Nadella scoffs at AI giants as Microsoft may host China’s DeepSeek
Having long played the role of elder statesman in the AI race, Microsoft CEO Satya Nadella has had enough, it seems. In a new interview, he offers blistering criticism of a small group of companies that have dominated the often gloomy conversation.
-
Nadella rejects AI doom narratives, arguing cheaper models and broader choice will better earn public trust.
-
Microsoft may host DeepSeek despite criticism, signaling cost pressures now outweigh loyalty to frontier incumbents.
-
The tech giant has already rolled out a series of low-cost models to drive down consumer prices.
In a conversation with The Wall Street Journal, Nadella described his vision for the AI industry’s future as one of cheaper models, greater user control, and targeted messaging aimed at winning back the public’s trust.
The latter point seems as important as ever. In a new Pew Research poll, only 16% of respondents in the US said they believed AI would have a positive impact on society.
Gen Z adults, ages 18 to 29, are the most wary of AI, with 48% believing it’ll be negative for society – and, of course, for their job prospects.
AI visionaries like Anthropic CEO Dario Amodei keep sending warnings that the technology will wipe out millions of jobs. Amodei said last year that new AI systems could eradicate half of entry-level jobs by 2029.
Nadella – even though Microsoft is a major OpenAI backer – is fed up, it seems. In the interview, he makes it clear he’s not a fan of a small group of firms making dire prophecies about safety and jobs, but insisting they need vast resources for limitless expansion.
“You can’t say, hey, all white-collar jobs are gone, and this could even be a weapon, and we will use all the power to build data centers,” Nadella told The Wall Street Journal.
According to him, the public wouldn’t tolerate just a few models and companies “doing all of the learning for the world.” That’s why Microsoft is allegedly trying to change the course of the AI race.
Check if your data has been leaked
The tech giant has already rolled out a series of low-cost models to drive down consumer prices. Microsoft also released Copilot Cowork, an AI agent allowing users to choose cheaper AI models.
Crucially, Microsoft is now reportedly considering introducing a fine-tuned version of the Chinese open-source model DeepSeek into Copilot Cowork. It’d be an obvious low-cost alternative to models from OpenAI and Anthropic.
This month, Microsoft limited employees’ use of Anthropic’s Claude Fable 5. The public reason is the company’s new data retention requirements, but large firms such as Microsoft have recently been pulling back from using frontier AI models as they’re burning through budgets at a frightening speed.
Both OpenAI and Anthropic previously called out DeepSeek, an AI provider based in China, for distilling, or copying, their top models.
But that smells of hypocrisy as Anthropic, now valued at $965 billion, is itself facing multiple lawsuits accusing the startup of illegally using copyrighted internet data to train its systems.
Unlock more exclusive Cybernews content on YouTube.
