Google Just Lost Its Nobel Prize Winner to Anthropic — and $225 Billion Vanished Overnight

John Jumper won a Nobel Prize for building AI that cracked one of biology's hardest problems. Last Friday, he walked out of Google DeepMind and into the arms of Anthropic — a company Google's own parent company holds a 14% stake in.

The move, announced on X on June 19, wasn't some quiet transfer. It came days after Noam Shazeer — Google's VP of Engineering and co-lead of its Gemini models — bolted for OpenAI. Two senior departures in one week. Alphabet's market cap reportedly shed $225 billion in a single trading session.

Not a typo. Two hundred and twenty-five billion. Gone.

Look, talent moves happen in tech all the time. But a Nobel laureate leaving the lab where he won it? For a rival that's partially funded by your own investor? That's not a talent move. That's a statement.

Who Is John Jumper and Why Should You Care?

John Jumper isn't a household name. He should be.

He co-created AlphaFold, the AI system that predicts 3D protein structures from genetic sequences. Before AlphaFold, figuring out a single protein's shape could take years of lab work and millions of dollars. AlphaFold predicted over 200 million of them. For context — that's basically every protein known to science.

In 2024, Jumper and DeepMind CEO Demis Hassabis shared the Nobel Prize in Chemistry for that work. The Nobel committee called it a breakthrough that "cracked the code." Biologists called it the end of a 50-year-old grand challenge in their field.

According to Business Insider, Jumper held the title of VP, Engineering Fellow at DeepMind. He'd been there nearly nine years. Hassabis reportedly gave him the AlphaFold team lead just six months after he finished his PhD.

Six months out of grad school. Leading the team that would win a Nobel.

That's the person Anthropic just hired. And Google just lost.

The Week That Wiped $225 Billion Off Alphabet

Here's what makes this more than a one-person story. Jumper's departure didn't happen in isolation.

Days earlier, Noam Shazeer left Google for OpenAI. Shazeer is a heavyweight in his own right — co-founder of Character AI, which Google acquired in a $2.7 billion licensing deal in 2024. After rejoining Google post-acquisition, he became VP of Engineering and co-lead of the Gemini AI models. You know, Google's flagship AI product line. The one they're betting the company on.

So in the span of about a week, Google lost:

• The co-lead of its main AI model (Shazeer → OpenAI)
• A Nobel Prize-winning researcher (Jumper → Anthropic)
• And somewhere in the neighborhood of $225 billion in market capitalization

Investors noticed. Whether the sell-off was purely about these exits or just the straw that broke the camel's back doesn't really matter. The signal was sent: Google's AI talent is leaking, and the market doesn't like it.

This isn't the first AI talent shake-up we've covered. Earlier this month, we wrote about how OpenAI's enterprise chief walked out after just five months. The talent carousel in AI is spinning faster than ever, and nobody seems to have a grip on it.

The Irony: Google Is Funding Its Own Rival

Here's where it gets genuinely weird.

Alphabet — Google's parent company — holds roughly a 14% stake in Anthropic. That stake came from Google's early investment in the AI safety startup. So when Jumper walks from DeepMind to Anthropic, he's technically moving to a company that Google partially owns.

Google is paying to fund the lab that's hiring away its Nobel laureates.

Let that sink in for a second.

Now, Google's investment in Anthropic was always a hedge. Bet on the incumbent (DeepMind) and the challenger (Anthropic) at the same time. Classic big-tech portfolio strategy. But the strategy assumes your talent stays put. When your best people start using your own investment dollars as a launchpad to leave? That hedge starts looking like a self-inflicted wound.

Anthropic, meanwhile, is on a tear. The company's revenue run-rate has reportedly surpassed $30 billion, up from about $9 billion at the end of 2025. Over 1,000 customers now spend $1 million or more annually. That count doubled in under two months. The company is hosting a science event on June 30, and rumors suggest Jumper's first public Anthropic appearance will happen there.

AlphaFold's Legacy — and What Jumper Actually Brings to Anthropic

AlphaFold wasn't just a cool science project. It fundamentally changed how drug discovery works. Pharmaceutical companies used to spend years on protein structure determination. AlphaFold compressed that to minutes. Researchers working on diseases from Alzheimer's to malaria cited AlphaFold as a accelerant for their work.

