The news: CEO Mark Zuckerberg has reorganized Meta Superintelligence Labs (MSL) into four units focused on research, superintelligence, products, and infrastructure, per The New York Times. 

• Meta further splitting its AI division, which it spun off in June, underscores both ambition and internal turmoil as it races rivals like OpenAI and Google. 
• Its restructuring shows that even with billions spent and a deep bench of AI talent, execution is messy. For example, there’s an internal debate over whether Meta should abandon open source and build closed models, per Business Insider.

Competitors will read this as validation that Meta is still searching for new directions in AI, which buys them time to solidify their own leadership positions.

Meta’s deep bench is proving difficult to manage: Leadership turnover is inevitable in realignment, as evidenced by recent exits to rivals like OpenAI, Cohere, and Figma. 

MSL, now estimated to have thousands of employees, could face cuts or role reassignments.  

Here are some other aspects of the restructuring process:

• Zuckerberg is open to using third-party AI models—a break from Meta’s reliance on its own tech.
• The company is pursuing existing AI solutions through reverse acqui-hires: It invested $14.3 billion in Scale AI, naming its CEO, Alexandr Wang, as chief AI officer. The investment resulted in Google cutting ties with the startup.
• Meta has reportedly paused AI hiring, per Yahoo. Indicating investor pressure on the unit's surging costs.
• It reportedly scrapped its previous frontier model, Llama 4 Behemoth, after early tests proved underwhelming, per DigiTimes.

Zooming in: Meta may be suffering from too much AI talent spread across competing agendas, with limited tangible output. 

Continuous changes and short-term uncertainty could result in a talent exodus, especially as AI rivals frame themselves as disciplined and results-driven while showcasing definitive product innovations.

Our take: Meta’s public growing pains show it won’t sit out the AI race, even if upheaval is the cost. Its future direction will have wider implications—if Meta leans into closed AI models, the shift could reshape how outside developers and partners interact with its platforms. 

For advertisers, the signal is clear: Expect fresh AI features in Meta’s ad products, but brace for volatility as Meta struggles to align its people, platforms, and technology.