Microsoft: A 52-Week Low and the OpenAI Stake the Market Isn't Pricing In

TL;DR Microsoft dropped 21% over the past year under ~$190B of capex, hitting a 52-week low of $349 in late June and now trading around $390. Its forward P/E has compressed from a 5-year average near 30 to about 23. But there's a card the price isn't reflecting — a 27% stake in OpenAI, which has confidentially filed for an IPO. My model's midpoint fair value is $550.

What happened

Microsoft has fallen 21% over the past year, and the reason is simple: it's spending big.

The company guided to roughly $190 billion in capital expenditures this year. The market hates that number. Invest at that scale and short-term margins get squeezed, investors get nervous, and they sell. The result was a 52-week low of $349 in late June; it's back around $390 now.

By the numbers

The forward P/E has compressed to about 23. Given a five-year average closer to 30, that gap is striking — and it's pulling a lot of investors back to the name.

Look at the scale. Market cap around $2.86 trillion, enterprise value about $3.06 trillion — a $200 billion gap, which is net debt. But Microsoft generated $73 billion in cash flow last year, with a five-year average of $67 billion. That $200 billion is well within reach.

The profitability is even more impressive. Gross margin of 68%, and a net margin that keeps climbing: roughly 34% over ten years, ~37% over five, and close to 40% last year. Returns on capital ran 22% a year over five years and 14% last year. One thing to watch is the dividend — under 1% yield, but $25 billion a year in cash out the door. Funding that huge capex budget and the dividend together means drawing meaningfully on the cash pile.

What Wall Street isn't talking about enough

Here's the key differentiator: Microsoft owns 27% of OpenAI.

OpenAI recently filed confidentially for an IPO, and some think it could be worth a trillion dollars. That stake is not properly reflected in Microsoft's price today. If OpenAI becomes a $1 trillion company and Microsoft owns 27% of it, that stake alone could pay off all the debt and still leave about $70 billion over.

On top of that, tools like Copilot and Fabric are embedding deeper into businesses every day. Cloud, AI, productivity software, gaming — Microsoft is effectively everywhere.

So what would I pay?

My assumptions for the next ten years: revenue growth of 7/10/13%, and net margin and FCF of 34/37/40%. The 37% and 40% are essentially what they already did last year. FCF will be dented near term by capex, but over the long run it should even out.

For the P/E ten years out, I used 20/23/26. Honestly, I'm not sure that's right — I could see higher. Microsoft is embedded in almost everyone's life, so a low-to-mid-20s multiple feels justifiable to me.

Run that with a 9% desired return and no margin of safety, and I get a low of $360, a high of $823, and a midpoint of $550. At $385, that's about a 13.5% return. Note that this return already includes the dividend — don't add it on top.

Getting paid to wait

While I wait for a better price, I still get paid. Selling a put about a month out at a $345 strike pays roughly $5 per share. With the stock near $390, I'm committing to buy $45 cheaper and collecting a cash yield in the 16% range.

The catch is emotional. If a bear market drags Microsoft to $300, I still buy at $345 (effectively $340 after the $5 premium). Buying a $300 stock for $340 is psychologically hard. But be honest: if I was willing to buy at $345 today, I'm down $45 anyway when it falls to $300. Having sold the put, I at least kept the $5 premium, so my real loss is $40, not $45. You need the discipline to actually see it that way.

What to watch

Is Microsoft a great business? In my view, yes. The real question, as always, is price.

Right now a world-class company is getting punished in the short term for building the future, trading at a valuation it hasn't seen in years. Whether that valuation is cheap enough is a question your own required return and margin of safety have to answer. But the fact that there's an OpenAI stake sitting on the table that the market isn't pricing — that's something I'll keep watching closely.