A portfolio manager at a Boston mutual fund began June with Microsoft as his largest holding, a position built over two decades of reliable earnings growth. He spent the past four weeks trimming it. Microsoft shares are on course for their worst month since 2008, having lost more than a fifth of their value at the lows, and the company has seen roughly $530 billion of market value erased at the worst of the slide.
The decline is remarkable for a company whose fundamentals most investors still rate highly. Microsoft reported strong revenue growth in its most recent quarter, and its Azure cloud business remains one of the two largest in the world. The problem is no longer whether Microsoft is growing, but what it is spending to grow. The company has committed tens of billions of dollars a year to AI infrastructure, and investors have begun asking when that spending turns into profit rather than cost.
The math of the June selloff is unusually stark. At one point this month, Microsoft’s shares had fallen more than 20 percent from their high, a drawdown that would rank as the stock’s worst monthly performance since December 2000. A two-day rally into the end of the month recovered part of the loss, and the shares rose 1.8 percent on Monday, but the stock remains on track for its weakest month since the financial crisis of 2008.
The pressure reflects a shift in how investors evaluate the AI trade. For the past two years, the companies building AI infrastructure were rewarded for spending, with capital expenditure treated as a sign of ambition. That consensus has cracked. Cloud margins have come under pressure as AI capacity ramps ahead of revenue, and Microsoft’s own disclosures have shown that the cost of running AI workloads is rising faster than the revenue they generate.
Microsoft’s position is complicated by its relationship with OpenAI. The company has invested heavily in the startup and provides much of its computing capacity, but OpenAI has diversified to other clouds, and the financial terms of the partnership have been renegotiated as the relationship has evolved. Investors who once saw Microsoft as the surest way to own the AI boom now worry that it is the one paying for the infrastructure while the value accrues elsewhere.
The questions facing Microsoft are the questions facing all of big tech. How much of the AI build-out will translate into revenue, and how quickly? Which companies capture the profit from the boom, the ones selling compute, the ones selling models, or the ones selling applications? Wall Street has been asking these questions in private for months; the June selloff moved them to the center of the market’s attention.
Analysts are split on the answer. Bulls argue that Microsoft’s scale, its enterprise distribution and its installed base of Office and Windows customers give it the best position to convert AI spending into subscription revenue. Bears counter that the company’s core businesses are mature, that AI competition is intensifying and that the capital spending has no guaranteed return. Both sides agree on one thing: the next two quarters of Azure growth and margin data will settle the argument.
The selloff has made Microsoft cheaper in valuation terms than it has been in years, but cheapness has not stopped the decline, because the market is not pricing the current business so much as the uncertainty about the future one. A stock that falls 20 percent in a month on unchanged fundamentals is a stock whose investors have changed their mind about what they are willing to pay.
The scale of Microsoft’s spending is the heart of the matter. The company’s capital spending has roughly doubled over the past two years, and executives have told investors the pace will continue as long as AI demand justifies it. The money is going into data centers, chips and the software that runs them, and the returns are not yet visible in the margin rate that investors watch most closely.
The competitive backdrop has complicated the story. Amazon and Google have matched Microsoft’s AI ambitions, and the three cloud giants are spending against one another while prices for AI services fall. Microsoft’s advantage in enterprise software, the Office and Windows franchises that still generate most of its profit, has not translated into an unassailable position in AI, where customers can switch providers more easily than they can switch operating systems.
The next test comes with earnings. Microsoft will report its fiscal fourth-quarter results within weeks, and investors will scrutinize two numbers: Azure growth, which has been the stock’s engine, and the margin rate on cloud services, which has been its worry. A quarter that shows AI revenue beginning to offset AI costs could reverse the June slide as quickly as it began.
For the Boston portfolio manager, the calculus is simple even if the market is not. Microsoft is still his largest holding, but it is no longer his largest conviction. Whether he adds back to the position will depend on the same thing the whole market is watching: evidence, in the company’s own numbers, that the AI spending eventually pays for itself.
