A Great Company Isn't a Great Stock: I Ran Microsoft's Fair Value Myself

TL;DR Microsoft's real price isn't $400 a share — it's a $2.95 trillion market cap. Running 7/10/13% revenue growth, 34/37/40% margins, and 20/23/26x exit multiples, the midpoint comes out at $550 at a 9% required return and $350 at 15%. Knowing a company is great and knowing the price is great are two completely different skills.

The real price isn't $400

Let me redefine the number on your screen before anything else.

Microsoft trades near $400 a share. That is not the real price. The real price is a $2.95 trillion market cap. The share price is just that figure divided by the share count. Get the order backwards and your entire sense of valuation goes with it.

Next I look at enterprise value: $3.15 trillion. The roughly $200 billion gap is net debt. That sounds enormous, but everything is relative. This company generated $73 billion in free cash flow last year — and that's free cash flow already beaten down by capex. Even at that level, it could retire all of its debt in under three years.

$125 billion in net income vs $73 billion in free cash flow

Last year's net income was $125 billion. Free cash flow was $73 billion. That gap is the first key to understanding this company right now.

Normally I get nervous when net income runs far above cash flow. Earnings can be shaped by accounting choices; cash is harder to dress up. Here, though, the cause is identifiable: capital expenditures.

Capex hits free cash flow in full on the day it's spent. It reaches net income slowly, spread over years as depreciation. So during a heavy investment phase, these two numbers are supposed to diverge. The real question is whether the investment eventually earns a return — and that's answered by time, not by accounting.

Return on capital, margins, growth — the underlying fitness

Return on capital has averaged close to 22% over the last five years. Last year alone it was 14%, largely because free cash flow took the capex hit.

What I like most is the margin trend:

• 10-year net margin: 34%
• 5-year net margin: roughly 37%
• Trailing one-year net margin: above 39%

Revenue is growing and the share of it that survives to the bottom line is growing too. Scale usually compresses margins. Going the other direction is a signal about business structure.

I also checked whether growth was bought. Microsoft spent $110 billion on acquisitions over the last five years — not a small figure. Yet revenue growth over 10-year, 5-year, and 3-year windows all sits in the 13–15% band. Acquisitions weren't driving all of it, and I weigh that heavily.

The dividend yields 0.88%, which looks negligible. But because of the company's size, that still consumes $25 billion a year. A low yield doesn't mean a light dividend burden.

Six checks, two X's on the eight pillars — and an X isn't a verdict

Running the checklist, six items pass: return on capital, shares outstanding, cash flow growth, net income growth, revenue growth, and debt.

The two that fail are the five-year average P/E and the five-year average price-to-free-cash-flow. Both are valuation metrics.

Here's the interpretation that matters. An X doesn't mean the company is expensive. It means the price sits on the higher side — and if revenue and profit growth can justify it, that's fine.

Take two companies both trading at 30x earnings. One grows 30% a year, the other 5%. The expensive one is the 5% grower. For the 30% grower, that 30x looks cheaper with every passing year. A P/E tells you nothing standing alone.

What analysts are expecting

Before plugging in my own assumptions, I looked at consensus.

Analysts see earnings per share going from $17 to $40 over the next seven years — roughly two and a half times, north of 10% annually. Revenue is modeled from $335 billion to $760 billion, again more than doubling and again above 10% a year.

So the market is still optimistic. I kept in mind that consensus sits near the upper end of my own range as I ran the numbers.

The assumptions I used

A lot of people say you must fully understand the story before touching the numbers. I disagree. I get a rough read on the story, a rough read on the numbers, and put them into the model early. If today's price lands anywhere near the output, I spend more time. If the stock is $400 and the model says it's worth a dollar, there's nothing left to research. Move on.

Here's what I assumed for the next ten years:

Assumption	Conservative	Base	Optimistic
Annual revenue growth	7%	10%	13%
Net margin / FCF margin	34%	37%	40%
Exit P/E and P/FCF in year 10	20x	23x	26x

On that exit multiple: the long-run market average is around 15–16x. But you have to pay a premium for genuinely good businesses, and I consider Microsoft one. That's why I used a band above the market average. Push the exit multiple to 30x or 35x and you can manufacture any answer you want — that's the most common way this kind of analysis quietly breaks.

The output: $550 at 9%, $350 at 15%

First I ran it at a 9% required return with no margin of safety. That isn't me saying I'd accept 9%. It's a way to see roughly what the business is worth before I apply any discount for my own comfort.

Required return	Low	Middle	High
9% (no margin of safety)	$360	$550	$822
15% (my own hurdle)	$234	$350	$515

The $550 midpoint implies roughly a 13% annual return from today's ~$400 price — dividends included.

My personal hurdle is different. I own real estate and operate businesses, so taking on the hassle of individual stocks only makes sense at outsized returns. I use 15%, and at that hurdle the midpoint lands at $350. My watchlist price is $345.

One thing worth stating plainly: there is no correct required return. Twelve percent is right for some people, twenty for others. It depends on your situation and how well you understand the business. But one implication is firm — if 9 or 10% would satisfy you, don't buy individual stocks. Buy a low-cost index fund. Individual names only make sense when you believe something is genuinely mispriced and you can beat the market by a wide margin.

What this exercise actually does for you

The value here isn't predicting the future accurately. Nobody nails a margin assumption ten years out.

What it does is one thing: it stops you from overpaying for a company everyone already agrees is great.

That's what cost me money before I learned it. Knowing a company is great and knowing whether the stock is great at this price are different skills, and most investors never build the second one. They see a name they trust, see someone smart buying, and click. In a bull market that works. It's also exactly how people end up underwater for three, four, five years in the best company on earth.

Having a concrete number instead of a vibe removes most of that risk. I walked through the same procedure in Nvidia's Valuation. The company changes; the process doesn't.

Where my math could be wrong

Let me point at my own weak spots.

First, the margin band. The 34–40% range is grounded in the last decade of results, but Microsoft is in the largest capex cycle of its history. Once depreciation flows through the income statement in earnest, margins could break below the bottom of my range.

Second, the exit multiple. Nobody knows whether the market will pay 20–26x for this business in ten years. If growth settles into single digits, 15x is plausible. The exit multiple swings the output more than anything else, and it's also the assumption easiest to pick out of thin air.

Third, the growth band. I placed 7–13% between consensus and historical results. But if the cloud market matures and competition compresses pricing, even the low end could prove optimistic.

Miss on any one of these and my $350 changes. So treat it as a baseline derived from my assumptions, not an answer. Your assumptions will produce a different number — that's normal, and having a number at all is the point.