To The Moon

MSFT fell another 3.46% today, extending a multi-day losing streak that has now pushed the stock down more than 24% in 2026, potentially its steepest June drop in company history.

This is not one bad headline. It is three separate fears compounding on top of each other, and untangling which ones are real versus which are noise is exactly the work that matters right now.

Fear One: Capex Is Eating Free Cash Flow

Microsoft is on track to spend roughly $190 billion on AI infrastructure in fiscal 2026, up from an earlier estimate near $165 billion. Q3 capex alone came in at $31.9 billion, with Q4 guided above $40 billion.

The consequence shows up directly in free cash flow. FCF fell to $15.8 billion in the most recent quarter, down from $20.3 billion a year earlier, against reported net income of $31.8 billion. That gap between accounting profit and actual cash generation is the single number every analyst note has fixated on. Every dollar spent on GPU clusters and data centers hits the cash flow statement immediately as cash out, but only converts into revenue gradually over years of depreciation.

This is the bear case at its most legitimate. If AI infrastructure demand stays strong but pricing power erodes, because AWS and Google Cloud are building at the same furious pace, margins may not recover the way the bulls are modeling. The entire long-term valuation thesis rests on free cash flow recovering sharply from fiscal 2027 onward. If that recovery is delayed or diluted by competition, the $190 billion becomes a drag rather than the moat management is selling it as.

Fear Two: The OpenAI Relationship Just Got More Complicated

Yesterday's blow came from reports that OpenAI is developing in-house chip ambitions, a direct signal that Microsoft's most important AI partner is working to reduce its dependency on Azure's infrastructure stack over time.

Today compounded it. The New York Times reported OpenAI is now seriously weighing a delay to its IPO until 2027, as CEO Sam Altman holds firm on a $1 trillion valuation rather than accept a lower price in shaky markets. OpenAI's most recent private valuation sits between $730 billion and $852 billion, and Altman has reportedly called any compromise on the trillion-dollar figure a "non-starter." The news triggered a 12% single-day plunge in SoftBank shares, given its large indirect OpenAI exposure, and pressured the broader AI trade.

For Microsoft specifically, this matters on two levels. First, roughly 45% of Microsoft's $627 billion contracted backlog traces back to OpenAI, including a $250 billion commitment. A delayed IPO does not cancel that contract, but it does mean the market has to keep pricing Microsoft's AI monetization timeline around a partner whose own path to public-market accountability just got pushed out by a year. Second, Microsoft retains real protection here that the headlines undersell: royalty-free access to OpenAI's IP through 2032, an equity stake of approximately 27%, and a revenue-share agreement that continues through 2030. The relationship has shifted from simple ally to something more like a complicated, still-valuable partner, not a broken one.

Fear Three: Is the Capex Actually Buying Growth?

This is the question that should matter most, and the data here is genuinely better than the stock price suggests.

Azure reaccelerated to 40% growth in the most recent quarter, defying the slowdown narrative that had spooked the market earlier in the year. AI revenue is now running at a $37 billion annual run rate, up 123% from a year ago. Revenue overall grew 18% year over year to $82.9 billion. Copilot queries per user grew nearly 20% quarter over quarter, and weekly Copilot engagement has reached the same level as Outlook, evidence of genuine usage depth rather than just seat-count expansion. Dynamics 365's LinkedIn talent solutions agents have already surpassed a $450 million annualized revenue run rate.

Microsoft is also actively diversifying away from single-partner AI risk. The company unveiled seven proprietary MAI models at Build 2026, with MAI Thinking 1 reportedly outperforming Claude Sonnet in blind evaluations. The Scout AI agent is live across Teams. Reported talks between Microsoft and Anthropic about using Microsoft's own AI chips are themselves a partial answer to the OpenAI concentration concern, since a more diversified counterparty book reduces the single-point-of-failure exposure investors are worried about.

Operating margins came in at 46% for the quarter, and CFO Amy Hood guided full-year operating margins to rise roughly one point year over year, inclusive of one-time restructuring costs. That is a genuinely strong result given the scale of infrastructure investment underway simultaneously.

Where Wall Street Actually Stands

The disconnect between sentiment and the analyst community is stark. Of 56 analysts covering MSFT, 53 are constructive,40 rate it outright Buy, 13 rate it Outperform, only 3 sit at Hold, and zero carry a Sell or Underperform rating. The mean price target around $561 implies roughly 31% upside from current levels, with the high-end target near $870 sitting close to where deeper DCF models land their mid-case valuation.

Not every voice is uniformly bullish. Stifel's Brad Reback lowered his target to $400 while maintaining a Hold, explicitly because of what the upcoming July earnings report might reveal about Azure's trajectory. That is the right level of caution: specific, tied to a clear catalyst, not a blanket bearish call on the entire AI capex story.

There is a genuinely interesting signal buried in prediction markets too. Markets currently assign only about 31.5% odds that Microsoft's market valuation tops the combined value of Anthropic and OpenAI by year-end. That is a real, measurable form of skepticism about whether Microsoft's massive infrastructure spend translates into AI leadership versus simply subsidizing the frontier labs building on top of it.

The Inflection Point: July Earnings

Everything converges on the upcoming Q4 FY2026 earnings report in late July. Three things decide which fear wins.

Azure growth needs to hold at or above 40%, or ideally reaccelerate further. Any deceleration below 35% would likely be severely punished given current valuation, according to multiple analyst notes flagging this exact threshold.

Capex commentary on trajectory matters enormously. If management signals infrastructure spending is approaching a plateau, with free cash flow margins beginning to recover into fiscal 2027, the multiple compression concern eases considerably. If guidance instead shows capex accelerating again with no clear ROI timeline, the bear case gets fresh ammunition.

Copilot and MAI adoption metrics need to show continued depth. Specifically watch whether MAI model adoption crosses 20% of Copilot users within six months of Build 2026, and whether GitHub Copilot's new consumption-based pricing model, effective from June 2026, shows the margin improvement management is promising.

Buy the Dip, or Wait for the Inflection?

At roughly 21 to 22 times forward earnings, with Azure growing 40%, an AI run rate at $37 billion growing 123%, and a $627 billion contracted backlog, Microsoft is not trading like an obviously broken business. It is trading like a company the market has temporarily confused with its own capex anxiety.

The case for buying the dip: this is a 3 to 5 year infrastructure cycle with messy free cash flow in the middle, and the market spent 2024 to 2025 pricing Microsoft as if AI revenue would arrive immediately and cleanly. The current selloff is the repricing for reality, not a verdict on whether the business model works. A DCF-implied intrinsic range as wide as $368 to $806 suggests meaningful margin of safety even at conservative assumptions.

The case for waiting: insider transactions have leaned net-seller across recent disclosed deals, the OpenAI relationship genuinely is evolving in ways that add real uncertainty, not manufactured uncertainty, and a name this large does not need to be chased into a falling knife when the July earnings catalyst is just weeks away and will resolve much of the ambiguity directly.

The honest framework: the AI spend narrative is currently overwhelming the AI monetization narrative in how the market is pricing this stock, but the monetization data, Azure at 40%, AI ARR at $37 billion, Copilot engagement matching Outlook, is real and arriving. July earnings is where the market either confirms the spend is converting or gets its clearest evidence yet that it isn't.

Scaling in gradually ahead of that report, rather than going all-in on either side of the trade, is the position that respects both the genuine risk and the genuine opportunity sitting in the same ticker right now.

I am not a financial advisor. Trade wisely, Comrades.