Microsoft CEO Satya Nadella has been leading the charge in artificial intelligence (AI) for years, thanks to his Forming long-term alliance with OpenAI’s Sam Altman and the pioneering work of his own AI CEO Mustafa Suleyman, in particular copilot tool. But Nadella hasn’t often talked about the concern that has rattled Wall Street for much of the second half of 2025: whether artificial intelligence is a bubble.
At the World Economic Forum annual meeting in Davos, Switzerland, Nadella had a dialogue with the interim co-chairman of the forum, BlackRock CEO Larry Fink explained that if AI growth comes entirely from investment, it could be a sign of a bubble. “If we’re just talking about tech companies, that’s a clear sign if there’s a bubble,” Nadella said. “If we’re just talking about what’s happening on the technology side, then it’s purely a supply side thing.”
However, Nadella proposed a solution to the productivity dilemma, calling on business leaders to adopt new approaches to knowledge work by changing their workflows to accommodate the structural design of artificial intelligence. “The mindset we should have as leaders is that we need to think about using technology to change work—work processes.”
growing pains
This change is not entirely unprecedented, as Nadella points out by comparing the current moment to the 1980s, when computing revolutionized the workplace, opened up new opportunities for growth and productivity, and created a new working class. “We invented this category of things called knowledge work, where people started really using computers to amplify what we were trying to achieve with software,” he said. “I think the same thing will happen in the context of artificial intelligence.”
Nadella believes that artificial intelligence has “completely turned upside down” the way information flows in enterprises, replacing slow, hierarchical processes and forcing leaders to rethink their organizational structures. “We have an organization, we have departments, we have this expertise, and information keeps coming,” Nadella said. “No, no, it actually flattens the entire flow of information. So once you start doing that, you have to redesign it structurally.”
For some Fortune 500 companies, the transition may be more difficult, as structural changes can come with troubling growing pains. Lean companies will be able to adopt AI more easily because their organizational structures are fresher and more malleable, Nadella said. Large companies, on the other hand, may need time to adopt new workflows.
Despite widespread adoption of artificial intelligence, the 29th edition PwC Global CEO Survey The study found that only 10% to 12% of companies said they saw revenue or cost benefits from the technology, while 56% reported no benefit. This follows up on more pessimistic findings about AI returns From August 2025: 95% of generative AI pilots fail.
Mohammed Kant, Global Chairman of PwC, was interviewed wealthDiane Brady in Davos talks about the finding that many CEOs are wary and lack confidence at this stage of the AI adoption cycle. “Somehow, AI is developing so fast…that people forget that technology adoption must focus on the basics,” he explained. The survey found that companies that see the benefits of AI are “laying the foundations.” He believes that execution is more important than technology, and good management and leadership are indeed important for future development.
“For large organizations,” Nadella told Fink, “there’s a fundamental challenge: Unless and until you change as fast as possible, you’re going to be educated by the smaller organizations that can achieve scale with these tools.”
New entrants have the advantage of “starting over” and building workflows around AI capabilities, while larger companies will have to contend with the flattening effects of AI on entire departments and areas of expertise.
To be sure, Nadella said, large organizations have the upper hand, especially when it comes to relationships, data and expertise. However, he insists that companies must understand how to use these resources to change management styles, otherwise this may pose a major obstacle.
“The bottom line is, if you don’t translate that into a new production function, then you’re really going to get into trouble,” he said.
This story was originally published on wealth network