Microsoft Corporation · MSFT · NASDAQ
My estimate (a forecast, not a fact)
USD 439.18 / share
Market price, 1 Sep 2026
USD 501.02
Difference
−12.3%
What Microsoft actually is
Forget Windows and Office. Microsoft made $331.8 billion last year, and 64.6% of it is cloud. That business has compounded at nearly 25% a year. Everything else — Windows, Xbox, devices, search — compounded at 4.5%. Microsoft transitioned into a cloud company.
How price behaved
Microsoft fell more than 35% from its peak in July last year. Then the fourth-quarter results landed, and price has run 27% since.
| Price | Diff from ATH | |
|---|---|---|
| High, 28 July 2025 | $555.45 | — |
| Low, early 2026 | $356.51 | −35.8% |
| Price used in this valuation, 1 September 2026 | $501.02 | −9.8% |
In the same year it spent $116 billion on data centers — 35% of revenue, 3 times its depreciation charge — and signed another $329 billion of datacenter leases that have not started yet.
Almost none of that has earned anything so far, and the cost shows. Return on invested capital has fallen from 47.4% to 23.6% in 3 years. Revenue per dollar of capital has fallen from 3 to under 1.
The drivers behind the number
Three assumptions produce most of that $439.18.
Revenue grows 17% next year and 15% for 4 years after — slightly slower than Microsoft has actually managed. That takes Microsoft past $1 trillion in 10 years, moving it from 18.9% of its addressable market to 21%. A 2-point share gain.
A long-run operating margin of 48%. That sounds cautious for a company whose margin has risen every year for a decade, but depreciation on $116 billion of new data centers has not arrived yet, and when it does it comes out of the margin.
And a return on capital in perpetuity of 10%, against a terminal cost of capital of 8.41%. The standard assumption is that competition eventually removes every excess return. For Microsoft I think that is too harsh. But be clear how much that judgement is worth: take that point and a half away and my value falls $43.76 a share, to $395.42.
The Valuation
$3.07 trillion of operating assets, of which 67.5% is terminal value. Take off $128 billion of debt, add back $77 billion of cash and $237 billion of non-operating assets, mostly the stake in OpenAI. Divide by 7,425 million shares.
$439.18, against market price of $501.02.
The full valuation below has all 10 assumptions, what building for demand looks like, the OpenAI stake and the rest of the argument.
| Files Subscribe to download |
|---|
| Valuation model The FCFF model behind the estimate — every input, every year, and the value it produces. XLSX · 23 KB |
| Sensitivity table Value per share across the two assumptions that matter most. XLSX · 44 KB |
| Market sizes The market-size estimates behind the growth check — every source, every caveat, and the date each was last refreshed. XLSX · 51 KB |
| Financial statements Microsoft's income statement, balance sheet and cash flow as reported, with the trailing-twelve-month column the base year is built from. XLSX · 37 KB |
Sensitivity
Estimated value per share in USD, across the two assumptions that move it most. Rows are revenue in year 10 ($bn); columns are target pre-tax operating margin (%). Shaded cells clear the USD 501.02 price at publication.
| Revenue in year 10 ($bn) | Target pre-tax operating margin (%) | ||||||
|---|---|---|---|---|---|---|---|
| 10 | 15 | 20 | 25 | 30 | 40 | 50 | |
| 720.6 | 93 | 125 | 157 | 190 | 222 | 286 | 350 |
| 875 | 95 | 133 | 171 | 209 | 247 | 323 | 399 |
| 1029.5 | 96 | 140 | 184 | 228 | 272 | 359 | 447 |
| 1183.9 | 97 | 147 | 197 | 246 | 296 | 395 | 495 |
| 1338.3 | 99 | 154 | 210 | 265 | 321 | 431 | 542 |
| 1647.1 | 102 | 169 | 235 | 302 | 369 | 503 | 636 |
| 2007.4 | 105 | 185 | 265 | 345 | 425 | 585 | 745 |
Disclosure
- Position
- As at 1 September 2026, I do not own shares in Microsoft. This covers shares held by people closely associated with me. It is a snapshot: if it changes, it changes in a new valuation, never by editing this one. I hold no position, long or short, exceeding 0.5% of Microsoft's issued share capital.
- Author
- Lucian Mesaroș, independent equity analyst, sole author of this publication, publishing in a personal capacity as an individual. I am not authorised or supervised by the ASF or by any other financial supervisory authority, I am not an investment firm, and I do not provide investment advice.
- Valuation date
- 1 September 2026. Production completed 7 September 2026. First published 7 September 2026.
- Market price used.
- USD 501.02 — NASDAQ close, 1 September 2026 on NASDAQ, 1 September 2026. The percentage difference on this page compares my estimate to that price on that date, and is not updated afterwards.
- What the number is
- USD 439.18/share is my estimate of intrinsic value, produced by a discounted cash flow (FCFF) model. It is a forecast and an opinion, not a fact, not a price target I undertake to defend, and not a prediction of where the share price will go. A Damodaran free-cash-flow-to-the-firm discounted cash flow. Revenue is forecast for 10 years and faded to the risk-free rate, the operating margin converges to a long-run target, reinvestment is set by a sales-to-capital ratio, and the resulting cash flow is discounted at a cost of capital that falls to a terminal rate. R&D is capitalised over 3 years and leases are treated as debt, so the base-year margin and invested capital are adjusted rather than as reported. The full framework is at Methodology.
- Key assumptions
- Revenue growth of 17% in year 1, then 15% a year through year 5, fading to the risk-free rate of 4.53% by year 10 · Operating margin of 49% in year 1, converging to a 48% target by year 5 · Sales to capital of 0.70 in years 1–5 and 1.54 in years 6–10 · Effective tax rate of 20% rising to the 25% marginal rate by the terminal year · Cost of capital of 10.91%, falling to 8.41% in perpetuity · Return on capital in perpetuity of 10%
- How wrong this can be
- Valuations are sensitive to assumptions I chose — the sensitivity table above shows what the estimate becomes when the two that matter most turn out differently. A small change moves the estimate materially. I can be, and have been, wrong.
- Time horizon
- At least 1 year, and normally far longer. I have no view on what the share price does over the coming weeks or months, and nothing here is a short-term call. Separately, this document stands until I next value Microsoft — see What I've Published.
- Updates
- I revalue a company when it publishes annual results, or when something material changes. There is no guaranteed schedule, and no undertaking to update this page.
- Conflicts of interest
- I have no agreement of any kind with Microsoft or with anyone connected to it, I have not been paid by anyone to produce this valuation, no one asked me to write it except through the public request page, no draft was shown to the company, and I am not a market maker, liquidity provider or adviser to it.
- Previous valuations of Microsoft
- None in the last 12 months.
- Sources
- Microsoft Corporation, Form 10-K for the fiscal year ended 30 June 2026 · OpenAI, "Accelerating the next phase of AI", the March 2026 funding round · market-data.xlsx, the market-size estimates published with this valuation · Press reports of OpenAI's confidential S-1 filing, June 2026 All figures are as reported by the company unless I say otherwise.
- Everything I have published in the last 12 months:
- What I've Published