So what does Anthropic want with the guy who built it?

A few possibilities, none of them confirmed because Anthropic hasn't disclosed Jumper's title or start date:

AI for science. Anthropic has been positioning itself beyond chatbots. A Nobel-winning scientist gives them instant credibility in the research world. If they're building models for scientific discovery — drug discovery, materials science, climate modeling — Jumper is the ultimate hire.

Model architecture. AlphaFold wasn't just a product. It was a masterclass in applying transformer architectures to a domain-specific problem. That kind of thinking transfers to building better foundation models.

Talent magnet. Nobel laureates attract other top researchers. If Anthropic is building a science team, Jumper is the anchor.

Signal to investors. Anthropic is reportedly heading toward its own IPO. A Nobel Prize winner on the team is the kind of thing that looks very good in an S-1 filing.

Google's Response: Polite, But the Subtext Is Loud

A Google DeepMind spokesperson told Reuters: "We are grateful for John's significant contributions to Google DeepMind's work in advancing science and AI. We wish him well in his next chapter."

That's the corporate equivalent of smiling through gritted teeth.

Hassabis himself replied to Jumper's departure post on X with something warmer: "What we achieved with AlphaFold changed the world, and showed the field what was possible with AI for science and medicine, lighting the way for how AI can benefit humanity."

Jumper, for his part, was gracious. He called DeepMind "a special place" and said he'd "still be excited to hear about what amazing things they discover next." No drama, no bridge-burning. Just a clean exit.

But the timing is brutal for Google. The company is already fighting on multiple fronts — Anthropic's regulatory battles with the US government, OpenAI's IPO momentum, and the ongoing question of whether Gemini can actually compete with Claude and GPT at the frontier. Losing two senior leaders in a week doesn't help any of those narratives.

Comparison: Google DeepMind vs. Anthropic — What Each Lab Is Working With

The Bigger Picture: AI Talent Is the New Oil

There's a reason these moves matter beyond the gossip column. The AI industry is in a phase where model quality increasingly depends on who builds them, not just how much compute you throw at the problem.

Sure, scaling laws still hold. More data, more chips, bigger models — that recipe still produces gains. But the marginal returns are shrinking. The next big leap in AI isn't going to come from doubling GPU clusters. It's going to come from architectural innovations, new training methods, and people who can see around corners.

That's why a guy like Jumper is worth more than a data center. He saw around a corner that the entire field of structural biology had been staring at for half a century. You can't buy that with compute.

And the market knows it. When the person who solved protein folding leaves the company that funded the solution, investors don't think "oh well, the IP stays." They think "the next AlphaFold won't be built here." Whether that's rational or not is almost beside the point. Perception drives valuations, and right now, the perception is that Google can't hold onto its best people.

Meanwhile, SK Hynix is preparing a $29.4 billion US IPO to fund more AI memory chip production. The entire supply chain — from chips to talent to models — is being reshuffled in real time. The companies that lock in both the hardware and the humans are going to win the next decade. The ones that lose either? They're going to be footnotes.

What to Watch: June 30 and Beyond

Anthropic's science event on June 30 is the next data point. If Jumper shows up and announces a research agenda — especially one focused on AI for scientific discovery — it'll confirm what many suspect: Anthropic is building a serious science division, and they just poached the guy to lead it.

For Google, the question is whether these exits are a trend or a blip. DeepMind still has Hassabis. It still has thousands of researchers. It still has more compute than anyone except maybe Microsoft. But morale matters. And when your Nobel laureate leaves for a rival that you're technically invested in, morale takes a hit that no press release can fix.

As for the broader AI talent market? Don't expect this to slow down. The stakes are too high, the money is too big, and the people who can actually build frontier AI are too rare. Every lab is one LinkedIn post away from a market-moving event.

We'll be watching.

Sources

• CNBC — "John Jumper to leave Google DeepMind for Anthropic" (June 19, 2026)
• Business Insider — "Nobel-Winning AlphaFold Pioneer Leaves Google DeepMind for Anthropic" (June 2026)
• AIToolsRecap — "AI News June 25, 2026: John Jumper Joins Anthropic, SK Hynix $29.4B US IPO, Claude Tag for Slack"