Morgan Stanley initiated coverage of Microsoft with an Overweight rating and a $600 price target on July 21, 2026, telling clients the software giant’s shares are undervalued at $402.29. The bull case rests on two near-term pivots: an imminent easing of Azure infrastructure constraints that have capped cloud growth, and a Copilot business evolving from a simple seat-based add-on into three distinct monetization engines. If the bank’s analysts are right, the stock offers roughly 49% upside over the coming 12 to 18 months.

What the Analyst Actually Said

The Morgan Stanley note, led by Adam Wood, argues that Microsoft is priced as if it were a slow-growth company despite earnings projected to expand more than 20% annually. The firm calculates a forward price-to-earnings multiple of about 16.9 times estimated fiscal 2028 earnings per share of $23.86—a figure that looks anomalously low for a company Microsoft’s size and growth profile.

That 16x multiple is not a conventional one-year forward P/E. It compares today’s share price with earnings that are still two fiscal years away. The thesis is less about a snap rerating and more about the market eventually paying a higher multiple for proven execution. Morgan Stanley believes a multiple closer to 25x (a 1.2 PEG ratio on 21.6% expected growth) would be more appropriate, yielding the $600 target.

The report highlights two catalysts that could force investors to reassess Microsoft’s valuation:

  1. Azure capacity is finally catching up with demand. For more than a year, Microsoft has been turning away cloud revenue because it lacked enough data centers, GPUs, and networking gear. CFO Amy Hood disclosed that if all newly deployed GPUs in fiscal 2026’s first half had been allocated to Azure, the quarterly growth rate would have topped 40%, not the reported 38%. As new capacity comes online, Morgan Stanley projects Azure revenue will hit $214.9 billion in fiscal 2028 and $305.9 billion in fiscal 2029—both ahead of consensus estimates.

  2. Copilot is becoming more than a $30 add-on. The original Copilot pitch was simple: sell a premium seat to every Microsoft 365 user. Morgan Stanley now identifies three revenue levers: (a) expanding paid seats across the installed base; (b) selling agent consumption, where enterprises are billed for the compute and tokens their custom AI agents use; and (c) embedding agents into premium workflows through Dynamics 365, Power Platform, and Fabric. The combination could eventually double average revenue per user, the analysts argue.

What It Means for You

For Windows Users and Enthusiasts

The AI buildout directly affects what you’ll see on your PC. Copilot+ PCs are designed to run certain AI tasks locally on neural processing units, while heavier workloads depend on Azure. Morgan Stanley’s thesis predicts a world where Microsoft can finally deliver enough cloud capacity to make hybrid AI feel seamless—instant local responses for personal tasks, and lag-free, powerful reasoning when the PC calls out to Azure. If Azure supply remains tight, features that require cloud inference may launch slowly or remain geographically restricted. Conversely, if capacity floods the market, expect a rapid rollout of AI features across Windows, Edge, and Office.

There’s also a subtle hardware implication: the more AI processing Microsoft can move to your PC’s NPU, the less it must spend on cloud inference for consumer features. That could keep premium Copilot features available without a mandatory subscription, but expect feature tiers to vary widely by device generation.

For IT Administrators and Enterprise Decision Makers

You should prepare for Copilot to shift from a per-seat licensing conversation to a consumption-based negotiation. Today, you buy a fixed number of licenses at a set monthly price. Under the expanding agent model, your costs could vary with how intensively your custom agents operate—every document retrieval, database query, and autonomous workflow step may generate a billable event. This is more like Azure compute than like traditional Office licensing.

Actionable steps:
- Start small and measure. Pilot Copilot with teams whose workflows are time-intensive and repetitive (developers, sales, finance). Establish baseline productivity metrics before deployment.
- Audit data permissions now. Copilot’s usefulness—and its risk—depends on what data it can access. Use Microsoft Purview, Entra ID governance, and Defender to ensure sensitive information is properly gated before agents can see it.
- Negotiate contract terms. Push for consumption caps, burst alerts, and clear service-level agreements on agent availability. If possible, lock in initial pricing for agent runtime in your Enterprise Agreement renewal.
- Watch Azure regions. New capacity arrives unevenly. If your workloads require data residency in a specific location, ask your Microsoft account team for an availability roadmap before committing to a project that hinges on GPU availability in that region.

For Developers

GitHub Copilot is the thin end of a wedge. Microsoft is building tools that let you extend Copilot into custom agents that interact with Azure services, corporate data, and third-party APIs. The three-engine strategy means your work may directly generate revenue for Microsoft each time your agent runs—so expect new Azure SKUs, monitoring dashboards, and cost controls in the Azure portal. Meanwhile, keep an eye on competition: if AWS or Google offer cheaper agent infrastructure, enterprises may route some workloads outside Microsoft’s walled garden.

How We Got Here

Microsoft’s transformation from Windows licensing giant to cloud and AI behemoth has been decades in the making. The shift accelerated under CEO Satya Nadella, with Azure becoming the company’s growth engine and Microsoft 365 locking in recurring enterprise relationships.

Generative AI added a twist: the infrastructure required is so capital-intensive that even Microsoft—with its enormous balance sheet—has been constrained. Data centers must be built, power secured, accelerators installed, and network topologies reconfigured. For the last several quarters, Azure growth was limited not by customer demand but by physical capacity. That unusual dynamic meant reported growth numbers understated the true appetite for Microsoft’s cloud services.

Meanwhile, Copilot launched as a seat-based add-on for Microsoft 365. Early feedback has been mixed: many enterprises are still assessing whether the productivity gains justify the cost per user. Morgan Stanley’s report signals that Wall Street now expects Microsoft to pivot toward a platform model, where Copilot becomes a universal interface through which Azure consumption flows—much like Windows was the interface that drove PC software licensing three decades ago.

What to Do Now

  • If you manage a team: Identify one high-value workflow (code review, sales forecasting, contract analysis) and pilot a custom agent. Track time savings and error rates for six months before scaling.
  • If you set IT policy: Review your data classification and access controls immediately. Over-permissioned SharePoint sites and stale distribution groups will become liability hot spots when agents can query them at scale.
  • If you follow the stock: The next few earnings reports are critical. Look for constant-currency Azure growth above 38%, accompanied by management commentary that supply constraints are easing. Ignore the headline revenue number; listen for whether Azure is finally able to serve the demand that has been waiting. Also watch Copilot renewal rates and any disclosure of agent consumption revenue—that will validate the three-engine thesis.

Outlook

The next six months will test Morgan Stanley’s thesis. Microsoft must demonstrate that easing Azure supply quickly translates into reported acceleration, not just prevented deceleration. Equally important, Copilot needs to show it can command premium pricing in a market where free AI assistants are everywhere. The $600 target is not a guarantee; it’s a bet that Microsoft can prove its AI investments generate cash flow comparable to traditional software. If Azure growth reignites and Copilot’s agent layer gains traction, the valuation gap could close quickly. If execution stumbles, the market’s current skepticism will look prescient.