Microsoft closed Thursday at $381.58, down for a third consecutive session and roughly 5 percent below where it traded before announcing an expanded partnership with the French AI company Mistral on July 21. The stock spent the weekend range-bound in the $380s — within shouting distance of its 52-week low of $349.20, and a long way from the $555.45 high it touched a year ago. On Wednesday, July 29, it reports fiscal fourth-quarter earnings.
That date carries more weight than a routine earnings call usually does. The Federal Reserve’s rate decision lands the same afternoon, followed by Chair Kevin Warsh’s press conference. Meta reports after the close, the same day as Microsoft. Apple and Amazon follow on July 30. Four of the most closely watched names in the market, plus a Fed decision, compressed into a 48-hour window — and Microsoft walks into it trading near its cheapest level in a year, with Wall Street’s own price targets implying 50 percent upside or more from current levels.
That gap — between where the stock trades and where the analysts who cover it say it should trade — is the story going into Wednesday. It did not appear overnight, and it will not resolve on the earnings call alone.
A Third Straight Losing Session, and Why
The proximate cause of the slide is identifiable and, on its face, should have been good news. On July 21, Microsoft announced an expanded strategic partnership with Mistral AI, the French AI lab: Microsoft will help fund and consume capacity from Mistral’s European data-center buildout, running on Nvidia’s next-generation Vera Rubin chips, aimed at enterprise and regulated-industry customers who want AI infrastructure that satisfies European data-sovereignty requirements. Mistral’s Medium 3.5 and OCR 4 models were added to Microsoft Foundry and Copilot Studio the same day.
The market’s read was not “Microsoft secured another AI distribution channel.” It was “Microsoft committed more multibillion-dollar capital to AI infrastructure a week before investors get to see what the last round of that spending actually produced.” The stock fell the day of the announcement and continued falling for two more sessions, a pattern consistent with a market that has grown warier of capex commitments than it was six months ago — not because the commitments are wrong, but because the return timeline on them remains the central unresolved question hanging over every hyperscaler’s stock, Microsoft’s included.
The capital-expenditure trajectory we have tracked through 2026 is the backdrop against which the Mistral deal landed badly. Estimates for Microsoft’s full fiscal-year 2026 capital spending have ranged as wide as $80 billion to $146 billion across different analyst models — a spread wide enough to indicate that even sophisticated coverage cannot pin down management’s actual spending intentions with confidence. A company whose investors cannot agree on its spending level, adding another undisclosed-dollar-amount commitment a week before earnings, is not a company set up for a calm week.
The Bull Case Has a Number: $600
Set against the sliding price is a specific, recently reiterated bull case. Morgan Stanley’s Adam Wood carries an Overweight rating on Microsoft with a $600 price target — implying roughly 54 percent upside from the stock’s current level — built on a 25-times multiple applied to fiscal-year-2028 earnings estimates. The thesis rests on Azure entering what Wood’s team frames as an acceleration phase, with growth in the low 40s percent range, and on Copilot’s monetization finally showing up in reportable numbers rather than adoption metrics alone.
Morgan Stanley is not an outlier. The broader analyst consensus on Microsoft remains solidly in Buy-to-Strong-Buy territory, with average price targets across full coverage clustering well above $550 — a gap of 40 percent or more from Thursday’s close that has persisted, and in some cases widened, as the stock has fallen through July. That combination — falling price, unmoved or rising targets — is either a market mispricing a fundamentally sound business, or a set of analyst models that have not yet caught up to a genuine deterioration in the growth or margin outlook. Wednesday’s report is the first hard data point in three months that can discriminate between those two explanations.
The Azure growth question is the crux of it. Microsoft’s cloud business grew revenue at roughly 40 percent year-over-year last quarter — a number that would be an unambiguous positive in almost any other competitive context. The complication is Google Cloud, which has been posting growth rates near 82 percent, more than double Microsoft’s pace. Azure remains larger in absolute revenue, and a smaller base naturally grows faster in percentage terms, so the comparison is not apples-to-apples on its own. But it has been enough to introduce a share-loss narrative into Microsoft coverage that did not exist a year ago, and Wednesday’s Azure growth number — not just whether it beats consensus, but whether it accelerates or decelerates from the prior quarter — will either feed that narrative or defuse it.
