Satya Nadella helped usher in the AI boom. Now he has a strong message for the companies leading the way.
Microsoft’s chief executive is joining a growing effort to take on artificial-intelligence giants OpenAI and Anthropic, outlining his vision for the next wave of the AI boom in an interview that includes cheaper models, more user control and political messages that win public trust.
Nadella offered a sharp critique of how the race for AI supremacy has shaped up, with a small group of companies capturing the value of world-changing technology as they make dire predictions about security risks and job losses, and insisting they need vast resources for unlimited expansion.
“You can’t say, hey, all the white-collar jobs are gone and this can also be a weapon and we will use all our power to build data centers,” Nadella told The Wall Street Journal. The public, he predicted, will not tolerate just a few models and companies “learning all the lessons for the world.”
While he didn’t directly name OpenAI, Anthropic or Alphabet’s Google — the three companies creating the most advanced proprietary models — he made clear that Microsoft is trying to move the AI race away from a future determined and controlled by frontier model-builders.
In just a few weeks, Microsoft has rolled out a suite of low-cost models to bring prices down for customers recovering from the sticker shock of skyrocketing AI bills. The company released Copilot Cowork, an autonomous AI “agent” that allows users to choose from a variety of AI models, including cheap ones, to complete long-running tasks.
Microsoft is considering whether to host a version of DeepSeek, an ultralow-cost AI provider based in China that OpenAI and Anthropic have asked to distill or copy their top models. Such a move would likely lead to a dramatic increase in usage for the Chinese model-maker, which could come at the expense of OpenAI and Anthropic, which face the prospect of a prolonged price war.
It’s a significant step for Nadella, who has long played elder statesman in the trillion-dollar AI race, joining a growing effort to turn the race away from the development of top models with ever-increasing capabilities.
Microsoft is one of OpenAI’s oldest partners and has invested billions to help build the company into the giant it is today. After years of tension, the two companies recently reached an agreement that allows OpenAI to expand its relationships with other big tech companies. Microsoft also signed a billion-dollar agreement with Anthropic last year.
Nadella said that every company has room to thrive and predicted that there will be very successful new companies. A Microsoft spokesperson said the company would continue to foster successful partnerships with OpenAI and Anthropic and that Nadella’s effort for an AI reset is not a “zero-sum game.”
Anthropic CEO Dario Amodei has made frequent comments about potential job losses caused by AI, including last year’s prediction that new AI systems could eliminate half of entry-level jobs by 2029. OpenAI’s Sam Altman also predicted significant job losses, but recently said he was “happy” to be wrong about that. OpenAI has published a series of policy proposals designed to help people see the potential benefits that AI can provide. Both leaders have raised dire warnings about security, and Anthropic has found itself in a tussle with the White House over the potential dangers of a wide release of a powerful new model.
Microsoft has lagged its competitors in developing its own AI systems. By late 2025, Copilot customers increasingly preferred other options like Google’s Gemini, according to market-research firm Recon Analytics.
Finding itself without a competing Frontier model, Microsoft has decided to use its deep pockets to join the ranks of companies trying to turn models into commodities. Axios previously reported that Microsoft was considering making a version of DeepSeek available on its Copilot platform.
Nadella’s interview outlining his plans to reshape the AI race follows an essay published on June 14 that outlined what AI-first companies will look like in the future.
In the interview he said the new model for AI deployment would be more democratic, in which social benefits would be widely shared and where companies would avoid reliance on a small group of marginal models.
Nadella criticized executives who have seen AI technology as a tool to reduce costs by eliminating jobs. “No, how do we think about reorganizing jobs?” He said. Referring to the main unit of measurement for AI output, he said companies should have both “token capital” – or in-house AI capability – and human capital. AI companies “should provide a recipe for how this can be done. Yes, it’s a lot of change management, it’s a lot of displacement, but there is a way,” he said.
Nadella described the approach to AI as a knowledge engine that helps companies leverage their employees and use their data, tapping a spectrum of models at different prices and capabilities. “All these models are going to climb a hill inside a machine that you control,” he said.
Companies will be characterized by “the tacit knowledge they have,” both human and AI, he said. In the future, companies will be “continuous learning systems” of both human knowledge and AI tokens. He said protecting intellectual property will be paramount to prevent businesses from becoming commoditized.
Fixing what’s wrong in the AI race will take more than a better message, he said.
Nadella pointed to the need to ensure that people feel they have agency and economic opportunity. “Now we have to work harder to earn social approval.”
Write to Bradley Olson at breadley.olson@wsj.com and Tina Li at tina.li@wsj.com.
