Microsoft's (MSFT.US) annual Ignite technology conference is being held from November 17 to 20 in San Francisco, according to reports.
Unlike the developer-focused Build conference, which emphasizes software development tools and new technologies, Ignite concentrates on enterprise customers, and Microsoft will use the event to showcase its latest advances in cloud computing, artificial intelligence, data, security, and agents.
Morgan Stanley views this year's Ignite as an important checkpoint for gauging the progress of Microsoft's enterprise AI strategy, rather than merely a product launch event. The firm maintains an "Overweight" rating on Microsoft with a price target of $650.
Morgan Stanley's core thesis is that Microsoft is building an "enterprise AI flywheel" composed of Azure, Fabric, Foundry, and Microsoft 365 Copilot: Azure provides the underlying compute, Fabric integrates enterprise data and business context, Foundry helps customers select and orchestrate models and build agents, and Copilot embeds AI capabilities into employees' daily workflows.
In Morgan Stanley's view, what truly deserves attention at this conference is not how many new features Microsoft unveils, but whether this technology stack can translate into more enterprise production deployments, higher product consumption, and more sustainable revenue growth.
Key Agenda: Nadella's Keynote Scheduled for Early Morning Beijing Time on November 18
Microsoft's official Ignite schedule shows the conference runs from November 17 to 20 at the Moscone Center in San Francisco, with online events also available. The opening keynote will be held at Chase Center in San Francisco, co-hosted by Microsoft CEO Satya Nadella and Judson Althoff.
Key public event times are as follows: Nadella's opening keynote on November 17 from 10:00 to 12:00 U.S. Pacific Time (November 18, 02:00 to 04:00 Beijing Time); Live at Microsoft Ignite: Cloud & AI Innovation on November 17 from 12:00 to 13:00 PT (November 18, 04:00 to 05:00 Beijing Time); the Azure and NVIDIA accelerated computing session on November 17 at 15:15 PT (November 18, 07:15 Beijing Time); the Foundry and NVIDIA Nemotron agent session on November 18 at 11:30 PT (November 19, 03:30 Beijing Time); and the cloud-to-edge agentic AI factory session on November 19 at 14:45 PT (November 20, 06:45 Beijing Time).
In addition, the Morgan Stanley report also lists investor meeting arrangements with Microsoft management: on November 17 at 16:00 U.S. Eastern Time, a session with Nadella; and on November 18 at 15:00 ET, a meeting with Chief Financial Officer Amy Hood and Vice President of Investor Relations Jonathan Neilson. These are investor events listed in the report, not public keynote sessions at the conference.
Enterprise AI Flywheel Already Has Earnings Support, but the Key Is Whether It Can Continue to Deliver
Morgan Stanley's optimistic view of Microsoft's enterprise AI strategy is first supported by the latest quarterly results. Microsoft reported fiscal first-quarter 2027 results on October 28. For the quarter ended September 30, revenue reached $92.3 billion, up 19% year over year; operating profit rose 21% year over year, with an operating margin of about 50%. Azure revenue grew 48% year over year, and Microsoft 365 commercial cloud revenue rose 18% at constant currency. Commercial remaining performance obligations (RPO) reached $677 billion, up 73% year over year. Of that, about 30% is expected to be recognized as revenue within the next 12 months. This massive backlog provides high visibility for Microsoft's future cloud business growth.
Notably, Azure's growth is not driven solely by AI infrastructure demand. Microsoft management said Foundry revenue more than tripled year over year, while additional compute from improved GPU cluster efficiency and better CPU resource utilization also contributed to Azure's growth. At the same time, Microsoft Cloud revenue reached $63.9 billion, up 30% year over year and 31% at constant currency.
Morgan Stanley believes Microsoft's advantage lies in its ability to connect cloud infrastructure, data platforms, model services, and office applications, allowing enterprises to complete multiple steps from data governance to agent deployment within a single ecosystem. However, whether this model can truly form a commercial closed loop still requires more evidence at this year's Ignite.
Five Key Takeaways: Copilot Monetization, Foundry Orchestration, and Compute Efficiency
1. Copilot Commercialization: From Seat Growth to ARPU Improvement
Whether Copilot can further raise average revenue per customer (ARPU) is one of the most important observations of this conference. Morgan Stanley believes Microsoft's monetization path is expanding from simply selling Copilot seats to three directions: increasing penetration of directly paid seats, pushing customers to upgrade to the higher-tier Microsoft 365 E7 package, and generating additional revenue from agents executing tasks through usage-based billing.
