Meta has recruited multiple former OpenAI researchers in recent days, according to reports from TechCrunch, The Wall Street Journal, and others. The moves center on Meta’s new AI “superintelligence” effort and include researchers linked to OpenAI’s Zurich office and work on AI reasoning and reinforcement learning.
Meta’s latest hires spotlight how intense the competition has become for top AI talent, with OpenAI CEO Sam Altman publicly claiming Meta has offered compensation packages worth more than $100 million to lure researchers. The reporting also shows Meta’s recruiting push is delivering some high-profile wins, even as other targets have reportedly turned the company down.
Trapit Bansal joins to work on reasoning models
TechCrunch reported that Meta hired Trapit Bansal, described as an influential OpenAI researcher, to work on AI reasoning models under Meta’s new AI superintelligence unit. OpenAI spokesperson Kayla Wood confirmed to TechCrunch that Bansal had left OpenAI, and Bansal’s LinkedIn page states he departed in June.
According to TechCrunch, Bansal joined OpenAI in 2022 and helped kickstart the company’s reinforcement learning work alongside OpenAI co-founder Ilya Sutskever. TechCrunch also said Bansal is listed as a foundational contributor on OpenAI’s first AI reasoning model, called o1.
TechCrunch wrote that Meta does not currently offer a public AI reasoning model, and that reasoning models are an important focus area because rivals have released highly capable reasoning systems. The report described reasoning models as systems trained to “work through problems” before responding, using extra time and compute to improve performance.
Three researchers linked to OpenAI Zurich move to Meta
A separate TechCrunch report said Lucas Beyer, Alexander Kolesnikov, and Xiaohua Zhai joined Meta’s superintelligence team, citing a Wall Street Journal report. TechCrunch described the trio as researchers who established OpenAI’s Zurich office.
SiliconANGLE also reported that Beyer, Kolesnikov, and Zhai joined Meta, and said the Wall Street Journal cited sources describing the trio as having helped establish OpenAI’s office in Zurich late the previous year. SiliconANGLE added that the three previously worked at Alphabet’s Google DeepMind.
Together, the reports show Meta’s hiring is not limited to a single recruit, but part of a broader push to staff the company’s new AI unit with experienced researchers from leading labs.
Big-money recruiting and a broader AI buildout
TechCrunch described Meta CEO Mark Zuckerberg as being on a hiring spree to build out Meta’s new AI team, and repeated Altman’s claim that Meta has offered $100 million compensation packages to researchers. TechCrunch said it was not clear what Bansal was offered.
TechCrunch also reported that Zuckerberg has personally reached out to many AI researchers, citing the Wall Street Journal’s description of WhatsApp outreach and recruiting dinners hosted at Zuckerberg’s homes in Palo Alto and Lake Tahoe. In the same TechCrunch report, Altman was quoted describing Meta’s efforts and saying he was “really happy” that, at least so far, none of OpenAI’s “best people” had accepted.
SiliconANGLE said the hiring comes after reports that Meta is forming a “superintelligence research group” tasked with developing AI models that can outperform humans across many tasks. SiliconANGLE also reported that the effort was launched after issues involving Llama 4 Behemoth, an internal large language model Meta previewed earlier, and that the company delayed the model’s launch because of performance concerns.
More names tied to Meta’s superintelligence effort
TechCrunch reported that Meta’s AI superintelligence unit includes leaders such as former Scale AI CEO Alexandr Wang. TechCrunch also said Meta has been trying to add former GitHub CEO Nat Friedman and Safe Superintelligence co-founder Daniel Gross.
SiliconANGLE separately reported that Meta hired Alexandr Wang and invested $14.3 billion in ScaleAI for a 49% stake. SiliconANGLE also said Meta was seeking to hire Daniel Gross and former GitHub CEO Nat Friedman, describing them as working at Safe Superintelligence, a company co-founded with former OpenAI chief scientist Ilya Sutskever.
The recent reports, taken together, paint a picture of a major staffing and investment push aimed at strengthening Meta’s AI capabilities—especially around “reasoning” systems—by bringing in researchers with experience building or advancing frontier models at OpenAI and other top labs.
