Noam Shazeer Leaves Google for OpenAI: Why It's a Big Deal

The "father of the Transformer" and co-lead of Google's Gemini just defected to OpenAI. It's one of the most significant talent moves of the AI era — here's who Shazeer is, the twisty backstory, and why this shakes the industry.

In the AI arms race, the most valuable resource isn't chips or data — it's people. And on June 17, 2026, OpenAI landed one of the biggest names alive: Noam Shazeer, the Google vice president and co-lead of its Gemini models, announced he's leaving Google to join OpenAI.

This isn't just another executive shuffle. Shazeer is a co-inventor of the Transformer — the breakthrough that made modern AI possible. Having the "father of the Transformer" jump from Google's flagship AI team to its fiercest rival, right as OpenAI heads toward an IPO, is a thunderclap moment. Here's the full story.

What Happened

Shazeer confirmed the move himself in a post on X: "I'm excited to share that I'll be joining OpenAI and look forward to working with the exceptional team there." Until now, he served as a VP of engineering at Google and co-lead of the Gemini models — about as senior and central a role in AI as exists.

The timing is striking: it comes less than two years after Google brought him back into the fold, and as OpenAI races to widen its lead ahead of a reported public listing.

Who Is Noam Shazeer?

If you've used any modern AI chatbot, you've used Shazeer's work. He's widely regarded as one of the most brilliant researchers in the field — the kind of individual whose contributions shaped entire product categories.

  • Co-author of "Attention Is All You Need" (2017) — the paper that introduced the Transformer.
  • Co-founder of Character.AI — a conversational AI startup that hit a $1 billion valuation and 20M+ monthly users.
  • Google VP & Gemini co-lead — until this departure, helping steer Google's most important AI effort.
Illustration of the Transformer neural network architecture powering modern AI models

The Transformer: Why He Matters

To understand why this is huge, you need to know about the Transformer. In 2017, Shazeer and seven colleagues published "Attention Is All You Need," introducing a neural-network design built around a mechanism called attention.

That architecture became the foundation of virtually every major AI model today — OpenAI's GPT series, Google's Gemini, Anthropic's Claude, Meta's Llama, and countless others. The "GPT" in ChatGPT literally stands for Generative Pre-trained Transformer. In other words, a paper Shazeer co-wrote is the blueprint for the entire generative-AI boom. That legacy is why his name carries so much weight — and why he's often called the "father of the Transformer." (For how today's models built on that foundation compare, see our best AI models of June 2026 guide.)

The Twisty Backstory

What makes this move fascinating is that Shazeer has now left Google twice.

Year Event
2017Co-authors "Attention Is All You Need" at Google
2021Leaves Google after it declines to release his "Meena" chatbot
2021–2024Co-founds Character.AI — reaches $1B valuation, 20M+ monthly users
Aug 2024Google strikes a ~$2.7B licensing deal for Character.AI tech, bringing Shazeer back
Jun 2026Leaves Google again — this time for OpenAI

There's irony here: Google reportedly paid around $2.7 billion in 2024 largely to bring Shazeer and his expertise back. Less than two years later, that talent is walking out the door to the company Google is trying hardest to beat.

Why This Move Matters

Beyond the headline, the move signals several things:

  • A prestige blow to Google. Losing a Gemini co-lead — and an icon of the field — is a morale and reputation hit at a critical competitive moment for Gemini, which now even powers features like Apple's new Siri (see our WWDC 2026 recap).
  • A momentum boost for OpenAI. Snagging Shazeer just before an IPO signals strength to investors and strengthens OpenAI's research bench.
  • Talent is the real battlefield. With compute increasingly commoditized, the scarce ingredient is brilliant researchers — and they now move between labs like star athletes.

The AI Talent War

Shazeer's jump is the latest, loudest shot in an all-out war for AI talent. Through 2025 and 2026, OpenAI, Google DeepMind, Anthropic, Meta, and xAI have offered eye-watering pay packages and poached each other's senior people relentlessly. Meta reorganized around AI; xAI absorbed top engineers; and OpenAI has steadily pulled in marquee names.

When a single researcher's move makes global headlines and rattles a trillion-dollar company, it tells you how concentrated AI's value has become in a small number of people. For OpenAI, adding the Transformer's co-inventor is both a practical upgrade and a powerful statement: the momentum, for now, is theirs.

Frequently Asked Questions

Who is Noam Shazeer?

Noam Shazeer is one of the most influential AI researchers alive. He co-authored the 2017 paper "Attention Is All You Need," which introduced the Transformer architecture that underpins virtually every modern AI model, including OpenAI's GPT series and Google's Gemini. He later co-founded Character.AI and, until June 2026, was a Google vice president of engineering and co-lead of the Gemini models.

Why is Noam Shazeer leaving Google for OpenAI?

Shazeer announced on June 17, 2026 that he is joining OpenAI, saying he looks forward to working with its team. While he didn't give detailed public reasons, the move reflects the intense AI talent war, in which OpenAI — heading toward an IPO — is aggressively recruiting top researchers, even from rivals like Google.

What is the Transformer and why does it matter?

The Transformer is the neural-network architecture introduced in the 2017 paper "Attention Is All You Need," co-authored by Shazeer. It is the technical foundation of nearly all large language models today — GPT, Gemini, Claude, Llama and more. That's why Shazeer is often called the "father of the Transformer," and why his move carries symbolic weight.

What was Shazeer's history with Google and Character.AI?

Shazeer first left Google in 2021 after the company declined to release his Meena chatbot. He co-founded Character.AI, which reached a $1 billion valuation and over 20 million monthly users. In August 2024, Google struck a roughly $2.7 billion licensing deal for Character.AI's technology that brought Shazeer and a small team back to Google. He is now leaving again, less than two years later, for OpenAI.

What does this mean for Google's Gemini?

Losing a co-lead of the Gemini models — and one of the field's most respected researchers — is a notable blow to Google, both practically and symbolically, just as Gemini is competing hard with OpenAI and Anthropic. Google has deep AI talent and Gemini development continues, but Shazeer's departure is a morale and prestige hit in a fierce competitive moment.

Why is OpenAI hiring so aggressively in 2026?

OpenAI is in a high-stakes race with Google, Anthropic, Meta, and xAI, and is reportedly heading toward an IPO. Securing marquee researchers like Shazeer strengthens its models, signals momentum to investors, and denies rivals their best people. Top AI talent has become one of the scarcest and most valuable resources in tech.

How significant is this for the AI industry?

Very. When the co-inventor of the Transformer leaves the company whose flagship model he helps lead to join its biggest rival, it underscores how brutal the AI talent war has become and how much momentum OpenAI carries into its IPO. Individual researchers now move markets and morale across the entire industry.

Top AI labs competing for elite researchers in the AI talent war

Final Thoughts

Noam Shazeer leaving Google for OpenAI is the rare personnel move that genuinely matters to the whole industry. He helped invent the technology the entire AI era is built on, he's steered one of the world's top models, and now his talent strengthens the company best positioned to capitalize on the moment.

For Google, it's a stinging loss at a delicate time. For OpenAI, it's a statement of momentum heading into its IPO. And for everyone watching, it's a vivid reminder that in 2026, the fiercest competition in tech isn't over models or money — it's over the handful of minds who can build the future. We'll keep tracking the talent war as it reshapes the AI landscape.