Microsoft Held Up as AI Chips Sold Off. Is It the Safer Megacap Bet Now?

Year to date, Microsoft (NASDAQ: MSFT) has trailed most of its big-tech peers, with shares down roughly 11% in 2026. But today, the stock did something the chip sector could not: it held up. Microsoft finished the session modestly higher, even as a wave of selling hit semiconductor stocks, with Broadcom sliding double digits and memory and semiconductor names like Micron and Arm Holdings dropping as much as 7% as investors pulled back from the hot group of stocks.

That divergence raises a question: As money rotates out of the hardware names that have led the AI trade, is Microsoft -- with its software-heavy, more diversified path to monetizing artificial intelligence (AI) -- the steadier way to own the theme?

Image source: Getty Images.

Two different kinds of AI exposure

The chipmakers that sold off on Thursday share a vulnerability. Their fortunes are tied to a build-out cycle: orders for accelerators, memory, and networking gear that can swing hard when customers pause to digest what they have already bought. And when the pace of spending merely holds steady rather than climbing, these stocks -- which are priced for an AI boom -- could fall sharply, which is roughly what happened after Broadcom only reiterated its $100 billion AI chip revenue guidance rather than raising it.

Microsoft, however, has a more insulated business. It sells the cloud capacity and software that the chips ultimately run, and it collects much of that revenue through subscriptions and consumption that recur quarter after quarter. The company doesn't have to sell a customer a new graphics processor every cycle to keep the money coming in.

And the AI piece of the software giant's business is now sizable. In its fiscal third quarter (the period ended March 31, 2026), Microsoft said its AI business reached an annual revenue run rate of about $37 billion, up roughly 123% year over year. That growth is spread across customers building on its Azure cloud, frontier model companies, and its own Copilot products rather than concentrated in a single product line.

The breadth shows up elsewhere in the quarter, too.

Total revenue rose 18% year over year to $82.9 billion, and earnings per share climbed 23%. In addition, Microsoft's "Azure and other cloud services" revenue grew 40%.

And Microsoft's commercial backlog (contracted work not yet recognized as revenue) reached $627 billion, nearly double the prior-year figure.

"We are at the beginning of one of the most consequential platform shifts that will change the entire tech stack as agents proliferate and become the dominant workload," said Microsoft chief executive officer Satya Nadella during the company's fiscal third-quarter earnings call.

Big spending

Of course, Microsoft isn't immune to the AI spending cycle. The company is pouring enormous sums into the data centers and chips that power its cloud, with fiscal third-quarter capital expenditures of $31.9 billion -- up about 49% from a year earlier. And management has guided for capital spending to top $40 billion in the fiscal fourth quarter and pointed to about $190 billion across calendar 2026.

But Microsoft said in its most recent quarterly update that its cloud demand continues to exceed available capacity, and it expects to stay supply constrained at least through the end of the year. That points to revenue waiting on the other side of all this spending.

Still, the risk isn't gone. If enterprises slow their AI adoption, or if all this contracted demand converts to revenue more slowly or less profitably than expected, the heavy spending could weigh on margins -- and earnings -- for some time. Microsoft's backlog also leans heavily on a particular customer, OpenAI, leaving the company with significant customer concentration risk.

As of this writing, the stock trades at a price-to-earnings ratio of about 26, below where it has spent much of the past decade. For a company still growing revenue in the high teens, with profit growing even faster and a diversified hold on AI demand, that valuation looks reasonable. And after the year-to-date decline, the stock arguably leaves more room for error than the chip names do -- and it could prove more resilient than chip stocks if the AI boom starts to slow.

Daniel Sparks and his clients have no position in any of the stocks mentioned. The Motley Fool has positions in and recommends Broadcom, Micron Technology, and Microsoft. The Motley Fool has a disclosure policy.