Microsoft Is Still Down 7% While Azure Surged 43%: 23% Base Case, or a 93% AI rerating?

The setup: a 7% drop leaves room for a 23% base case and a 93% high-end rerating

The setup is simple: Microsoft still sits at $451.10 while Wall Street's consensus target is $555.77, or about 23% upside, and Arete's Street-high target is $870, which is where the 93% rerating case lives. That upside matters because the market has been pricing Microsoft more as an AI spending risk even though Azure grew 43% and the company reports $678 billion in commercial remaining performance obligations. The core bullish idea is that the infrastructure build was real, and the commercial demand is already visible.

This is not a clean buy-the-headline stock. After a drawdown of more than 30% from peak to trough, Microsoft staged a sharp rebound after its July 29 earnings report. But even after that move, the stock was still red on the year. The debate was never just whether AI was real; it was whether Microsoft could spend aggressively now and still show returns quickly enough to satisfy investors.

That is why the setup still matters. With shares down 7% for the year while Azure accelerated, the story looks more like an infrastructure-to-monetization rerating than a finished narrative. Consensus is the base case. The high-end target is what could happen if AI revenue converts faster than capex concerns fade.

Why the selloff looked more like a timing trade than a quality reset

The selloff was not mainly about losing confidence in Microsoft's core business. It was mainly about timing: when AI spending would stop looking like a cash drain and start looking like a revenue engine. The bear case had a clear trigger: free cash flow fell 6.46% to $66.99 billion while Microsoft accelerated AI compute buildout. Bears saw early cash-flow pressure. Bulls saw a profitable platform still funding a major investment cycle.

What changed after the earnings rebound

The shift came as the market moved from worrying about spending to modeling conversion. The scale is hard to ignore: Azure has crossed $100 billion in annual revenue, and Microsoft 365 Copilot has passed 30 million paid seats. Once those figures are in front of investors, the debate changes from whether Microsoft has real AI products to how quickly those products and existing enterprise relationships can turn into recognized revenue.

Why the OpenAI concessions matter

The renegotiated OpenAI agreement was not a clean win on every point. Microsoft's ownership decreased from 32.5%, and it no longer has a right of first refusal on new OpenAI cloud workloads. But the broader deal still left Microsoft with a 27% equity stake in OpenAI's new entity, valued at about $135 billion, while OpenAI committed to another $250 billion in Azure usage. That helps explain why the market kept paying for contracted demand, platform reach, and software distribution even while giving monetization more time to mature.

What bears are still watching: monetization may take longer than bulls hope

That is the real trap in the bullish story: the market can be right about demand and still be too early about economics.

The timing risk has not disappeared

The cleanest bear argument is not that AI demand is fake. It is that Microsoft may still be paying for the buildout before it fully controls the payoffs. The stock was once down 30% from record highs, and Forbes specifically flagged the Bear Case For 2026, including concern that Microsoft is spending heavily on AI cloud capacity while investors wait longer for stronger enterprise software returns. If conversion remains gradual, heavy infrastructure spending can look like overbuilding for longer than bulls want to admit.

The OpenAI renegotiation sharpens that point. Microsoft kept important commercial terms: OpenAI remains its frontier model partner, and Microsoft retained exclusive IP rights and Azure API exclusivity until AGI. Even so, the reduced stake and the loss of the cloud-workload right of first refusal show the partnership became less tightly aligned than it once was.

That is why the economics still need closer reading. Some investors see the renewed alliance as full confirmation; others see Microsoft paying heavily and then settling for access rather than complete rent-sharing control filings, not headlines.

The AGI verification clause is a real watchpoint

There is also a more tactical risk. Microsoft says the relationship remains strong and central, but the updated agreement now requires an independent expert panel to verify an AGI declaration. If that milestone arrives and shifts bargaining power quickly, today's exclusivity may not protect margins as cleanly as the most bullish version of the thesis assumes.

How to think about Microsoft from here

This still looks more like a staged setup than a chase. The stock offers roughly 23% upside to consensus, with a $870 high-end target only if monetization starts moving faster than capex concerns fade. That asymmetry supports owning some exposure now, while additional size makes more sense only if execution keeps improving rather than on another AI headline alone.

What would confirm the thesis

• Azure and Copilot adoption keep translating into revenue faster than capex worries recede.
• Commercial backlog continues to convert into recognized revenue without a long lag.
• The OpenAI relationship keeps delivering Azure demand and commercial leverage without further erosion of key advantages.

What would challenge the thesis

• AI spending continues to pressure cash flow while enterprise monetization stays slow.
• OpenAI's renewed independence gives the partner more flexibility and weakens Microsoft's exclusivity value over time.
• Investors decide the market for Microsoft's AI output is good, but not good enough to justify the current buildout.