The Zhitong Finance App learned that Microsoft (MSFT.US)'s Ignite conference opens next week, and Morgan Stanley treats it as an inspection, not a press conference. The forward-looking report, released on Thursday, maintains an “overweight” rating and a target price of $650. The core of the report is not a new product, but whether a closed loop can be proven to be closed: Azure provides computing power, Fabric provides enterprise context, Foundry orchestration models and agents, and M365 and Copilot bring intelligence into employees' everyday workflows. Damo refers to these four layers as the “enterprise AI flywheel,” and believes that Ignite's value is to show that it is being transformed into wider production deployment and higher full-stack consumption.
The conference will be held in San Francisco from November 17 to 20. The core meeting is scheduled for November 17 at 1 p.m. EST (November 18 at 2 a.m. Beijing time), followed by seller Q&A with CEO Satya Nadella (Satya Nadella) at 4 p.m.; on November 18, at 3 p.m., Damo will meet Chief Financial Officer Amy Hood (Amy Hood) and Vice President of Investor Relations Jonathan Neilson (Jonathan Neilson).
The bottom line is good, but I need to explain where it is good
The report is based on the first fiscal quarter of the 2027 fiscal year published on October 28. Revenue for the quarter was US$92.3 billion, up 19% year over year, Azure grew 48% (6 percentage points faster than month on month), M365 Commercial Cloud accelerated to 18% (fixed exchange rate), and the operating profit margin was about 50%. According to Damo, this combination supports its judgment of “maintaining a ten percentage point higher and accelerating towards 20%” revenue growth, as well as profit growth of more than 20%.
However, 48% of Azure needs to be taken apart: the two market segments of enterprises and SMEs are strengthening at the same time, and core cloud services contribute together with AI infrastructure, and are not driven by a single customer type. Remaining commercial fulfillment orders (RPO) reached US$677 billion, an increase of 73% over the previous year, of which about 30% are expected to be confirmed within the next 12 months. This is a source of revenue visibility and a reason to dare to continue increasing capital expenditure.
Damo proposed five points of observation for this conference. These five points are actually its dismantling of the “flywheel”:
First, Copilot's next trip is ARPU, not the number of seats. The report's three pillars are: direct seat penetration, migration to the E7 package, and pay-as-you-go billing. The significance of the third pillar is that it can push Microsoft's revenue boundary from a “software budget” to a “knowledge work and business process budget.” The two figures quoted in the report are: Copilot's daily active users increased by about 50% month-on-month, and customers actively spending credits increased by more than 75% month-on-month.
Second, the value of Foundry shifts from “giving you a model” to “no matter how you use the model.” Foundry's revenue in the first fiscal quarter more than tripled year over year, and customers building multi-model agentic solutions increased 5 times year over year. According to this, Damo believes that more model choices are a good thing rather than a bad thing for Microsoft — because the more customers switch between multiple models, the more they need a unified orchestration layer.
Third, Fabric provides an “enterprise context.” When the model itself becomes a commodity, the governed enterprise data and context become a source of differentiation. More than 19,000 customers used both Fabric and Foundry in the first fiscal quarter, compared to 17,000 in the previous quarter. The growth rate is not that fast, but it points to the step where “customers are willing to move data in.”
Fourth, the popularity of agents has turned security into an entry point. As agents begin to access enterprise data and perform actions across applications, customers must manage their identity, permissions, and behavior. The correlation between Entra, Defender, and Purview increases, which also supports the rationality of upgrading M365 to E5/E7. Agent 365, launched by Ignite last year, was the starting point of this route.
Fifth, Maia 300 is a variable on the cost curve. On the premise that demand is still greater than supply, the deployment roadmap of self-developed chips, supported workloads, and cost advantages over third-party solutions are the keys to whether the gross margin of the Azure product layer can be improved.
