In a striking two-day stretch, Google DeepMind lost two of its most prominent scientists to direct competitors — a wave of departures analysts are calling the most consequential talent shift in AI research history. Alphabet’s shares dropped more than 5% on June 22 as investors processed the news.
Who Left and Where They Went
Noam Shazeer announced on June 18 that he is joining OpenAI as Lead for Architecture Research. Shazeer co-authored the 2017 paper “Attention Is All You Need,” which introduced the transformer architecture that underlies virtually every major language model in use today. He had returned to Google just two years ago — reportedly at a cost of $2.7 billion to the company — after co-founding the chatbot platform Character.ai.
Within 48 hours, John Jumper — who shared the 2024 Nobel Prize in Chemistry for developing AlphaFold, an AI system that predicted the three-dimensional structure of nearly every known protein — announced his own move to Anthropic after almost nine years at DeepMind.
Competitive Pressure at the Top
The departures land at an uncomfortable moment for DeepMind. According to a June 23 analysis by Fortune, the lab’s top models — Gemini 3.5 Flash and Gemini 3.1 Pro — currently sit outside the top five on widely watched AI benchmarks, trailing offerings from OpenAI and Anthropic. DeepMind took roughly four months to release Gemini 3.5 Pro after its previous frontier model; over the same window, Anthropic shipped two Claude Opus updates and launched an entirely new Mythos model family.
Former colleagues and industry observers point to growing friction between DeepMind’s scientific culture and the product-cycle pressures of Alphabet’s broader business. “There’s a sense that science is now less of a priority for Google DeepMind than in years prior to ChatGPT’s launch,” Fortune reporter Jeremy Kahn wrote, citing people familiar with the matter.
Why It Matters for Georgia
Georgia is working to build its own AI research ecosystem. According to the UNESCO Global AI Ethics and Governance Observatory, the country’s 2026 state budget sets aside approximately $18.4 million for an AI research and competence centre, research infrastructure, and talent-development initiatives running through 2029.
The DeepMind episode offers an instructive lesson for that effort: world-class researchers follow environments that prioritize scientific ambition over corporate caution. For Georgia’s new centre to attract and retain serious talent, it will need to create the conditions — intellectual freedom, ambitious research mandates, long-term funding horizons — that the world’s best AI scientists find worth staying for.