By Dan Gallagher and Asa Fitch
Even in the turbulent market of 2026, going from "AI winner" to "AI loser" status in six months seems excessive.
Unfortunately for Microsoft, the complexity of the company's business model also means its new label can't be shaken off quickly. But patient investors should be rewarded, as the world's largest software company by revenue still has notable strengths in the artificial-intelligence world.
While Microsoft's stock has rebounded some recently along with other megacap techs, it remains down about 17% over the past six months. The company has lost more than $800 billion in market cap from its peak in late October, putting the stock at under 24 times projected earnings. That is around the same multiple it fetched just before OpenAI launched ChatGPT in late 2022, according to FactSet data.
That launch made Microsoft the presumed early winner of the AI race, given its prescient investment of billions of dollars in Sam Altman's startup. But three years in, that race has racked up huge investment bills with less impressive business contributions -- at least apparent ones.
Looming large among Microsoft's concerns now is its suite of Copilot AI tools. Copilot got off the ground in late 2023, bringing AI features to Microsoft's ubiquitous productivity software, among other offerings. But while paying users of corporate Copilot software have grown to about 15 million in its latest quarter, they represent only about 3.5% of its enormous user base, according to Bank of America analysts.
That isn't tremendous 2 1/2 years in, especially when considering how the competition is faring. AI-developer Anthropic's Claude Cowork AI tools, which autonomously handle common corporate tasks like file management and compiling data, have taken off in recent months -- to the extent that Microsoft recently felt the need to add them to its Copilot suite. Paying fees to Anthropic when users tap into Cowork surely makes Copilot less lucrative for Microsoft.
The trouble with Copilot is to some degree a consequence of Microsoft's decision early in the boom to back OpenAI -- a much-hyped mass-market AI developer that didn't make corporate AI its top priority -- instead of waiting for an Anthropic to emerge.
That is a mistake, however, that Microsoft can overcome. The company has already demonstrated its adaptability by integrating Anthropic models into software offerings despite its close ties to OpenAI. And Microsoft's huge sales force and existing base of millions of customers gives it commercial leverage that upstart AI developers could only dream of. If and when Microsoft does find a successful formula for corporate AI, it will be off to the races.
Microsoft's other big problem is that growth in its Azure cloud-computing business has been slowing. Azure revenue grew 39% in its most recent quarter, down a percentage point from the previous quarter and below Wall Street's expectations. Analysts aren't expecting a big acceleration this quarter, either.
The limiting factor there has been capacity. Microsoft hasn't been able to build AI data centers fast enough to simultaneously run its own software offerings, develop its own AI models and lease out computing power at an accelerating rate. That isn't a good thing, to be sure: It means competitors are eating its lunch in serving AI workloads. But it is far better than a slowdown caused by lagging demand. Microsoft is also working to develop its own AI models, which consumes computing resources now but should lead to greater independence from OpenAI later.
Given the time it takes to build new data centers and fill them with servers, Azure's capacity issues aren't likely to be resolved soon even as Microsoft further grows capital spending, Evercore ISI analysts said in a note last month. A return to acceleration for Azure revenue is likely in the second half of the year, they said.
Getting Copilot off the runway will also likely take time -- years even. Investors will be looking impatiently at the strategy when Microsoft reports quarterly results next Wednesday.
The good news is Microsoft has time to figure it all out. Its corporate software offerings are broad, and customers can't readily abandon them for flashy AI alternatives. Its diverse stable of businesses, from cloud computing to the LinkedIn corporate social network, are collectively doing well enough to support more spending on its AI efforts.
After riding the crest of the first wave of the AI boom, Microsoft has lost momentum. But its prospects in the AI era haven't faded as much as many investors seem to think.
Write to Dan Gallagher at dan.gallagher@wsj.com and Asa Fitch at asa.fitch@wsj.com
(END) Dow Jones Newswires
April 23, 2026 05:30 ET (09:30 GMT)
Copyright (c) 2026 Dow Jones & Company, Inc.
Comments