Meta has cut 600 roles within its artificial intelligence division as part of a major restructuring of its Superintelligence Labs. The move follows years of extensive hiring under CEO Mark Zuckerberg and aims to create a more “organized and rapid” AI operation.
The layoffs come after Meta’s latest push to position itself as a leader in AI research and infrastructure. According to The New York Times, the cuts were concentrated in the company’s FAIR research unit and teams managing AI infrastructure and product integration.
Chief AI Officer Alexander Wang said the goal is to remove “organizational bloat” and make decision-making faster.
“Fewer conversations will be required to make a decision, and each person will have more scope and impact,” Wang wrote in a memo to staff, obtained by The New York Times.
The newly formed TBD Lab, a small, elite team developing Meta’s next-generation foundation models, was not affected. Reuters reported that affected workers were encouraged to apply for other positions across Meta’s AI ecosystem.
Meta’s leadership insists the layoffs are not a retreat from AI; instead, they represent a shift toward speed, efficiency and competitive focus. The company has spent the past year revamping its AI efforts after internal missteps and challenges surrounding the launch of its Llama 4 model.
In June, Meta reorganized its AI units under the Superintelligence Labs umbrella and appointed Wang, former CEO of Scale AI to lead the overhaul.
Scale AI is a San Francisco based startup specializing in data labeling and infrastructure for training large models. Meta acquired a 49% stake in the company and Zuckerberg also invested $14.3 billion, securing both its expertise and top talent.
Since then, Meta has been on a costly recruitment drive, hiring over 50 engineers and researchers from OpenAI, Google and Apple, according to The Wall Street Journal. Many of these hires joined the TBD Lab, now leading Meta’s most advanced model development.
Alongside these cuts, Meta struck a $27 billion financing deal with Blue Owl Capital to fund its largest data center project. Reuters noted that this agreement allows Meta to pursue massive AI expansion while shifting much of the cost and risk to external investors.
Despite internal turbulence, Meta’s strategy shows Zuckerberg’s growing focus on “superintelligence,” which is AI that could one day surpass human cognitive abilities. The company views this as its best chance to redefine its position in the global AI race.
For years, Meta fell behind rivals like OpenAI and Google, whose models dominated both consumer and enterprise use. The restructuring suggests a recognition that success in AI depends not only on research but also on execution and discipline.
Following the restructuring, Meta faces a clear test: turning its smaller division into a faster, more focused driver of innovation. The coming months will reveal whether streamlining talent can accelerate the company’s push toward superintelligence.
