Blab News
Google DeepMind Loses Two Leading Researchers to Rival AI Firms
In a headline‑making shakeup, two of DeepMind’s most influential engineers announced they are leaving the lab for rival AI firms. Noam Shazeer, co‑author of the transformer architecture and vice‑president of engineering, will join OpenAI. John M. Jumper, vice‑president of research and the 2024 Nobel Prize‑winning lead author of AlphaFold, is moving to Anthropic.
Both departures were reported in late June 2026. Shazeer’s exit was confirmed by several technology outlets, while Reuters, Bloomberg and TechCrunch covered Jumper’s announcement. Neither DeepMind nor the receiving companies has released details about the terms of the transitions.
DeepMind, a subsidiary of Alphabet Inc., has been a driving force behind recent AI breakthroughs. The lab pioneered reinforcement‑learning programs such as AlphaGo and AlphaZero, and created AlphaFold, a protein‑folding model that earned its creators a share of the 2024 Nobel Prize in Chemistry. Shazeer helped design the transformer model that underpins large‑language models like ChatGPT, Claude and Gemini.
The timing of the two exits highlights the broader AI talent war. OpenAI, Anthropic, Microsoft and Meta are all vying for researchers capable of building next‑generation models. OpenAI has recently expanded its product line beyond chatbots, adding scheduled‑task functionality to ChatGPT. Anthropic has added editing and synchronization features to its Claude design tools.
Some observers suggest that DeepMind’s broad portfolio—spanning Gemini, search, Android, cloud services, robotics, scientific research, AI agents and productivity tools—may dilute focus and make it harder to retain top talent. Shazeer and Jumper were key figures in projects that have defined DeepMind’s public profile.
Despite the departures, Alphabet continues to invest heavily in Gemini. The company rolled out a new Gemini 3.5 Flash release in early July, correcting a text‑looping bug and resetting weekly usage quotas for all users. The update demonstrates that DeepMind remains committed to improving model quality.
The exits do not signal a slowdown in DeepMind’s research agenda. The lab still operates multiple research centers in the United Kingdom, the United States, Canada, France, Germany and Switzerland, and it continues to develop generative AI tools such as Imagen, Veo and Lyria.
The broader AI industry is experiencing rapid innovation as firms compete for talent, users and market share. Developers have benefited from new tools such as GitHub Copilot, which was recently made generally available.
The impact of Shazeer’s and Jumper’s moves on the industry remains to be seen. Both researchers have shaped modern AI, and their future work at OpenAI and Anthropic could influence the direction of large‑model research for years.
For now, DeepMind’s leadership and infrastructure remain strong. Alphabet’s continued investment in Gemini and other AI initiatives suggests that the lab will continue to be a major player in the field.
The situation highlights the intensity of the AI race, with companies vying for researchers, technology and users. The outcome will likely affect the pace of innovation and the tools available to businesses and consumers.
The story is developing, and further details about the departures and DeepMind’s strategic plans are expected in the coming weeks.
Both departures were reported in late June 2026. Shazeer’s exit was confirmed by several technology outlets, while Reuters, Bloomberg and TechCrunch covered Jumper’s announcement. Neither DeepMind nor the receiving companies has released details about the terms of the transitions.
DeepMind, a subsidiary of Alphabet Inc., has been a driving force behind recent AI breakthroughs. The lab pioneered reinforcement‑learning programs such as AlphaGo and AlphaZero, and created AlphaFold, a protein‑folding model that earned its creators a share of the 2024 Nobel Prize in Chemistry. Shazeer helped design the transformer model that underpins large‑language models like ChatGPT, Claude and Gemini.
The timing of the two exits highlights the broader AI talent war. OpenAI, Anthropic, Microsoft and Meta are all vying for researchers capable of building next‑generation models. OpenAI has recently expanded its product line beyond chatbots, adding scheduled‑task functionality to ChatGPT. Anthropic has added editing and synchronization features to its Claude design tools.
Some observers suggest that DeepMind’s broad portfolio—spanning Gemini, search, Android, cloud services, robotics, scientific research, AI agents and productivity tools—may dilute focus and make it harder to retain top talent. Shazeer and Jumper were key figures in projects that have defined DeepMind’s public profile.
Despite the departures, Alphabet continues to invest heavily in Gemini. The company rolled out a new Gemini 3.5 Flash release in early July, correcting a text‑looping bug and resetting weekly usage quotas for all users. The update demonstrates that DeepMind remains committed to improving model quality.
The exits do not signal a slowdown in DeepMind’s research agenda. The lab still operates multiple research centers in the United Kingdom, the United States, Canada, France, Germany and Switzerland, and it continues to develop generative AI tools such as Imagen, Veo and Lyria.
The broader AI industry is experiencing rapid innovation as firms compete for talent, users and market share. Developers have benefited from new tools such as GitHub Copilot, which was recently made generally available.
The impact of Shazeer’s and Jumper’s moves on the industry remains to be seen. Both researchers have shaped modern AI, and their future work at OpenAI and Anthropic could influence the direction of large‑model research for years.
For now, DeepMind’s leadership and infrastructure remain strong. Alphabet’s continued investment in Gemini and other AI initiatives suggests that the lab will continue to be a major player in the field.
The situation highlights the intensity of the AI race, with companies vying for researchers, technology and users. The outcome will likely affect the pace of innovation and the tools available to businesses and consumers.
The story is developing, and further details about the departures and DeepMind’s strategic plans are expected in the coming weeks.
