微软2026年第三季度：表现良好的季度，但争议仍在继续

Mark L. Moerdler, Ph.D.+1 917 344 8506mark.moerdler@bernsteinsg.com Firoz Valliji, CFA+1 917 344 8316firoz.valliji@bernsteinsg.com Shelly Tang, CFA+1 917 344 8342shelly.tang@bernsteinsg.com Price Target MSFT 646.00 USD(641.00OLD) Microsoft 3Q26: a good quarter, but the debate continues While this was a solid quarter beating consensus on every major metric, this was not thedefinitive Azure acceleration qtr we had hoped for. At a high level Azure was in-line; CAPEXwas slightly lower; while Copilot significantly beat. Q4 guide was similar with CAPEX in-line,Azure slightly above; and Copilot growing strong. It was a solid quarter but not enough tomake those sitting on the sidelines to return to Microsoft. But one must look a bit deeper to see that the global Gen AI disruption cycle is good forMicrosoft: 1) AI is a driver of growth for Microsoft and not a headwind; 2) Copilot strengthshows AI is not eating Microsoft’s apps/ SaaS business; 3) Operating margins are increasingwhen AI worries point to deceleration; and 4) The Azure growth acceleration is not off thetable, it layers in over time. The company is spending aggressively on CAPEX and more specifically on datacenterhardware to meet demand which of course will give many investors pause, but we wouldargue that for a stock that was trading at >30x future EPS and now at 23x, Microsoft deliversthe combination of AI resilience / upside plus strong GAAP EPS growth and softwareexposure. Will everyone flock back into the stock over the next few days? We doubt it, but forthose ready to add to software exposure this is a solid / safe place to add that exposure atvery good valuation. Investment Implications Microsoft (MSFT, $646 TP, Outperform)delivered a quarter with solid Azure growth andstrong Copilot momentum. Combine that with 1) ability to deliver stable GAAP margins in themid 40s and 2) downside risk protection even if we see some bubble like behavior in partsof AI, this is a stock that every investor should have significantly in their portfolio. We raisedour price target from $641 to $646 as we roll forward our estimates and lower our valuationmultiple from 29.5x to 29x, to reflect peer step-down in peer group valuation. VALUATION COMPS TABLE DETAILS While Microsoft delivered a solid quarter beating consensus on every major metric (Exhibit 2), this was not the definitive Azureacceleration quarter we had hoped and discussed in our recent note (Microsoft: Could the bear case in fact be the bull case?). Ata very high level, Azure was in-line with buy-side expectations; in-quarter CAPEX was slightly lower than expected; while Copilotsignificantly beat expectations. Q4 guidance was similar with CAPEX roughly line-in with expectations, Azure guidance is likelyslightly above expectations and Copilot commentary much stronger than expectations. In other words, a good / solid quarterbut not enough to make those sitting on the sidelines return to Microsoft in droves. But to appreciate the quarter requires looking a bit deeper into the numbers and their implications. There are, we believe, 4 veryimportant data points that really show that the global Gen AI disruption cycle is good for Microsoft: 1.AI is a driver of growth for Microsoft and not a headwind to Microsoft growth: The company stated that Microsoft AIARR was > $37B growing 123% YoY. This means that > 11% of Microsoft’s total revenue is growing triple digits and if youplug that into your Microsoft model you would have AI, by itself, delivering ~12% revenue growth for all of Microsoft and thatis excluding the rest of the business that is doing quite well and overall growing nicely. •It is important to call out the size of Microsoft’s AI business. $37B makes Microsoft’s AI revenue stream much bigger thanits hyperscaler peers and in-line with the best in class frontier model business and that $37B does not include the non-GPU revenue that AI at enterprise customers are generating. In fact, as Agentic AI goes mainstream the positive impacton the non-GPU Azure business will be substantial and potentially another leg of growth. 2.AI is not eating Microsoft’s apps/ SaaS business: We have written extensively about the SaaS Apocalypse concernsimpacting software valuations. Office is the largest apps business at Microsoft and Office 365 Copilot, which many investorshave been concerned about, grew from 15M paid seats in Q2 to > 20M seats in Q3 and management is guiding to morenew paid seats in Q4. Note that in May M365 E7 launches which incorporates Copilot and should, we believe help drive thisstrong adoption. •Note that GitHub Copilot is also seeing acceleration and the company is now moving to a hybrid model for both GitHubCopilot and M365 Copilot with subscription plus consumption to assure Microsoft can capture the revenue upside andprotect the margin from excess usage. 3.Operating margins are increasing: Microsoft delivered another quarter in which operating margins increased. AI iscurrently negati