our arithmetic from S-1 p. 164-169 and F-64).
5. Limits
- Depreciation and exit are not yet one rule across the three companies. Microsoft's 18.3% steady depreciation and 20-year long life stay its own, against Alphabet's 14.2% and 25 years; Microsoft's exit capitalises free cash flow, Alphabet's and Amazon's operating profit. At Alphabet's rate Microsoft is worth $609.
- The long-run return holds FY2031's capital per revenue dollar for good; that is a judgment, the case's own level rather than a rising or a falling one.
- The Buy is within 3.7% of the price: a small move in the price or in any one input changes the label.
6. Funding the build (msft-4 amendment, unchanged)
Added to msft-4 the same day, before the 2026-10-28 print; it changes no
fair value, rating or forecast and stands as written. The arithmetic
is in ../2026-10-26-googl-5/funding.py (output funding.txt there).
Microsoft still pays for its build out of operating cash flow. Fact (10-K FY2026 less the 10-Q to March 31, 2026): April-June 2026 operating cash flow $55.4B, additions to property and equipment $35.8B, 65% of it; 66% with $0.9B of finance-lease principal; 74% counting the $5.1B of assets taken on finance leases instead, which is closer to Microsoft's "capex including finance leases". In fiscal 2026 Microsoft issued no debt or stock, repaid $3.0B of debt, bought back $22.3B of stock and paid $26.4B of dividends; cash and short-term investments fell from $94.6B to $76.8B (10-K). There is no gap to fund.
Q3 and the fund's bear reading (two or more of Microsoft, Alphabet,
Amazon and Meta with Q3 capex above operating cash flow, with debt or
stock funding the gap). Meta is already there (Q3 capex 103.5% of
operating cash flow, 106.5% with finance leases; 10-Q, 2026-10-22).
Alphabet, the same evening, and Amazon, the next day, are likely to join
it (funding.txt). Microsoft is unlikely to: our $53B of capex including
finance leases for Q1 FY2027, less a finance-lease share of 13-32% (the
shares in FQ4 and FQ1 of fiscal 2026), is $36-46B of cash capex against
operating cash flow we infer at $50-60B (FQ1 FY2026 $45.1B, FQ4 $55.4B),
60-92%; and it raised no debt or stock in fiscal 2026. The numbers that
would put Microsoft in: Q1 FY2027 "Additions to property and equipment"
above "Net cash from operations", and a "Proceeds from issuance of debt"
or stock line that is not zero. Even on Microsoft's own basis, including
non-cash finance leases, $53B sits inside the $50-60B range we infer for
operating cash flow; that basis is not the reading's, and the funding test
would still fail.
Revision history
| Report | View | Pessimistic | Central | Optimistic | Central change | Price | Blinding |
|---|---|---|---|---|---|---|---|
| 2026-10-15-msft | hold | 249 | 497 | 715 | first | 539.45 | partly unblinded |
| 2026-10-22-msft-2 | buy | 357 | 545 | 640 | +48 | 521.08 | unblinded |
| 2026-10-26-msft-3 | hold | 292 | 482 | 737 | −63 | 529.26 | unblinded |
| 2026-10-26-msft-4 | buy | 292 | 541 | 793 | +59 | 529.78 | unblinded |
| 2026-10-26-msft-5 | buy | 298 | 549 | 801 | +8 | 529.78 | unblinded |
Sources
- msft-3's sources (
../2026-10-26-msft-3/report.md), unchanged: Microsoft's 10-K FY2026, Q4 FY2026 call (2026-07-29), 8-K of 2026-09-02; Anthropic's S-1; FundaAI; market data and Robinhood market data. - The long-run-return rule:
valuation.md("Mature return: the case's own capital held") and the study 2026-10-26-hyperscaler-return-rule; Microsoft's 10-K FY2026 balance sheet for net property ($313.1B) and intangible assets ($18.6B). - The shared capital rule: msft-4,
../2026-10-26-googl-5/and../2026-10-26-amzn-5/. - Robinhood quote, 2026-10-26 16:16 UTC.
- Section 6: Microsoft's 10-K FY2026 and 10-Qs to 2025-09-30 and
2026-03-31 (SEC EDGAR, pinned in
data-manifest.json); the other companies' sources listed in../2026-10-26-googl-5/funding.py.
This work was done with Claude, a model made by Anthropic, whose S-1 msft-3 reads: a conflict of interest.
Details
- Report date
- Oct 26, 2026
- Price source
- Robinhood last regular-session trade 2026-10-26 16:16:14 UTC (official close 2026-10-23: $524.51, sip-list-exchange-close), Oct 26, 2026, 9:16 AM PDT
- Bear / bull
- $298 / $801
- Exit price
- $775 (on Jun 30, 2031)
- Confidence
- 31%
- Price against our scenarios
- The price is justified in over 50% of our scenarios4,000 simulated outcomes, each drawing the uncertain drivers from ranges we chose: our judgement, not a measured chance
- Worlds
- 3 stories about the businessEach outcome first picks one of these stories by its weight, then draws that story's own driver ranges; the answer above is their mix.
- World: base (50%)
- the base drivers. The price is justified in over 50% of its scenarios; middle outcome $542, 10th to 90th $301 to $1,040Weight: Our judgment, matching the central case's 50% weight.
