Four People Left Google DeepMind. Alphabet Lost Billions.

Welcome to Staten News, where Google just learned that paying someone hundreds of millions of dollars doesn’t always mean they’re staying, and Alphabet’s market cap found that out the hard way.

Noam Shazeer announced on X that he was leaving Google for OpenAI.

Two days later, a Nobel Prize winner did the same thing.

One high-profile departure makes headlines. Two within 48 hours starts raising questions.

By the time Bloomberg reported that two more Gemini researchers were headed to Anthropic, Alphabet had lost roughly $269 billion in market value over several trading sessions between June 18 and June 24. While multiple factors influence a stock that size, the timing put an uncomfortable spotlight on Google’s ability to retain the very people building its AI future.

🧠 Who Left—and Why It Matters

Noam Shazeer co-authored Attention Is All You Need, the landmark 2017 paper that introduced the Transformer architecture powering nearly every major large language model today.

Google paid roughly $2.7 billion to acquire Character.ai in 2024 largely to bring him back as a Gemini co-lead.

Now he’s leaving again—this time for OpenAI.

He’s not alone.

John Jumper, recipient of the 2024 Nobel Prize in Chemistry for his work on AlphaFold, is reportedly heading to Anthropic after nearly a decade at DeepMind.

Joining him are Jonas Adler and Alexander Pritzel, both researchers who contributed to AlphaFold while also working on Gemini.

That’s not ordinary turnover.

That’s one research lab losing both the architect behind modern AI models and a Nobel Prize-winning scientist in the same week.

⚙️ It’s About More Than Money

The timing makes the story even more interesting.

Shortly before Shazeer’s departure, reports suggested Google reassigned compute resources from one of his projects to another DeepMind team in London.

In today’s AI race, compute is almost as valuable as talent.

If researchers can’t access enough GPUs to build frontier models, compensation alone won’t keep them happy.

That’s the real takeaway.

Shazeer and Jumper were already wealthy.

The bigger draw appears to be a combination of greater research freedom, access to computing resources, and pre-IPO equity packages that public companies simply can’t match in the same way.

⏳ Google’s Clock Is Ticking

The departures arrive at an awkward moment.

At Google I/O, Sundar Pichai said Gemini 3.5 Pro would arrive in June.

That deadline has come and gone, with the model still sitting in limited enterprise preview.

If it slips into July, Google risks launching alongside GPT-5.6 and Claude Opus 4.7 instead of beating them to market.

Meanwhile, Gemini 3.1 Pro has reportedly slipped behind competing models on several independent AI benchmarks.

Losing top researchers during a winning streak is a footnote.

Losing them while your flagship product is trying to regain momentum sends a much louder message.

🚀 The Winners

OpenAI lands one of the original architects of the Transformer just as it continues pushing frontier models toward its next phase of growth.

Anthropic adds a Nobel laureate alongside two more DeepMind veterans, strengthening its growing AI-for-science effort.

Both companies also offer something Google structurally cannot.

Private equity.

For elite AI researchers, that’s more than compensation.

It’s the possibility of turning career success into generational wealth if another IPO arrives.

💭 Final Thoughts

Google isn’t short on capital.

The company plans to spend roughly $190 billion annually on AI infrastructure.

But infrastructure alone doesn’t build breakthroughs.

People do.

This week wasn’t simply about recruiting.

It was a reminder that the engineers and scientists closest to the technology are making career bets on where they believe the next wave of AI innovation will happen.

And in Silicon Valley, sometimes the biggest market signal isn’t a product launch.

It’s who quietly changes the name on their email signature.

— The Bandicoots 📱🔌