NEW YORK — Stifel analyst Brad Reback called it “getting back on track.” What the note actually said, in the numbers that enterprise sales teams care about, was more specific: Microsoft’s Copilot has crossed the engagement threshold that separates tools that survive budget reviews from ones that get quietly cut.
Reback upgraded Microsoft Corporation (NASDAQ:MSFT) from Hold to Buy on Tuesday, lifting his price target from $530 to $575, as Seeking Alpha reported. The call arrived on a day that was otherwise difficult for technology shares, with the Nasdaq falling 1.1% and the 10-year Treasury yield climbing to 5.11%, its highest level since 2007. MSFT ended the session at $500.84, up 0.57%, holding ground that most of the sector surrendered.
The engagement data is what makes the upgrade more than a routine target revision. Weekly active Copilot usage, Reback noted, now matches Outlook and Teams. In an enterprise software context, those are not arbitrary benchmarks. Outlook and Teams are the applications workers open at 8 a.m. and close at 6 p.m., the floor-level productivity infrastructure of modern office work. Copilot reaching the same weekly engagement rate means it has moved from optional AI layer to habitual work tool, at least among the companies that have already deployed it.
The pricing structure tells a more complicated story. Microsoft is offering 30% discounts for deployments of 1,000 or more Copilot seats and 50% for 10,000 or more, both taking effect in October. A discount of half on a flagship AI product, one whose engagement metrics are described as peer-level with Teams, is not a sign of weakness. It is a calculated bet on adoption velocity over near-term margin, timed to coincide with the rollout of a redesigned Copilot the company has confirmed will arrive this quarter as a unified “super app” incorporating AI coding tools and autonomous agent capabilities, as Fortune first reported. The logic: get seats deployed at scale, let the stickiness follow.
Azure remains the structural floor under the entire thesis. Reback cited accelerating Azure revenue growth as the primary driver for the target increase, alongside the OpenAI revenue-sharing arrangement that is beginning to compound. Artificial intelligence inference and training workloads are increasingly the high-margin use case that separates cloud winners from commodity providers, and Azure’s early partnership with OpenAI gave it a structural advantage in that segment. The arrangement means Microsoft doesn’t simply sell the infrastructure that OpenAI runs on; it participates in OpenAI’s commercial upside.
NVIDIA‘s chips power the compute layer that Azure depends on, and the GPU market’s trajectory shapes what Microsoft can offer customers on inference pricing. That dependency is baked into every cloud AI business model, Microsoft’s included.
The rate backdrop is the one variable no Copilot engagement metric addresses. The Federal Reserve has provided no rate-cut signal for the remainder of 2026, and with the 5-year Treasury touching 5% for the first time since 2007, the discount rate applied to high-multiple growth stocks is unfavorable by almost any historical measure. MSFT trading above $500 against that backdrop is a claim that the earnings case is compelling enough to absorb the yield headwind, not a sign that the market is ignoring it.
Sam Altman‘s OpenAI remains the variable in Microsoft’s AI narrative that no internal metric resolves. The commercial success of OpenAI’s products, and the pace at which enterprises choose OpenAI’s models over competing offerings from Anthropic, Google DeepMind, or Meta AI, directly shapes how much the revenue-sharing arrangement contributes to Azure growth. Microsoft’s structural position in enterprise AI predates most competitors by 18 months. What it does not control is OpenAI’s commercial ceiling, which means the Azure AI thesis is partially a bet on a company Microsoft doesn’t fully own.
Reback’s $575 target implies roughly 15% upside from Tuesday’s close. At current Treasury yields, that is a compressed risk-adjusted case, the kind of upgrade that reflects a changed near-term earnings view rather than a structural conviction on where MSFT settles in 2028. The next earnings report, expected in late October, will be the first real read on whether the Azure acceleration and Copilot engagement data Stifel described are showing up in revenue lines the way the engagement numbers suggest they should.
Until then, “getting back on track” is the thesis. What it means for the bottom line is the question the quarter answers.
