Why Burry Bought Microsoft: The AI Capex Debate in Numbers

It is rare for Microsoft to become a buy on fear. Rarer still for a $3 trillion company.

The stock fell to $411, with intraday lows near $355. That is where Michael Burry initiated. Tracking the gap between what the market sees and what he sees, in actual numbers, makes this look less like "buy quality cheap" and more like a specific, dated thesis.

The one thing the market fears

It comes down to a single line item: $80 billion a year in AI capex with an unclear payback window.

That's why free cash flow now prints meaningfully below net income. Capex is depreciated slowly on the income statement but exits the cash flow statement immediately. For any other company that gap would invite skepticism. Microsoft gets a pass because the spend has a clear destination — data centers and GPUs. But the longer the revenue catch-up takes, the more the multiple wobbles.

Fundamentals at a glance

The core dashboard:

• Revenue: ~$318B
• 5-year ROIC: ~19% (slightly softening recently)
• Net margin: 39% (1y), 36.7% (5y), 34% (10y) — margins still expanding over time, which is rare at this scale
• Revenue growth: accelerating across 3, 5, and 10-year windows
• P/S: ~9.5x
• P/E: ~34x on 5-year average
• P/FCF: ~45x on 5-year average
• Dividend yield: 0.84% (~$26B annually)

On an eight-pillar screen Microsoft passes six and fails two. Both fails are valuation. The business itself has no flaws on the screen — it is just expensive.

What consensus sees

Analyst consensus models the next four years of EPS growth at 24.5%, 14.5%, 17.5%, 22%. Revenue growth at 17.5%, 15.5%, 16.5%, 17%, 14%. If a company this size sustains those rates, EPS roughly doubles in four years. The point is that an aircraft carrier is still accelerating.

A more conservative scenario

Hard to take consensus at face value. Assume AI doesn't pay off as cleanly as bulls expect — revenue growth at 7%/9%/11%, 10-year operating and FCF margins at 34%/37%/40%, exit P/E at 20x/23x/26x. The intrinsic value picture comes out:

• Low: ~$360
• Mid: ~$506
• High: ~$702

With the stock at $411 and a recent low near $355, a window briefly opened where the price sat below even the conservative low case. If Burry assumes anything close to consensus growth, his perceived margin of safety is much wider than this.

Bottom line

Microsoft is no longer an easy buy. But even cutting consensus revenue growth roughly in half, the current price falls inside a reasonable intrinsic value band. The capex concern is a short-term momentum killer, but on a multi-year view it can be the very trigger that creates the entry. Burry seems to have taken the latter view.