Hey there, bargain hunter. Here’s an unusual situation. Microsoft posted Q3 fiscal 2026 revenue of $82.9 billion, up 18% year-over-year. Azure grew 40% (39% in constant currency). The AI business crossed a $37 billion annual revenue run rate, up 123% year-over-year. Diluted earnings per share came in at $4.27 on a GAAP basis, up 23%. Every major metric beat expectations. And the stock is still down roughly 19% year-to-date as of June 20, 2026, sitting near $379.

That’s the disconnect active traders need to understand right now.

The valuation compression has been driven by a specific concern. For calendar year 2026, Microsoft said it expects to invest roughly $190 billion in capital expenditures, including approximately $25 billion from the impact of higher component pricing. The market looked at that number and pushed the stock lower even as the underlying revenue accelerated. The 52-week range tells the story: high of $555.45, low of $356.28. Right now the stock is sitting much closer to the floor than the ceiling.

What’s interesting is that the capex number is large, but it’s not undisciplined. More than 90% of the Fortune 500 are using Microsoft AI. The switching costs on that installed base grow with every model trained and every API configured. The capex is building a moat, not chasing a hype cycle.

The AI Business Nobody Is Talking About

That $37 billion AI annual revenue run rate deserves more attention than it’s getting. The 123% growth rate means this is still in early-stage acceleration, not deceleration. CEO Satya Nadella said in the earnings release that Microsoft is focused on delivering cloud and AI infrastructure and solutions for the agentic computing era, and the numbers back that up.

The agentic layer is where the next revenue step comes from. Microsoft introduced Scout at Build 2026. The company has also signaled it is exploring additional model options for Copilot experiences, including the possibility of using DeepSeek for certain use cases. Microsoft also moved one of its enterprise Copilot agent offerings to usage-based pricing. It’s early to quantify the impact, but the direction is clear: monetization is shifting toward consumption, which scales faster than headcount.

Slight tangent, but it matters. Gross margin came in at 67.6%, the narrowest since 2022. That’s what depreciation does when you’re mid-cycle in a massive infrastructure build. The market is reading that as a warning sign. The counterargument is that the revenue being generated on top of that compressed margin is accelerating, not stalling.

Copilot and the $627 Billion Backlog

Microsoft now has over 20 million Microsoft 365 Copilot paid seats, up from 15 million in January. The add-on runs $30 per user per month on top of existing M365 enterprise licenses. With hundreds of millions of commercial Office users globally, even modest attach rate expansion translates to billions in incremental high-margin revenue annually.

The number that really jumps out, though, is this one. Commercial remaining performance obligations hit $627 billion at the end of Q3 FY2026, up 99% year-over-year, with a weighted average duration of roughly two and a half years. That’s contracted revenue not yet recognized. It’s the enterprise pipeline already locked in. Roughly 25% of that will convert to revenue in the next 12 months, itself up 39% year-over-year. The old email referenced a Q1 figure of $392 billion. The current Q3 number is $627 billion. That’s a meaningful update.

The AI infrastructure build is largely pre-sold. This isn’t speculative spending into a demand vacuum.

Where the Stock Actually Sits

• MSFT trading near $379 as of June 20, 2026
• 52-week range: $356.28 low (March 30, 2026) to $555.45 high (July 2025)
• YTD total return: approximately -19%
• Trailing P/E: approximately 23x based on current pricing
• Market cap: approximately $2.82 trillion
• Q4 FY2026 earnings: confirmed for July 28, 2026, after market close

The Capex Overhang and Why It May Be Peaking

The honest tension in the MSFT story is whether roughly $190 billion in annual infrastructure spend eventually generates the free cash flow to justify it. The depreciation on that capex begins hitting the income statement over the next 2-3 years. Q4 operating margin is already guided down to approximately 44%, from 46.3% in Q3. Azure capacity constraints are expected to persist through 2026 with only modest improvement in the second half of the calendar year. The market is pricing in some risk that the returns on this investment cycle arrive later than management implies.

That said, the $627 billion commercial RPO provides a floor under near-term revenue expectations that most companies simply don’t have. Microsoft said on the Q3 call that it remains confident in the return on its investments given higher demand signals and increasing product usage. Whether you believe that or not probably comes down to how much weight you put on the backlog versus the margin compression.

Three Scenarios Into July 28

Bull Case: Azure sustains 39-40% growth in constant currency in Q4, meeting or beating its own guidance. Copilot paid seat adds accelerate sequentially, as management suggested they expect on the Q3 call. Usage-based pricing metrics from the new agent billing model give analysts confidence in FY2027 revenue. Stock recovers meaningfully off the current lows. The consensus analyst price target in the $525-$560 range starts to look achievable on a 12-month basis.

Base Case: Azure meets but does not beat guidance. Copilot adoption continues at the current pace without a breakout acceleration. Free cash flow remains under pressure from capex. Stock moves modestly higher post-earnings but doesn’t trigger a multiple expansion event. The 2026 AI infrastructure overhang keeps a lid on sentiment until the depreciation math becomes clearer in FY2027 guidance.

Bear Case: Azure growth decelerates below 38% on continued capacity constraints. Margin pressure from the capex cycle accelerates faster than expected. Copilot monetization proves slower to scale than management implied at Build 2026. FY2027 EPS estimates get revised lower, compressing the forward multiple. Stock tests or breaks toward the March lows near $356.

What Traders Are Actually Watching

July 28 is the near-term fulcrum. Between now and then, MSFT is essentially a waiting trade. The fundamentals are strong. The sentiment is poor. That’s the environment where patient, disciplined traders typically find the best risk/reward opportunities, not because momentum has turned, but because the asymmetry tends to favor those willing to tolerate the overhang.

The core question isn’t whether Azure is growing. It clearly is, at 40%. The question is whether the market will re-rate the stock based on that growth once the capex peak becomes more visible and the AI revenue run rate continues to climb. With an AI business at a $37 billion annual run rate growing at 123%, and $627 billion in contracted future revenue sitting on the books, the downside math is arguably more limited than the 19% YTD decline implies.

Watch the $356-$360 zone as the line that changes the conversation. A decisive close below there would be a different kind of signal. For now, the real bet is whether July 28 gives the bulls the catalyst they’ve been waiting for since January.

The Cheap Investor