According to Bloomberg Business, Microsoft’s stock dropped as much as 8.1% after its latest earnings report, even though sales and profit beat expectations. The problem? Capital expenditures hit a staggering $37.5 billion for the quarter, blowing past analyst forecasts of $36.2 billion. This massive spending, aimed at building AI infrastructure, is now a core concern for investors like Morgan Stanley’s Keith Weiss, who openly questioned the return on investment. The stock has fallen about 11% since its peak last October and had been flat for 2026 until recently. Meanwhile, Microsoft’s exclusive AI partnership with OpenAI has cooled, with OpenAI now also working with Oracle, Google, and Amazon, and Microsoft itself investing in rival Anthropic.
The ROI question is getting louder
Here’s the thing: spending billions on capex is one story. Proving it’s worth it is another. And right now, Wall Street isn’t seeing the proof. The narrative has shifted. It’s no longer about who has the flashiest partnership; it’s about who can actually monetize this tech and integrate it into products people love. Google, with its Gemini model, is apparently “crushing it” with consumer use cases and benefiting from its own, cheaper chips. Anthropic is making waves with its Cowork agent, which is so good that, according to The Information, even Microsoft’s own engineers are baffled by how well it controls Excel and PowerPoint. That’s a brutal reality check. Microsoft owns those programs and the mega-platform GitHub, yet it’s playing catch-up on the AI assistant front. So where exactly is all that Azure and capex money going if not into creating a definitive lead?
The AI landscape is polyamorous now
Microsoft’s early bet on OpenAI was a masterstroke that defined the last two years. But that era is over. OpenAI is shopping for compute elsewhere. Microsoft is hedging with Anthropic. Basically, exclusivity is dead. This “AI polyamory,” as Bloomberg puts it, spreads risk but also completely erases that first-mover halo. When everyone has access to top-tier models, your advantage has to come from something else—integration, cost, or a killer application. And at the moment, Microsoft seems to be losing ground on all those fronts. The sparkle is gone. Now it’s just another giant in a brutally expensive arms race, trying to justify its spending to skeptical shareholders.
Can Microsoft pull a Google?
There’s a weird irony here. Not long ago, everyone was writing Google’s obituary, saying OpenAI had made DeepMind irrelevant and doomed its search business. Look what happened. Google got its act together, streamlined, and is now the team to beat. The lesson? Corporate giants can turn it around when they feel enough heat. The pressure is now squarely on Satya Nadella. He has to prove Microsoft can innovate with the urgency of a startup, not just write checks. The blueprint is right there. But executing it within the layers of a behemoth like Microsoft is the real challenge. If they can’t, that capex number will keep growing, and the ROI question will become a deafening roar.
