The AI capex that crushed Microsoft's FCF is the exact reason it's mispriced the market made the identical judge on Amazon for two decades. 39x on FCF is it really expensive?
The market calling Microsoft at 39x FCF expensive reminds me of Amazon. For two decades Amazon looked expensive on FCF while it poured money into data centers and logistics. The same overreaction looks like it is hitting Microsoft now: its worst month since 2000, already down two time around 34% from its October high, mostly because a AI Capex spend has crushed reported free cash flow.
The bear case is real:
* **Is the capex actually turning into revenue?** So far, yes Azure grew around 40% last quarter, the revenue is accelerating, the AI business went from a $13B to a $37B run rate, and committed backlog hit $627B, nearly double a year ago.
* **But 45% of that backlog is OpenAI,** currently a loss making company looking to turn profitable in 2029-2030. That single customer concentration is exactly what broke CoreWeave, Microsoft was around 65% of its revenue, but the difference here is OpenAI is 45% of Microsoft backlog not revenue, tbh the toughest point to debate with the bears and I wont disagree.
* **And the real question is whether that capex keeps climbing or finally normalizes**. I personally think the capex will remain in the high twenty's in FY27 and FY28 still.
Running a DCF model over the following assumptions:
* **Revenue** compounds 20% next year, slowing to 17-18% by FY31 as the AI layer keeps the top line growing
* **Operating cash flow** holds near 50% of revenue
* **Capex** stays pinned near 30% of revenue through FY27 an FY28 (no fast normalization on my perspective), then eases toward \~21% by FY31
Discount that back at 8.8% with a 4% terminal growth rate, and the base case is worth about $540, roughly 45% above today.
Coming back to Amazon. For two decades Amazon looked expensive on FCF while it poured money into data centers and logistics. FCF metric never saw it, because FCF subtracts every dollar of capex. Owner's earnings I think is the right lens here. Most of MSFT capex today is the data center build: growth spending that won't repeat, not the cost of staying in business. Strip it out, and the stock that looks about 21x owner's earnings. The market is penalizing Microsoft for building its next moat, and that's usually the best time to own one.
Full detailed analysis on my Substack for free:
[https://open.substack.com/pub/equivara/p/microsoft-deep-dive-microsoft-isnt](https://open.substack.com/pub/equivara/p/microsoft-deep-dive-microsoft-isnt)
Are you holding or buying Microsoft?
Disclosure: This is my personal thesis, not investment advice. I am not a registered investment adviser. Do your own research and size positions according to your own risk tolerance
