Microsoft's $37B AI Machine Is Just Getting Started — and It's Trading at a Discount

Let's cut through the noise. Microsoft just posted $82.9 billion in quarterly revenue — up 18% year over year — while its AI business quietly exploded to a $37 billion annual run rate. That's up 123% from a year ago. This isn't a company catching a wave. This is a company that built the ocean.

Azure growth re-accelerated to 40% — the strongest print in years. Microsoft Cloud pulled in $54.5B for the quarter alone, up 29%. Operating income hit $38.4B, up 20%. Net income? $31.8 billion. EPS came in at $4.27, also up 23%.

And the stock? Down 17% year-to-date. Trading at 20x forward earnings. For context, that's a discount MSFT hasn't offered in years.

Here's the number that should make you sit up: commercial remaining performance obligation of $627 billion. That's nearly double last year. It means enterprises are pre-committing to Microsoft services at a pace we've never seen. Every dollar of that $97B capital expenditure is getting booked as future revenue before the data center lights even turn on.

The Azure story isn't just about cloud compute anymore. It's about being the platform layer where enterprise AI actually runs. Copilot is embedded across M365 — commercial cloud subscriptions grew 19%. Dynamics 365 grew 22%. Consumer M365 cloud? Up 33%. The AI monetization is happening at every layer of the stack simultaneously.

Yesterday, Microsoft committed $2.5 billion and 6,000 employees to a new "Frontier" AI implementation unit. They're also merging the enterprise and consumer Copilot chatbots into one unified product by August. This isn't experimentation. This is execution at scale.

Now let's address the OpenAI elephant. The narrative that Microsoft is hopelessly dependent on OpenAI? It's getting stale. At Build 2026, they showed 7 in-house MAI models — purpose-built for different workloads. The OpenAI P&L impact this quarter was just $14 million, compared to $583M a year ago. That's a 97.6% reduction. The revenue share payments are capped as of April 2026. This is becoming a distribution and platform deal, not a dependency.

Michael Burry went long around June 27. The stock jumped 5% the next session. When the guy famous for predicting the 2008 crash starts buying mega-cap tech at 20x earnings, pay attention.

Analysts are overwhelmingly on board. Of 55 coverage analysts, 53 rate it Buy or Strong Buy. The mean price target sits at $561 — that's 44% upside from where we are. The high target? $870. Only two downgrades all year, both to Hold, not Sell. Even Stifel's CapEx concerns didn't stick — the $627B RPO answers every question about ROI.

Jefferies noted on July 1 that Microsoft and Amazon are "standing out" in cloud spending. Deutsche Bank turned bullish on the AI story the same week. The smart money is rotating back in while retail is still staring at the YTD red.

Let's talk about what $390 actually buys you. You're getting $318 billion in trailing revenue growing at 18%. A 46.3% operating margin. $72.9 billion in free cash flow. And an AI business that tripled in a year while the enterprise customer base is locking in nearly $630 billion in future contracts.

The 52-week range tells you everything: $349 to $555. We're sitting closer to the bottom. The Windows OEM business is down 2% and Xbox is off 5%, but nobody's buying MSFT for gaming hardware anymore. The enterprise flywheel is the story, and it's accelerating.

Microsoft returned $10.2 billion to shareholders in Q3 alone through dividends and buybacks. The company is spending $97B on AI infrastructure while simultaneously funding massive shareholder returns. That's not a startup burning cash — that's a mature monopoly printing money and reinvesting it at the same time.

The market is pricing Microsoft like a slow-growth utility at 20x forward earnings. The fundamentals say it's an AI infrastructure monopoly growing revenue 18% with 40% cloud growth and a $37B AI business that just tripled. Something has to give.

Disclosure: The Signal holds no position in MSFT. Positions may change. This is not financial advice.