Satya Nadella, the CEO who once rebranded Microsoft as a “learn-it-all” company, now admits the empire he runs has become a “massive disadvantage” in the AI race. In a recent interview with Axel Springer CEO Mathias Döpfner, reported by Business Insider, Nadella confessed he spends his weekends studying how startups build products, because Microsoft’s sprawling bureaucracy has slowed its ability to move at AI speed. “This entire weekend, I spent all the time trying to get myself to understand how new companies are building products,” Nadella said, lamenting that at Microsoft, decisions are filtered through three divisional heads instead of one scrappy table of engineers.
That’s a striking admission from the man who helped turn Microsoft into a cloud powerhouse. But it also underscores a deeper truth: Microsoft was first to loudly trumpet its AI advantage through its OpenAI partnership, yet it now looks like it’s stalling out.
Microsoft’s Copilot branding was supposed to be the killer app of generative AI. Instead, adoption has plateaued. The user base hasn’t exploded in the way Redmond hoped, and the novelty of AI assistants is wearing thin. Meanwhile, as I recently reported, Google is speeding past Microsoft’s biggest AI partner, OpenAI, with Gemini’s rapid iteration cycles and tighter integration into its ecosystem. Nadella’s “study hall mode” feels less like visionary leadership and more like a CEO scrambling to relearn agility while his rivals sprint ahead.
The bigger issue is that most AI companies still haven’t found a profit lane. Despite trillions sunk into GPUs, data centers, and model training, the economics remain murky. Nadella himself warned that treating AI like a traditional IT upgrade is “going to fail by definition”. He’s right, but the irony is that Microsoft, with its layers of management and legacy systems, is precisely the kind of company most likely to make that mistake.
Let’s be blunt: Microsoft’s size is both its moat and its anchor. It can afford to pour billions into OpenAI and Anthropic, but it can’t replicate the nimbleness of a startup where engineers, scientists, and infrastructure teams sit at the same table. Nadella’s quote about needing to “unlearn the things that made you successful” is telling. The problem is, unlearning at Microsoft’s scale is like trying to turn an aircraft carrier with a canoe paddle.
And let’s be honest: with close to 200,000 employees, Microsoft will probably never be as nimble as a startup. But nimbleness isn’t even the larger issue here. The bigger question is whether AI itself is a sinking ship. If the economics don’t improve, Microsoft will need to figure out when to jump ship to mitigate billions in losses. Nadella’s weekend study sessions won’t change the fact that scale can’t save you if the entire industry is barreling toward an unprofitable future.
