Morgan Stanley maintains an "Overweight" rating on Microsoft Corporation (MSFT.US): the more AI models there are, the more valuable the "orchestration power" in Microsoft Corporation's hands becomes.

Title context: Morgan Stanley maintains an "Overweight" rating on Microsoft Corporation (MSFT.US): the more AI models there are, the more valuable the "orchestration power" in Microsoft Corporation's hands becomes. Text: At a juncture when the AI race is shifting from "whose model is the strongest" to "who can truly turn models into enterprise productivity," Morgan Stanley cast its vote for Microsoft Corporation (MSFT.US) with a latest CIO survey. The core conclusion of the survey can be summed up in one sentence: the deciding factor in enterprise AI is not the strength of any single model, but the three layers surrounding models"context, orchestration, and applications"and Microsoft Corporation happens to be the only company that can string these three layers into a closed loop. Based on this, Morgan Stanley maintained its "Overweight" rating on Microsoft Corporation and a $600 price target. A "multi-model world" is Microsoft Corporation's home turf Morgan Stanley's conclusion is built on a premise that is being confirmed by data: enterprises are no longer betting on a single model. The survey shows that currently 50% of CIOs have no AI workloads running on open-weight models, and that share is expected to plunge to 11% within three years; conversely, the proportion of enterprises expected to have more than 20% of workloads handled by open-source models three years from now will jump from 8% today to 33%. As models become "more numerous and more diverse," what gets elevated instead are the layers beyond the modelsenterprises need someone to provide data context, someone to select and configure the most suitable model by task, and someone to feed results into business processes. This is exactly Microsoft Corporation's three-piece combination of Fabric (enterprise context), Foundry (task orchestration and model routing), and Copilot (endpoint delivery). Morgan Stanley particularly highlighted one new data point: when asked about their preferred vendor for the "AI control plane/orchestration layer," 20% of CIOs chose Microsoft Corporation, exceeding the combined 17% of all large language model vendors and far surpassing Alphabet Inc. Class C (5%) and Amazon.com, Inc. (4%). In Morgan Stanley's view, this layer of "orchestration power" is a structural advantage that Microsoft Corporation has earned through enterprise penetration, CIO trust, and installed base as cloud vendors' capabilities converge. Incremental GenAI budgets: Microsoft Corporation remains the biggest beneficiary Another main thread of the survey is where the money is going. In 2026, 30% of CIOs expect Microsoft Corporation to receive the largest incremental share of GenAI spending, up 3 percentage points quarter over quarter, ahead of large language model vendors combined (16%), Alphabet Inc. Class C (8%), and Amazon.com, Inc. (7%). On a three-year horizon, Microsoft Corporation remains first at 28%, although its lead over model vendors has slightly narrowed from 13 percentage points to 7 percentage pointswhich precisely confirms that enterprises are rolling out "multi-model strategies." The old story of cloud migration has not changed either. On the net score for expected incremental IT budget share in 2026, Microsoft Corporation is +43%, ahead of Amazon.com, Inc. at +26% and Alphabet Inc. Class C at +4%; Azure firmly ranks first among public clouds with a 79% usage rate. The share of public cloud spending accounted for by AI/ML is expected to rise from 3.8% today to 9.9% in three years, providing long-term endurance for Azure demand. Token economics and the open-source wave The research report did not shy away from risks. 42% of CIOs are still unable to assess the return on AI token investment, 28% report returns below investment, 43% have already implemented token controls, and 31% plan to implement them before the second half of 2027. Morgan Stanley believes that this "growing caution in spending" will suppress consumption in the short term, but will instead benefit Foundryenterprises will need even more an orchestration layer that can schedule models according to cost and capability matching. On the earnings model, Morgan Stanley provided three scenarios: the base case of 25x FY28e EPS ($24.06) corresponds to $600; the bull case of about 29x corresponds to a target price of $795; the bear case of about 11x corresponds to a target price of $250. Under its base forecast, Microsoft Corporation's Azure growth, M365 commercial cloud growth, and operating margin all show resilience, with EPS expected to reach $24.06 in fiscal year 2028.