Google is losing the people who built Gemini. Top AI researchers Jonas Adler and Alexander Pritzel are leaving the company for Anthropic, according to TechCrunch AI, which cites a Bloomberg report. Both men played key roles in developing Google’s Gemini model, and their exit lands in the middle of a talent exodus that’s starting to look less like coincidence and more like a pattern.
What stands out here is the timing. These departures aren’t isolated. They’re the latest in a run of high-profile exits that have hit Google’s AI ranks over just a few weeks.
Who’s walking out the door
TechCrunch AI lays out a striking sequence of recent losses:
- Jonas Adler and Alexander Pritzel are heading to Anthropic. Both helped build Gemini.
- Noam Shazeer announced last week he’s leaving for OpenAI. Shazeer had been at Google since 2000, apart from the three years he spent building his chatbot startup, Character.AI. Google effectively acqui-hired that company for $2.7 billion, partly to bring Shazeer back to work on Gemini. Now he’s gone anyway.
- John Jumper, a Google DeepMind director, said days after Shazeer that he’s also joining Anthropic. Jumper won the 2024 Nobel Prize in Chemistry alongside DeepMind CEO Demis Hassabis for AlphaFold, the system that predicts 3D protein structures from amino acid sequences.
That’s a Nobel laureate, a 25-year company veteran, and two core Gemini contributors, all leaving in a matter of weeks. TechCrunch reached out to Google for comment.
Why this matters
Talent is the whole game in frontier AI. The number of people who can actually push a model like Gemini forward is tiny, and the companies fighting for them know it. When researchers of this caliber move, they don’t just leave a gap. They carry hard-won knowledge straight to a competitor.
Look at where they’re going. Anthropic is pulling in two of the four named researchers, plus Jumper. OpenAI grabbed Shazeer. These are Google’s two most direct rivals in the model race, and they’re stocking up on the exact people who know how Gemini was made.
This is significant because Google has spent enormous sums to keep this kind of talent close. The $2.7 billion Character.AI deal was, in large part, a Shazeer retention play. He left anyway. That tells you money alone isn’t holding the line.
The equity factor
There’s a clear reason the timing favors the challengers right now. Both OpenAI and Anthropic are preparing to go public, according to TechCrunch AI. That changes the recruiting math in a big way.
A pre-IPO offer comes with equity that could be worth a fortune if the listing goes well. Google can match a salary. It can’t easily match the upside of getting in before two of the most-watched IPOs in tech. For a researcher weighing a move, that’s a powerful pull, and it’s one Google can’t fully counter.
TechCrunch AI frames it plainly, with both companies heading toward the public markets, it’s a great moment to recruit top AI talent on the promise of equity. Expect the poaching to keep going.
What to watch next
A few things worth tracking from here:
- Whether Google responds. Counteroffers, new retention packages, or reorganizations inside DeepMind would signal how seriously leadership is taking this.
- Where the next exits land. If more Gemini and DeepMind names surface at Anthropic or OpenAI, the trend hardens into a genuine structural problem.
- The IPO windows. As OpenAI and Anthropic move closer to listing, the equity pitch gets sharper, and the recruiting pressure on Google rises with it.
Google still has deep research talent and the resources to rebuild. But losing the architects of your flagship model to the two rivals you’re racing isn’t a great look, and it’s happening fast. The next few weeks will show whether this is a rough patch or the start of something Google can’t easily stop. You can read the full report at TechCrunch AI.