Two of Google's most important AI researchers left for rivals in a single week. Days later, Alphabet’s stock had its worst day in a year
Two of Google's most important AI researchers left for rivals in a single week. Days later, Alphabet’s stock had its worst day in a year.
On June 18, Noam Shazeer left for OpenAI. He co-led Gemini, Google's flagship model, and co-wrote "Attention Is All You Need," the 2017 paper behind nearly every large language model today. Google had paid $2.7 billion in 2024 to bring him back through Character.AI, and lost him inside two years. The next day, John Jumper left DeepMind for Anthropic. He shared the 2024 Nobel Prize for AlphaFold with Demis Hassabis, DeepMind's own CEO, and walked after nine years.
Alphabet fell about 7% on Monday and shed roughly $250 billion. The exits were not the whole story. AI-valuation nerves and a broad tech sell-off were already driving the tape. But the two departures crystallized the fear underneath all of it: that Google is losing the people who define the frontier.
Money was never the constraint. Google paid a premium to keep Shazeer and lost him anyway. The most credentialed researchers go where they believe the important work is happening, and two of them chose elsewhere in the same week.
This is key-person risk, but not the kind boards are built to see. We track it at the CEO level. The market has mature tools for that. It has almost none for the researchers two or three layers below the C-suite, whose grasp of frontier systems is, for now, close to irreplaceable. That gap is why the risk stays underpriced until a week like this forces it into view.
Alphabet has guided to about $190 billion of capital spending this year, funded by over $100 billion of new debt and equity. That capital buys compute. It does not buy the people who decide what to build with that compute.
If you own Alphabet, which line do you actually value most: the $190 billion of compute, or the small group of people who know what to point it at?
🔗 Source: Alphabet Has Its Worst Day in Over a Year on AI Concerns After High-Profile Exits