What the Numbers Need to Show
Consensus estimates for the quarter sit at earnings per share of $4.21 to $4.24, up roughly 15.3 percent from $3.65 a year earlier, on revenue growth estimated near 15 percent. Microsoft has beaten consensus EPS in each of its last four reported quarters — a streak that raises the bar for what counts as a genuine beat this time, since the market has come to expect one. A number that merely meets consensus, after four straight beats, risks being read by a jittery market as a disappointment relative to trend, even though it would represent solid double-digit growth in absolute terms — a reminder that the market’s reaction function into this print is calibrated against Microsoft’s own recent history, not against some neutral external benchmark.
Copilot’s monetization gap, which we documented in detail earlier this year, remains the second pillar of the bear case alongside capex uncertainty. As of the most recently disclosed figures, Microsoft 365 Copilot had reached roughly 20 million enterprise seats, up 250 percent year-over-year — a genuinely large adoption number. What the market has not yet seen is a clean disclosure connecting that seat count to incremental revenue at a scale that justifies the AI infrastructure spend funding it. Wednesday’s report is an opportunity to close that gap, or to leave analysts modeling Copilot’s contribution the same way they have been: as a qualitative growth driver rather than a quantified one.
Three specific disclosures will do most of the work of resolving the bull-bear tension. First, Azure’s growth rate and whether it moved toward or away from Google Cloud’s pace. Second, any dollar-denominated Copilot revenue disclosure, rather than a seat count alone. Third, forward capital-expenditure guidance for fiscal 2027 — a number that, given the $80 billion to $146 billion uncertainty already surrounding fiscal 2026, the market will treat as at least as important as the quarter just reported.
Reading the Consensus Gap Correctly
A 40-to-54 percent gap between a stock’s trading price and its average analyst target is unusual enough to warrant a closer look at how it happened, because the two most obvious explanations — the market is wrong, or the analysts are wrong — are not the only possibilities.
A third explanation is timing mismatch. Morgan Stanley’s $600 target is built on a 25-times multiple applied to fiscal-year-2028 earnings, which means it is not, strictly speaking, a forecast of where the stock should trade today. It is a forecast of where a $600 valuation becomes justified once two more fiscal years of growth have compounded. Expressed as upside from today’s price, a target like that will always look dramatic, because it is discounting a future earnings base rather than the current one. That does not make the target wrong, but it means the “54 percent upside” framing overstates how quickly the market is expected to close that gap. A more precise reading is that Morgan Stanley believes Microsoft’s business, two years from now, supports a materially higher valuation than the market is currently pricing in — a different and more modest claim than “the stock is 54 percent underpriced today.”
That said, even adjusting for the multi-year discounting, the gap between price and consensus has not been this wide for Microsoft in the past several quarters, and gaps of this size do not typically persist through an earnings report without some resolution. Either the quarter provides evidence that the two-year growth story analysts are pricing is on track, in which case the stock likely re-rates upward even without matching the full target immediately, or the quarter introduces new doubt about that growth story, in which case the targets themselves are the ones due for revision.
Options markets offer one more data point on how large a move traders expect. Implied volatility on Microsoft into the July 29 print has been running above its typical pre-earnings level, consistent with a market that expects a genuinely consequential reaction rather than a routine beat-and-hold outcome — itself a reflection of how much is riding on this particular quarter relative to Microsoft’s usual earnings cadence.
Why the Fed Shares the Stage
The scheduling collision with the Federal Open Market Committee is not Microsoft’s doing, but it materially changes how the market will be positioned going into the print. The Fed’s decision is due at 2:00 PM Eastern on July 29, with Chair Warsh’s press conference following at 2:30 — hours before Microsoft’s own report lands after the closing bell. The same oil-driven inflation pressure we tracked in Bitcoin’s price action this month has pushed rate-hike odds for the Fed’s September meeting up sharply, and any hawkish signal from Warsh’s press conference — language suggesting the committee views the recent oil shock as feeding into persistent rather than transitory inflation — would tighten financial conditions hours before investors get Microsoft’s own numbers, compressing the multiple the market is willing to assign to whatever growth Microsoft reports.
The reverse is also true. A Fed decision read as dovish, or a Warsh press conference that treats the oil-driven inflation spike as temporary, would loosen the discount-rate pressure on every long-duration growth stock in the market simultaneously — Microsoft included, regardless of what its own numbers show a few hours later. In an ordinary earnings week, Microsoft’s stock would trade almost entirely on Microsoft’s numbers. This week, roughly half of the trading day’s information flow arrives from an institution that has nothing to do with Redmond.