On its latest quarterly earnings call, Microsoft disclosed that Copilot daily active users grew nearly 50% quarter over quarter and more than fivefold year over year; the number of customers actively using usage-based Copilot Credits grew more than 75% quarter over quarter. In addition, the number of customers with more than 50,000 Copilot seats grew more than eightfold year over year. This means Microsoft is attempting to expand its AI business model from traditional software subscriptions to charging based on actual usage. As agents take on more workflows, enterprise spending is expected to extend beyond IT department software procurement budgets to business unit operating budgets.
However, Microsoft also changed the way it discloses Copilot business data in the latest quarter. The company had previously disclosed paid seat numbers on a consecutive basis, saying in July 2026 that Microsoft 365 Copilot paid seats exceeded 30 million; but in the latest quarter, Microsoft did not update that figure, instead emphasizing daily active users, customer usage, and revenue momentum. CFO Hood explicitly said on the earnings call that the company did not disclose an update to the Copilot seat number this quarter. Although usage metrics have grown, the lack of an updated paid seat base makes it difficult for outsiders to accurately judge how much of the increase in user activity has translated into new paid revenue.
Changes in pricing models are also worth watching. Microsoft's Microsoft 365 E7 package is priced at $99 per user per month, integrating Microsoft 365 E5, Copilot, Agent 365, and Entra, among other products. At the same time, Microsoft plans to enable usage-based billing by default for newly purchased Microsoft 365 Copilot Business licenses sold through the Cloud Solution Provider channel, with a default allowance of 4,000 Copilot Credits per user per month. That arrangement has been adjusted to roll out gradually in the third quarter of fiscal 2027, namely January to March 2027.
Therefore, the key question this year's Ignite needs to answer is: can Microsoft prove that Copilot not only has more active users, but can also continuously increase revenue through seat upgrades and usage-based billing?
2. Foundry: In the Multi-Model Era, Can Microsoft Control the Agent Orchestration Gateway?
If Copilot is responsible for delivering AI to end users, then Microsoft Foundry plays the platform role for enterprise AI application development and deployment. Microsoft said in its latest quarter that Foundry has more than 120,000 customers, with related revenue more than tripling year over year; the number of customers building agent solutions using multiple models grew fivefold year over year, while revenue from open-source models on Foundry grew tenfold year over year.
Morgan Stanley believes that an increasingly rich array of model choices does not necessarily weaken Microsoft's platform value. On the contrary, when enterprises need to choose among different models and adjust task allocation based on quality, cost, and response speed, a unified model management and orchestration platform may become even more important. Microsoft also said on its earnings call that 90% of Copilot queries use Auto mode, with the system dynamically routing among different models to balance quality, speed, and cost. This also reflects how Microsoft is embedding multi-model scheduling capabilities into its own product system.
What is worth watching at this year's Ignite is whether Microsoft can further demonstrate Foundry's actual progress in model selection, agent development, deployment, and operational management, and whether these capabilities can bring more enterprise workloads and cloud service consumption.
3. Fabric: Can Enterprise Data Become a Differentiated Advantage in the AI Era?
As foundation model capabilities become increasingly widespread, an enterprise's true competitive advantage may depend more and more on whether it can make AI understand its own business data and workflows. Microsoft Fabric is an important part of this strategy. The platform aims to integrate enterprise data and provide agents with callable business information and context.
Microsoft disclosed in its latest quarter that paid Fabric customers exceeded 44,000, up more than 50% year over year; customers using both Fabric and Foundry exceeded 19,000. This means more and more enterprises are beginning to combine data platforms with AI application development platforms. Morgan Stanley believes that governed enterprise data and business context are important advantages distinguishing Microsoft from pure model providers. If enterprises are willing to migrate more data and workflows to Microsoft's platform, the synergy between Fabric and Foundry is expected to further drive Azure consumption growth.
4. Agent Adoption Drives Security Demand, Agent 365 and Enterprise Governance Worth Watching
As AI agents move from answering questions to executing tasks across applications, managing their permissions to access enterprise data, invoke tools, and perform operations will become more important. Microsoft is integrating identity authentication, permission control, behavior monitoring, and data protection capabilities into its agent management system. Products such as Entra, Defender, and Purview are expected to gain more enterprise demand as a result, while Agent 365 plays the role of managing, monitoring, and governing AI agents. Microsoft previously included Agent 365 in the Microsoft 365 E7 package. On the latest earnings call, Nadella said the number of agents registered with Agent 365 had exceeded 50 million, up more than 30% quarter over quarter.
