Meta Acquires Three Top Researchers from OpenAI's Zurich Office in a Major Talent Shift}
Meta has aggressively recruited top AI researchers from OpenAI's Zurich office, including three ViT authors, amid ongoing competition in AI talent acquisition, as reported by WSJ.
Congratulations to Meta.
Recently, The Wall Street Journal published an exclusive report revealing that Meta has essentially absorbed the newly established OpenAI Zurich office, which was formed last year.
Specifically, Meta CEO Mark Zuckerberg recruited three researchers: Lucas Beyer, Alexander Kolesnikov, and Xiaohua Zhai. These researchers have been close collaborators, notably involved in important studies like ViT. They left Google DeepMind's Zurich office in December last year to join OpenAI and establish the Zurich branch. Details can be found in our previous report: Just now, three Google Vision Transformer authors officially joined OpenAI.
The WSJ states: "An OpenAI spokesperson confirmed that these three researchers have left the company."
Following recent model release failures, Zuckerberg has been actively recruiting top talent, aiming to regain Meta’s edge in AI. Recently, Meta has been spending heavily to attract researchers.
According to reports, Zuckerberg has been emailing and messaging top AI talents daily via email and WhatsApp, personally reaching out to hundreds of researchers, scientists, infrastructure engineers, product stars, and entrepreneurs to join his new superintelligence lab. His offers are highly attractive, including up to $100 million in benefits for some.
Some recipients were shocked and thought it was a prank, not replying for days.
Recently, Meta invested $14 billion in AI startup Scale and hired its CEO, Alexandr Wang, to lead its new superintelligence team. Details are in our previous report: Just now, Scale AI CEO Alexandr Wang officially announces: Meta’s massive investment and poaching.
Zuckerberg also attempted to recruit OpenAI co-founders Ilya Sutskever (founder of Safe Superintelligence Inc.) and John Schulman, but was unsuccessful.
OpenAI CEO Sam Altman commented at an event on Tuesday that he was not worried about Zuckerberg’s blitz: "It’s like Zuckerberg is doing some new crazy stuff. So what?" Last week, Altman said that his top talents had not left for Meta. Despite this, OpenAI is also countering by providing more funding and growth opportunities for its researchers.
Here’s a brief introduction to the three main figures in this news:
Xiaohua Zhai
Personal homepage: https://sites.google.com/view/xzhai
According to LinkedIn, Xiaohua Zhai joined OpenAI in 2025. Previously, he was a senior research scientist and manager at Google DeepMind Zurich, leading a multimodal research team focused on multi-modal data, open-weight models, and inclusivity.
He earned his PhD in computer science from Peking University in 2014. He worked at Google for three years as a software engineer, then joined DeepMind as a research scientist in December 2017, working there for 7 years before moving to OpenAI.
His Google Scholar citations exceed 80,000, mostly from their joint ViT paper: “An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale”.
Lucas Beyer
Personal homepage: https://lucasb.eyer.be/
He graduated from RWTH Aachen University in Germany in 2018, interned at Google, worked as an AI engineer at Kindred.ai, and was a research assistant at RWTH. He joined Google and later DeepMind, then moved to OpenAI in 2024.
His Google Scholar citations exceed 80,000.
Alexander Kolesnikov
Personal homepage: https://kolesnikov.ch/
He earned his master’s degree from Moscow State University in 2012, then obtained a PhD in machine learning and computer vision from the Austrian Institute of Science and Technology. He worked at Google Brain and DeepMind before joining OpenAI in 2024, focusing on multimodal AI research.
His Google Scholar citations exceed 90,000.
Reference links: