Meta faces internal rift as elite AI lab challenges company priorities

Wednesday 10 December 2025 - 15:50

By: Dakir Madiha

Meta faces internal rift as elite AI lab challenges company priorities

The rise of Meta’s elite artificial intelligence division has deepened internal divisions over the company’s strategic direction, exposing tensions between innovation and commercial focus. As CEO Mark Zuckerberg accelerates his push toward superintelligence, competing visions within the company are testing its cohesion at a decisive moment.

Power shift within Meta

Meta’s advanced research group, known internally as TBD Lab, was established earlier this year after the company’s $14.3 billion investment in Scale AI. Led by Chief AI Officer Alexandr Wang, the unit operates near Zuckerberg’s office in a space designed to symbolize transparency and innovation. Yet behind the glass walls, senior executives have locked horns over how Meta should balance fundamental AI research with its core business operations.

Chief Product Officer Chris Cox and Chief Technology Officer Andrew Bosworth reportedly pressed Wang’s team to use data from Facebook and Instagram to advance the next-generation AI model, code-named Avocado. Their vision aims to strengthen Meta’s advertising and recommendation algorithms. Wang, however, argued that such product-driven goals could limit the company’s progress toward building true superintelligence. He believes Meta must first close the gap with rivals such as OpenAI and Google before translating breakthroughs into consumer-facing tools.

Budget battles and strategic pivots

The dispute has stretched into budget planning. Bosworth’s Reality Labs, responsible for virtual and augmented reality projects, is facing cuts of up to 30 percent, following cumulative losses surpassing $60 billion since 2020. The funds are expected to be redirected toward AI initiatives.

Meanwhile, the Avocado model, once slated for release in late 2025, has been delayed until early 2026 due to significant testing and performance setbacks. The delay has also revived internal debate over whether Meta should continue its open-source tradition or move toward a proprietary AI model an abrupt shift after the lukewarm reception of Llama 4 earlier this year.

Recent months have seen prominent departures, including pioneering AI scientist Yann LeCun, who left in November after 12 years with the company. About 600 employees from Meta’s AI research and infrastructure arms were also laid off in October, though TBD Lab’s roughly 100 researchers remained unaffected.

Retention and recruitment wars

After initial instability, TBD Lab regained momentum following its first vesting milestone on November 15, with only two members departing. Meta’s competitive recruitment push earlier this year reportedly including personal gestures from Zuckerberg himself toward OpenAI employees underscored the intensity of Silicon Valley’s AI talent war.

In an official statement, a company spokesperson denied reports of internal conflict, affirming that leadership remains aligned in developing superintelligence while improving key business areas. Meta reiterated that its continued investment in AI fuels innovation across advertising systems and platform recommendations central to its global ecosystem.