OpenAI Lures Google's AI Pioneer Noam Shazeer Away Again

OpenAI has pulled off a major talent coup by hiring Noam Shazeer, the former Google vice president of engineering and one of the key builders of Gemini. Shazeer said on X that he is joining OpenAI and is “looking forward to working with the extraordinary team there.” He added that leaving was “a difficult decision,” saying he was proud of “the amazing team at Google and everything we built together.” Sam Altman replied that Shazeer was one of the people he had wanted to work with since OpenAI was founded, joking that it only took 10 years.

The move is especially notable because Google had paid $2.7 billion to bring Shazeer back, only to lose him again less than two years later. In August 2024, Google returned Shazeer and researcher Daniel De Freitas to DeepMind through a partnership tied to Character.AI, the startup they founded after leaving Google in 2021. They had originally departed after Google declined to aggressively back a chatbot project they supported.

Shazeer is widely regarded as a pioneer in AI. He began working at Google US in 2000 and became one of the company’s most important AI figures. He helped develop the Transformer architecture, a foundational building block for modern AI models and the “T” in GPT. Several of his papers are considered milestone contributions to AI development.

His departure comes amid an intense race among technology companies for AI talent. It also follows Google’s recent developer conference, where it unveiled new products including Gemini 3.5 Flash and the Gemini Spark AI agent. OpenAI, which makes ChatGPT, has also recently filed confidentially for an initial public offering, and is expected alongside Anthropic and SpaceX to be one of the most significant IPO stories in recent years.