For Microsoft, growth in the number of agents means not only potential cloud computing demand, but also that enterprises need new security and governance tools. If Microsoft can combine agent management with its existing enterprise security products, it may further encourage customers to upgrade to higher-priced enterprise software packages.
5. Maia 300: Can In-House Chips Improve AI Infrastructure Costs?
AI compute demand continues to grow, and data center construction and chip procurement are also putting enormous capital expenditure pressure on Microsoft. Therefore, whether in-house chips can reduce unit compute costs is an important variable affecting Microsoft's long-term profitability. Microsoft announced its next-generation in-house AI inference chip, Maia 300, in October, and plans to expand deployment of the Azure Maia series chips starting in 2027. The company expects Maia 300 to deliver about 40% more tokens per megawatt than the previous generation, with rack-level system token throughput reaching up to seven times that of Maia 200.
However, improved chip performance does not necessarily mean overall costs will fall immediately. Investors still need to watch its actual deployment pace, the workloads it supports, and its cost and efficiency advantages relative to third-party GPU solutions. Morgan Stanley believes that as Microsoft expands deployment of in-house chips, improved AI infrastructure efficiency is expected to ease gross margin pressure and enhance the long-term returns of the Azure business.
Gross Margin Under Pressure, but Can Operating Leverage Offset Cost Pressure?
The other side of Microsoft's rapid AI business growth is continually rising capital expenditure and infrastructure costs. Microsoft's first-quarter capital expenditure reached $45.2 billion, about three-quarters of which was for short-term assets such as CPUs, GPUs, memory, and storage. The company expects second-quarter capital expenditure to exceed $50 billion and calendar year 2026 capital expenditure to be about $170 billion. In the same period, the company's free cash flow was $18.6 billion. Microsoft said it expects free cash flow to remain positive in fiscal 2027, but also noted that it will continue to flexibly use financing channels based on funding costs and market conditions.
Morgan Stanley's model shows Microsoft's gross margin may gradually decline from 67.9% in fiscal 2026 to 65.9% in fiscal 2027, 64.5% in fiscal 2028, and 63.5% in fiscal 2029. However, the firm expects operating margins of 46.8%, 46.6%, 46.7%, and 47.2% over the same period, remaining broadly stable. This means Morgan Stanley expects Microsoft to offset some of the gross margin pressure from AI infrastructure investment through revenue growth, platform efficiency gains, and expense control. For investors, what really matters is not whether gross margin declines in the short term, but whether AI investment can generate enough revenue and operating leverage to keep operating margins resilient.
Behind the $650 Price Target: Valuation Depends on AI Commercialization Delivery
On valuation, Morgan Stanley maintains its "Overweight" rating and $650 price target on Microsoft. The base case is based on fiscal 2028 earnings per share of $24.41, corresponding to about 26 times earnings. Morgan Stanley's bull case target price is $800, corresponding to fiscal 2028 EPS of $27.41 and about 29 times earnings; the bear case target price is $300, corresponding to EPS of $22.25.
The firm believes Microsoft is expected to continue expanding its share of the enterprise AI market through the synergy among Azure, Foundry, Fabric, and Copilot. But that judgment also means Microsoft needs to keep proving that the commercialization progress of its AI products can match market expectations for its growth and profits.
The Biggest Test: Can Revenue Growth Keep Pace With Capital Expenditure?
Behind this year's Ignite, two conditions must hold true at the same time: on one hand, after expanding from seat subscriptions to usage-based billing, Copilot must truly drive revenue growth; on the other hand, Microsoft must improve AI infrastructure utilization to avoid capital expenditure continuing to climb without a corresponding improvement in investment returns.
In the first quarter, Microsoft's capital expenditure reached $45.2 billion, and second-quarter capital expenditure is expected to exceed $50 billion. Although Azure and Foundry continue to grow strongly, AI infrastructure investment remains enormous. Microsoft needs to prove to the market that every additional dollar of capital expenditure can generate more monetizable compute and higher revenue returns.
Therefore, what is most worth watching at Ignite is not a particular new model or a specific agent feature, but three more commercially meaningful developments: whether Copilot can raise ARPU through seat upgrades and usage-based billing; whether Foundry can become the core platform for enterprise agent deployment in a multi-model environment; and whether in-house chips and data center efficiency improvements can reduce unit compute costs.
For Microsoft, competition in enterprise AI is no longer just a contest of model capabilities, but a contest of full-stack commercialization ability from infrastructure to applications and from data to security. Whether Ignite can provide a clearer implementation roadmap will become an important window for the market to assess the return on its AI investment.
Comments