Gross margins are declining, but the story isn't entirely here
Improvements at the Azure product level partially offset the structural pressure brought about by the increase in Azure's share of revenue; more importantly, incremental production capacity freed up by efficiency improvements was distributed to Foundry and quickly monetized — this also supported revenue growth and operating leverage.
According to Damo's model, the direction of this curve is clear: gross margin fell from 67.9% in fiscal year 2026 to 65.9% in fiscal year 2027, 64.5% in fiscal year 2028, and 63.5% in fiscal year 2029, but operating profit margins hardly declined — 46.8%, 46.6%, 46.7%, and 47.2%. In other words, the decline in gross margin has been eaten up by operating leverage — this is the key to whether this round of AI investment can be tolerated: the market accepts the decline in gross margin, provided that operating profit margins do not decline.
Target price and three scenarios
The $650 benchmark scenario corresponds to roughly 26 times earnings of $24.41 per share for fiscal year 2028. According to the report's caliber, 26 times is a small premium for the big-market software industry, and 1.2 times PEG is lower than Microsoft's history and peers. A bull scenario of $800 corresponds to about 29 times that of $27.41; a bear scenario of $300 corresponds to $22.25.
The probability measured by the implied volatility of options is to hit about 23.0% above $650, hit 6.1% above $800, and fall below $300 about 6.8%. The seller consensus was almost undivided: 97% increase, 3% neutral, 0% reduction.
Overlooked risk points
This forward-looking blind spot is exactly what the market is most concerned about this week: Microsoft stopped announcing the number of paid Copilot seats.
For the previous three consecutive quarters, paid seats (more than 30 million were last disclosed) were the most direct scale for the outside world to measure the progress of its AI implementation. However, in the first fiscal quarter earnings press release and conference call, this figure disappeared. Chief Financial Officer Hood told analysts that “no updates to Copilot figures were disclosed” this quarter, pointing instead to “revenue and usage momentum.” There is no basis for Nadella's “nearly 50% month-on-month increase” and “more than 75% month-on-month increase” for credit-consuming customers.
This coincided with Microsoft's pricing shift. The $99 M365 E7 kit and Agent 365 have pushed customers to a hybrid model of “seat fee+metering”; Copilot Business, sold through cloud solution provider channels, uses pay-per-use billing by default for new purchases and a default limit of 4,000 Copilot Credits per user per month — an arrangement that was originally scheduled to take effect on December 1, and was postponed to January to March 2027 after partner feedback.
Damo uses “daily activity +50%” as evidence of demand, but in the absence of seat numbers, what is the denominator of these two percentages determines how much they can prove. Forrester chief analyst Tracy Woo told the media that many organizations “are still trying to figure out what exactly this Copilot is useful for us” — “a fancy search tool” or an agent embedded in the workflow. This question is exactly the part that Ignite must answer, and Damo assumes it will be answered.
Two conditions that must be established at the same time
The report's logic requires two things at the same time: after Copilot switched to pay-as-you-go billing, revenue could keep up with consumption growth, and capital expenditure intensity did not increase. The capital expenditure for the first fiscal quarter was 45.2 billion US dollars (including financial leasing). The company indicated that the second fiscal quarter exceeded 50 billion US dollars and the full year of fiscal year 2027 was about 170 billion US dollars; free cash flow was 18.6 billion US dollars, down 27% year on year. It is the only one of the three major cloud vendors that remains positive.
So what is really worth listening to on Ignite is not how many proxy templates have been released, but three sentences: how does Copilot's measurement charge, can Foundry maintain the orchestration layer in a multi-model world, and can self-developed chips reduce the cost per unit of computing power. The first two sentences determine the ceiling of revenue, and the third sentence determines the floor of profit.
Citi gave a target price of $680 and a “buy” rating in another report. The logic is close to this report: Ignite is close to the strong first fiscal quarter, and Azure has reached an inflection point. The conference will revolve around infrastructure, its own cutting-edge model, and Foundry and CoPilot+IQ strategies.