Microsoft has been the quiet name in the Magnificent Seven for most of 2026. The stock lagged. The AI spending kept climbing.

Investors kept waiting for the payoff. One Wall Street firm just said it started.

Stifel upgraded Microsoft from Hold to Buy on Sept. 23 and raised its price target to $575 from $530, implying roughly 15% upside from Tuesday’s close of $498.

Shares were up about 1% in premarket Wednesday. Analyst Brad Reback said Microsoft has “clearly turned the corner” after its June quarter, Investing.com reported.

This is not a firm chasing momentum. Stifel was on Hold. It watched the June quarter, ran its numbers, and changed its mind. That is a more meaningful signal than a firm that has been bullish all year adding another price target bump.

What flipped Stifel’s view on Microsoft

Three things moved in the June quarter. Azure growth came in stronger than expected. The cost of large-language-model research started falling. OpenAI’s revenue contribution stepped up.

All three at once was enough for Stifel to change its rating.

Reback is now calling for mid-to-upper-teens revenue growth over several years. Not a one-quarter read. A multi-year thesis built on the idea that Microsoft’s AI infrastructure decisions from three years ago are only now starting to show up in the numbers, Investing.com reported.

More Microsoft:

• Morgan Stanley resets Microsoft stock forecast ahead of earnings
• Bank of America doubles down on Microsoft stock ahead of earnings
• Citi revamps Microsoft stock price target for the rest of 2026

Stifel sees 200 to 300 more basis points of Azure upside ahead. Chips are getting more efficient. Model costs are falling. Software is getting tighter. All of that flows into better Azure margins.

The firm also made a counterintuitive point about open-weight models. Most people read them as a threat to Microsoft. Stifel reads them as a tailwind. More AI workloads in the world means more compute demand on Azure regardless of which model is running them.

CFO Amy Hood said dock-to-live times fell by more than 50% over the past year. That is the time between hardware going live and actually generating revenue. Cutting it in half means Microsoft gets paid faster on every dollar it invests in infrastructure.

At the spending levels Microsoft is running, that matters a lot.

Copilot seats and the GitHub play

Thirty million Copilot seats in the fourth quarter. That is up 10 million in a single quarter, the fastest seat gain since Copilot launched.

GitHub Copilot separately hit 50 million users in the same period. Both numbers landed on the same earnings call.

The ceiling people keep pointing to is Microsoft 365 itself, which has around 464 million commercial seats. Fair point. But Stifel’s answer to that ceiling is GitHub. Not Copilot. GitHub.

GitHub is moving away from fixed seat pricing toward consumption-based billing. A developer that runs AI coding tools heavily pays more than one who barely touches them.

The seat is the same. The bill is not. Microsoft earns more from the heavy user without selling a single additional license.

Microsoft has never run a developer tool at this kind of scale on consumption pricing before. The better the AI coding tools get, the more developers lean on them, the higher the bill. Stifel is betting that GitHub becomes a bigger part of the revenue story than most people are currently modeling.

Seats and usage. Both growing. Both inside the same productivity business. Stifel needs both to hit its numbers and it thinks both are running.

The margin story nobody expected

Earlier in the year Stifel had concerns about Microsoft’s margins. It now says those concerns were too negative.

In April, Microsoft revised its contract with OpenAI and cut the revenue-share payments it had been making. That alone removed a structural drag on gross margins.

The company also extended the useful life of certain assets from 15 years to 25 years. Annual depreciation falls when you spread the cost over a longer period. The asset does not get cheaper. The accounting looks better, Stocktwits reported.

Azure efficiency is moving the same way. Stifel now thinks Microsoft can fund the AI buildout from internal cash without going to external markets. Six months ago, that was not the assumption.

Headcount is down. Spending is tighter. The AI investment is still enormous but the company around it is running leaner.

What the risks still look like

Stifel has Microsoft hitting positive free cash flow in fiscal 2027. It also flags something worth noting: Some of that improvement might not come from discipline. It might come from constraints. Power. Land. Chips. Construction.

Microsoft cannot build faster than the physical world allows. When capacity limits cap spending, free cash flow improves whether the company wants it to or not.

Azure could slow if enterprises cut cloud budgets. That risk is sitting over every hyperscaler right now, not just Microsoft. Copilot could stall before monetization gets deep enough to matter.

OpenAI’s contribution already changed once in April. It could change again. And Alphabet and Amazon are not going anywhere.

Execution risk on the infrastructure side is real too. Data centers need power, specialized chips and physical space. Delays push revenue out. Hood’s dock-to-live improvement addresses that directly but does not eliminate it.

The upgrade comes down to whether June was a turning point or a bounce. Stifel says turning point. The next two quarters will settle the argument. But for a firm that sat on Hold all year, moving to Buy now is not a small call.

Related: Wells Fargo resets Microsoft stock price target for 2026