Davos: Why AI's Top Leaders Started a Public Brawl

The heads of the three biggest AI labs — Google DeepMind, OpenAI, and Anthropic — traded public jabs this week at the World Economic Forum in Davos. The Verge compared the scene to the US presidential primaries: instead of the usual talk about the good of humanity, the industry's leaders waged a reputation war. The trigger was OpenAI's plan to test ads in ChatGPT. Google DeepMind CEO Demis Hassabis didn't miss the chance to hint that his rival is simply short on money.

Hassabis' remark: "Strange that they went for it so early"

It all started on Tuesday. In an interview with Alex Heath, author of the Sources newsletter, which runs in The Verge once a week for subscribers, Demis Hassabis commented on OpenAI's decision to test ads in ChatGPT. "It's interesting that they've gone for that so early," Hassabis said. "Maybe they feel they need to make more revenue." The line was textbook passive aggression. In a world where training a single model costs hundreds of millions of dollars, hinting at a rival's financial trouble is an attempt to scare off investors and poach talent. Hassabis was effectively saying that the market leader's business model is coming apart at the seams.

Amodei piles on: "the end of the romantic era"

The next day, Wednesday, Anthropic CEO Dario Amodei joined the conversation. His company has long pitched itself as the safe alternative to OpenAI. Once the heads of the three biggest labs — DeepMind, Anthropic, and OpenAI — ended up in the same snowy town, the friendly masks came off. Observers describe it as the end of AI's romantic era, when everyone pretended to be working toward a shared future for humanity. Now it's a hard-nosed business where market share matters more than ethics statements, and lab leaders behave like primary candidates, less focused on selling their own product than on undermining the competition's reputation.

Why now: the money is running out faster than the breakthroughs arrive

The answer is simple. Investors are starting to ask awkward questions about monetization and returns on the billions they've poured in, while new technical breakthroughs arrive more slowly than the money gets spent. As the pioneer, OpenAI takes every hit. If the company introduces ads, it will wreck the user experience people are used to; if it doesn't, it may not be able to cover Microsoft's cloud computing bills. Hassabis and Amodei understand this perfectly well and are hitting the sorest spot, trying to cast their own companies as more stable, cleaner players next to a "desperate" leader.

From existential risk to ad revenue

The shift in the industry's rhetoric is telling. The conversation used to be about existential risk and a superintelligence capable of taking over the world. Now it's about ad revenue, user acquisition costs, and quarterly reports. This return to earth shows the industry is maturing, but it also means losing the mystique that once helped it raise capital without too many questions. Davos showed that unity in the AI community is a myth built for regulators. Behind closed doors and on discussion panels, the labs are fighting over who becomes the face of progress in the eyes of governments, and PR strategy is becoming just as important as neural network architecture.

What did Hassabis say about ads in ChatGPT?

On Tuesday, in an interview with Alex Heath (the Sources newsletter for The Verge), Demis Hassabis said OpenAI's decision to test ads in ChatGPT looks premature: "It's interesting that they've gone for that so early. Maybe they feel they need to make more revenue." It was the first public remark, and it kicked off a wave of mutual jabs between the labs in Davos.

Why have AI lab leaders started publicly criticizing each other?

Because training models costs hundreds of millions of dollars, and investors want answers about monetization and returns. As the market pioneer, OpenAI was the first to face the dilemma: introduce ads and lose user trust, or skip them and struggle with its cloud computing bills. Its competitors, DeepMind and Anthropic, are exploiting that vulnerability by publicly hinting at OpenAI's financial instability so they look more reliable to investors and regulators.

What does this mean for the AI market?

It marks a shift from a showy united front for "saving humanity" to open competition for market share and reputation. The AI industry is coming back down to earth: instead of talk about superintelligence, it's talk about ad revenue and quarterly reports. Davos exposed that, behind the scenes, the labs are fighting for status — over who becomes the face of progress in the eyes of governments — and that fight could reshape the industry as soon as the end of the year.