Meta’s earnings, landing the same evening as Microsoft’s, add a third variable. Both companies are, in different ways, defending AI capital-expenditure programs to investors who have grown less patient with unquantified return timelines than they were a year ago. A Meta report that reassures the market on AI-spending discipline would likely lift risk appetite for Microsoft’s own capex story by extension; a Meta report that spooks investors on the same theme would do the opposite, independent of anything Microsoft itself discloses. The two companies are not competitors in most of their core businesses, but on July 29 they are, functionally, co-defendants in the market’s ongoing trial of AI capital spending.
What a Genuine Azure Deceleration Would Mean
It is worth walking through the downside scenario explicitly, because the market has spent enough of 2026 pricing hyperscaler AI stories on faith that the specific mechanics of a disappointment are easy to lose track of.
Azure growing at 40 percent is, by any historical standard for a cloud business of Microsoft’s scale, an exceptional number — the kind of growth rate that, five years ago, would have been treated as unambiguous evidence of category dominance. What has changed is not Azure’s absolute performance but the competitive frame around it. Google Cloud’s 82 percent growth rate, even acknowledging its smaller revenue base, has introduced a genuine question about whether Microsoft is ceding share in the specific, high-value segment of the market — large enterprise AI workloads — where competitive dynamics matter most for long-term positioning. A deceleration in Azure’s growth rate on Wednesday, even a modest one, would be read by the market not as “Azure grew slightly less than last quarter” but as “the share-loss narrative has data behind it now,” and stocks tend to reprice sharply on narrative confirmations even when the underlying number moves only marginally.
The capital-expenditure side compounds this risk. If Azure decelerates while capex guidance for fiscal 2027 comes in at the high end of the $80 billion to $146 billion range that has circulated in analyst models, the combination is close to the worst-case framing for the stock: slowing growth in the business the spending is meant to accelerate, alongside confirmation that the spending itself is not moderating. That combination — deceleration plus unchecked capex — is precisely the scenario that would validate the bear case this narrative has tracked since earlier in the year, and it is the scenario Wednesday’s report needs to avoid most directly if the stock is to find a floor rather than test new lows.
The Structural Backdrop This Series Has Tracked
The AI-squeeze dynamic we identified earlier in this narrative — Microsoft caught between capital-intensive AI infrastructure commitments and a monetization timeline that keeps extending — has not resolved itself in the months since. If anything, the Mistral announcement is a data point in favor of the squeeze thesis rather than against it: a company already facing capex-uncertainty questions added another open-ended international commitment in the same week analysts were finalizing their earnings previews.
The end of Microsoft’s exclusivity arrangement with OpenAI, which this publication covered as it unfolded, remains relevant context for Wednesday’s report in a way that is easy to lose amid the capex and Azure questions. Microsoft’s Copilot products no longer draw on OpenAI’s models as an exclusive, protected input — competitors can access comparable underlying model capability through other arrangements, including AWS’s Bedrock platform. That erosion of a structural moat does not show up as a single line item in an earnings report, but it is the kind of condition that makes every subsequent quarter’s Copilot monetization numbers more important, not less: without exclusivity as a competitive shield, Copilot has to win on product execution and distribution alone, and Wednesday’s disclosures are the market’s best available window into whether that execution is working.
Xbox restructuring, the subject of extensive coverage in this narrative earlier in the year, is a smaller but still-live thread. Arkane Lyon’s fate — sale, management buyout, or closure, subject to the French Works Council consultation process required before Microsoft can finalize any of those outcomes — remains unresolved as of this writing. It is unlikely to feature prominently in Wednesday’s call, since gaming is a minor segment relative to Azure and productivity software in Microsoft’s overall revenue mix, but any update on the restructuring’s cost or timeline would be a data point for the broader thesis that Microsoft is simultaneously expanding aggressively in AI infrastructure and contracting in businesses that do not fit the AI narrative.
What the Dividend and the Databricks News Do and Do Not Tell Us
Two smaller items from the past week are worth placing in context, precisely because they illustrate how little the routine news flow around a company like Microsoft says about the questions that actually move the stock.