- World: bear_lean (25%)
- Azure slows as OpenAI moves compute elsewhere and the build stays heavy. The price is justified in 20-50% of its scenarios; middle outcome $463, 10th to 90th $257 to $883Weight: Our judgment, matching the pessimistic case's 25% weight.
- World: bull_lean (25%)
- contracted backlog is consumed on schedule and component prices ease. The price is justified in over 50% of its scenarios; middle outcome $584, 10th to 90th $325 to $1,104Weight: Our judgment, matching the optimistic case's 25% weight.
- What moves the answer
- the exit price (market outcome) 85%How much of the spread in value each driver explains on its own; drivers that move together share credit, so the shares need not add to 100%.
- Leaning cases
- 2 other views we ranThe same simulation with some ranges moved to another view, to show how much the answer depends on our ranges.
- Case: bear_lean
- Azure slows as OpenAI moves compute elsewhere and the build stays heavy. The price is justified in 20-50% of this case's scenarios; middle outcome $463, 10th to 90th $257 to $883
- Case: bull_lean
- contracted backlog is consumed on schedule and component prices ease. The price is justified in over 50% of this case's scenarios; middle outcome $584, 10th to 90th $325 to $1,104
- Cost of equity (original)
- 10.1%
- Original figures
- fair value $549 (bear $298, bull $801), exit $775; Buy
- Street
- average $598, median $600, low $440, high $725; 29 targets in the last quarter; Street consensus targets, Oct 26, 2026, 12:13 PM PDT
- Restated
- Oct 26, 2026, method case-spread-2026-10-24; forecast.json valued by method.py at market inputs 2026-10-01
- Cost of equity calibration
- +0.0 points
- Calibration source
- none: the 1-point cut set 2026-10-16 to pull the batch toward the Street was removed on 2026-10-18 (Dong); the cost of equity is risk-free plus beta times the premium
- View rule
- strong buy when the price is at or below 0.80x fair value; buy when at or below 0.90x fair value; avoid when fair value is below 0.85x the price; otherwise hold
- Blind step
- Unblinded; the model was committed in e107a9e before the record
Original call, Oct 26, 2026 (before the Oct 26, 2026 restatement)
A Buy: fair value $549 a share against $529.78, 3.7% above the price; weighting the pessimistic, central and optimistic cases 25/50/25 gives $550. This revises this morning's thin Buy at $541 (msft-4). Nothing about the business or the forecast changed. One input did: the long-run return on new capital after FY2031. Our four hyperscaler reports set it four ways, and the owner approved one rule for all of them: each case earns, on growth after FY2031, its own year-five return on its net capital, so growth needs the capital per revenue dollar the case itself reaches in FY2031. For the central that is 32.2%, against 22% set by hand (+$8.28). The answer still turns on capital judgments: keeping the FY2028 plant per dollar for good gives $489 (Hold), a gradual reviewer path $542 (Buy), 14.2% depreciation instead of our 18.3% $609 (Buy). Capital 4.2% heavier than our path, or Azure growth 2.9 points a year lower, takes the value to the price. Low confidence.
MSFT — Microsoft Corporation — research view (market: Buy)
As of 2026-10-26 · unblinded: a same-day revision of msft-4 under one long-run-return
rule the owner approved for MSFT, GOOGL, AMZN and META (assumptions.md). Model
committed before this report (e107a9e). Revises 2026-10-26-msft-4 ($541, Buy),
which stays on the record.
Fair value today $549 per diluted share (cost of equity 10.08%). Weighting the
pessimistic, central and optimistic cases 25/50/25 (our weights)
gives $550
Judgment range $489 to $609 on the capital and depreciation judgments (section 1);
Buy at or above $529.78, Avoid below $450.31
Price $529.78, Robinhood last trade 2026-10-26 16:16 UTC (2026-10-23 close
$524.51) → fair value 3.7% above: a Buy
Five-year exit price $775 on 2031-06-30: the mean of the long fade ($558) and the outside
view on free cash flow ($992); at 10/20 years $756/$791
Five-year return 11.0% a year from $529.78 to the central exit, payouts included
Bear / bull today $298 / $801, each one consistent business outcome
Confidence 31% of our simulated values fall within ±20% of $549 (odds.json;
subjective, not calibrated)
Moat Wide: Microsoft 365 is the default work suite and Azure a top-two cloud;
the open question is the capital each new dollar of revenue needs
Market expects A little less than our path: plant per dollar of revenue 4.2% heavier
than ours, or Azure growth 2.9 points a year below ours, takes the value
to the price
Price needs any one of: capital +4.2% every year; Azure growth −2.9 points a year in
FY2028-32; an exit at 32.6x EV/FCF against our 34.8x; Agents and Infra
margin −1.5 points
Odds in our scenarios P(price justified): 50%, the share of simulated scenarios with fair value
at or above the price (odds.json; subjective; simulated mean $609)
Key driver Plant per dollar of revenue: keeping the FY2028 level for good is −$60;
10% heavier every year is −$46. Depreciation: +$59 at 14.2% of net
PP&E instead of 18.3%
Street $597 average target, 59 analysts, Robinhood, updated 2026-10-03 (read
2026-10-23, not refreshed; an expectations reference, not validation)