Microsoft CEO Satya Nadella is in study hall mode.
Speaking with Mathias Döpfner, CEO of Axel Springer — Business Insider’s parent company — Nadella said he’s spending his weekends studying how startups build products. The reason, he said, is simple: Microsoft’s vast size has become “a massive disadvantage” in the race to build AI at startup speed.
“This entire weekend, I spent all the time trying to get myself to understand how new companies are building products,” Nadella said in an interview on the “MD MEETS” podcast, hosted by Döpfner and that aired on Saturday.
At young companies, he said, everyone involved in product development — from scientists to engineers to infrastructure teams — is “all sitting in one little table.” It means they’re able to make decisions on product, science, and infrastructure, and iterate instantly.
At Microsoft, he has “three divisional heads who manage those three things.”
Nadella’s comments come as many major tech companies — from Meta to Google to Amazon — are aggressively trimming middle layers of management in the name of speed.
It reflects a broader shift in Silicon Valley, where thick hierarchies are increasingly viewed as impediments to swift product decisions and AI experimentation.
Unlearning success to stay relevant
Nadella told Döpfner that thriving in the AI era requires unlearning habits that once made companies successful.
Big organizations and their leaders need to abandon the “know-it-all” mindset and adopt a “learn-it-all” approach, he said.
“The most important skill set for long-term relevance is — how do you be a learn-it-all and not a know-it-all,” Nadella said.
The challenge, he added, is that “you have to unlearn the things that made you successful to learn something new.” He also emphasized that empathy and emotional intelligence have become increasingly important in the AI era.
Internally, Nadella has already been pushing this cultural shift.
A leaked organizational chart reviewed by Business Insider showed how he has reorganized executives and has 16 direct reports — a lineup of handpicked lieutenants tasked with breaking down silos and accelerating the company’s AI shift.
The 4 things you need to fix first
Nadella said the majority of corporate AI projects collapse because companies approach AI like a traditional IT upgrade — a mistake he believes “is going to fail by definition.”
To make AI work, he said, organizations must be prepared to do four key things: rethink their workflows from the ground up, adopt modern AI tools, train employees to use them, and make sure company data isn’t stuck in old legacy systems.
Only companies that rebuild these foundations, Nadella said, will see meaningful gains from AI — and only leaders willing to unlearn old habits will be able to pull it off.