Microsoft declared its regular quarterly dividend of $0.91 per share, with an ex-dividend date of August 20 — an announcement so ordinary it barely registers as news, and it should not be read as any kind of signal about capital-allocation priorities shifting toward shareholder returns and away from AI infrastructure spending. Dividend declarations at this level are calendar events, not strategy statements; Microsoft has raised or maintained this payment through every phase of its AI buildout, and Wednesday’s capex guidance will tell the market far more about capital priorities than the dividend schedule does.
The expanded Databricks partnership, announced the same week, is more substantive but still secondary to the core earnings questions. Deepening ties with a leading data-and-AI platform company reinforces Microsoft’s positioning as infrastructure of choice for enterprise AI workloads — consistent with, rather than contradictory to, the Azure growth story analysts are modeling. But partnership announcements of this kind have become frequent enough across the entire hyperscaler category in 2026 that the market has largely stopped treating them as independently market-moving; investors want revenue attribution, not distribution-channel announcements, and Wednesday’s report is the venue where that attribution either appears or does not.
Neither item changes the fundamental setup. They are the kind of news that fills a quiet earnings-week calendar without altering the two or three numbers that will actually determine how the stock trades on Thursday morning.
The Setup, Stated Plainly
Three days before its most consequential earnings report in recent memory, Microsoft trades near a one-year low while the analysts who cover it maintain price targets implying 40 to 54 percent upside. That gap will not persist indefinitely — either the stock rises to meet the targets, the targets fall to meet the stock, or Wednesday’s disclosures split the difference by giving both camps partial vindication.
The specific numbers to watch, in order of likely market impact: Azure’s growth rate and its trajectory relative to Google Cloud’s; any dollar-denominated Copilot revenue figure; fiscal 2027 capital-expenditure guidance; and, layered on top of all three, whatever the Federal Reserve signals about the path of rates a few hours before the numbers land. Meta’s own report that evening will color how the market interprets whichever of those four signals Microsoft delivers.
None of this resolves in advance. What can be said with confidence is that the setup itself — a stock at a one-year low, a Street consensus well above the current price, and a calendar collision that puts the Fed, Microsoft, and Meta in the same 24-hour window — is unusual enough that Wednesday’s outcome will be read as more than a single company’s quarterly result. It will be read as a referendum on how much patience the market still has for AI capital spending in mid-2026, with Microsoft as the marginal name whose earnings call happens to fall on the day that referendum comes due.
How This Compares to Microsoft’s Last Earnings-Week Setup
It is worth noting how different this setup is from the one Microsoft faced heading into its prior quarterly report, if only to underscore why this particular print carries more weight than a typical quarter’s. Three months ago, Microsoft entered earnings trading near its highs, with a consensus that had largely converged around a shared bullish narrative — Azure growth, Copilot adoption, and capex all pointing in the same direction with little daylight between the bull and bear cases. The report that followed was, by most accounts, a confirmation of an already-agreed-upon story rather than a re-rating event.
This quarter offers none of that convergence. The stock has fallen roughly 30 percent from its 52-week high. The bull case and bear case have genuinely diverged, with billions of dollars in capital-expenditure uncertainty sitting between them. And the scheduling accident that places the Fed’s decision and Meta’s own earnings in the same 24-hour window means that even a clean, unambiguous Microsoft result could be overwhelmed by information the market receives from elsewhere on the same day. Three months ago, Microsoft’s earnings were the headline event of the week they landed in. This time, they are one of at least three headline events competing for the same news cycle — a genuinely unusual position for a company of Microsoft’s market capitalization to occupy, and one that reflects how crowded and consequential this particular stretch of the 2026 earnings calendar has become.
Investors positioning ahead of Wednesday are, in effect, making a bet not just on Microsoft’s numbers but on the sequencing of information across a single trading day — whether the Fed’s afternoon signal, Microsoft’s after-hours report, and Meta’s simultaneous release reinforce or contradict one another. A clean sweep in one direction, all three sending the same message about growth, rates, and AI spending discipline, would produce a sharper move than any single report could generate alone. A split outcome — the Fed hawkish while Microsoft beats, or Microsoft disappoints while Meta reassures — would likely produce a choppier, more uncertain session, with the stock’s ultimate direction depending on which signal the market decides to weight most heavily. Given how compressed the calendar is, that weighting decision may not resolve cleanly even by Thursday’s opening bell.
