Meta Lays Off 600 Employees in AI Unit as Restructuring Deepens
Table of Contents
Meta Platforms has confirmed plans to eliminate roughly 600 positions across its artificial intelligence division. The company announced the cuts on October 22, 2025, affecting workers in its Superintelligence Labs. Chief AI Officer Alexandr Wang communicated the decision through an internal memo. This marks another chapter in Meta’s ongoing efficiency drive under CEO Mark Zuckerberg’s leadership.
Understanding the Meta AI Unit Restructuring
The job cuts target specific teams within Meta’s AI unit, including the Fundamental AI Research group and infrastructure divisions. Meta’s AI unit had expanded rapidly over the past three years. The company hired aggressively to compete with rivals like OpenAI and Google. However, leadership grew concerned that existing efforts were not delivering expected breakthroughs.
Meta lays off these employees as part of a strategic reorganization. The company wants to move faster in the competitive AI race. Alexandr Wang stated that reducing team size would require fewer conversations to make decisions. Each remaining person would carry more responsibility and have greater impact.
Which Teams Face Job Cuts
The layoffs affect workers across multiple divisions within the AI unit. The Fundamental AI Research team, known as FAIR, will see significant reductions. Product-related AI teams and infrastructure units will also lose staff members. The cuts spare Meta’s newly formed TBD Lab unit entirely.
TBD Lab focuses on developing next-generation AI models. This division remains central to Meta’s superintelligence ambitions. The company continues recruiting and hiring for this specific team. Meta has offered lucrative compensation packages to attract top researchers from competitors.
Timeline of the Announcement
Employees learned about the layoffs on October 22, 2025. U.S. workers received notifications by 7 a.m. Pacific time. The company sent emails to both work and personal accounts. Affected staff members have until November 21 as their official end date.
During this period, employees cannot access company systems. They remain on nonworking notice but can apply for other internal positions. Meta encourages impacted workers to seek alternative roles within the organization. The company expects most will find new positions internally.
Severance and Support Packages
Meta lays off workers with relatively generous severance terms. Affected employees receive 16 weeks of base pay. The company adds two additional weeks for each full year of service. This package exceeds the nonworking notice period already provided.
Health insurance coverage extends for six months. This includes coverage for family members on employee plans. The company also provides access to external career services. Immigration support is available for workers on visas. Meta has assembled a dedicated recruitment team to help affected staff find internal positions.
The AI Unit Reorganization Strategy
Wang emphasized that cuts do not signal decreased investment. The company will continue hiring industry-leading AI talent. The goal is enabling Meta Superintelligence Labs to move faster. Leadership believes smaller teams can make decisions more efficiently.
CEO Zuckerberg grew frustrated with the pace of AI progress. Existing initiatives were not producing needed breakthroughs or performance improvements. The company decided to consolidate efforts under a unified vision. This led to the creation of Meta Superintelligence Labs earlier in 2025.
Meta’s Investment in AI Infrastructure
Despite workforce reductions, Meta continues massive AI spending. The company projects capital expenditures between $66 billion and $72 billion for 2025. This represents nearly 70 percent growth from 2024. Another significant increase is expected in 2026.
Meta announced a $27 billion agreement with Blue Owl Capital on October 21. This deal finances the Hyperion data center in rural Louisiana. Zuckerberg stated the facility would be large enough to cover significant portions of Manhattan. The company plans multiple multi-gigawatt clusters across the United States. For additional insights, explore our complete guide about Amazon Introduces AI Smart Glasses to Assist Delivery Drivers.
The Alexandr Wang Appointment
Alexandr Wang joined Meta in June 2025 as Chief AI Officer. He co-founded Scale AI, a data labeling company. Meta invested $14.3 billion in Scale AI as part of the arrangement. Wang brought several Scale AI employees with him to Meta.
The 28-year-old entrepreneur leads Meta Superintelligence Labs. His appointment signaled Zuckerberg’s determination to revitalize AI efforts. Wang previously testified before Congress about AI challenges. He has met with world leaders to discuss global AI cooperation.
Competition Drives AI Unit Changes
Meta faces intense competition in artificial intelligence development. OpenAI, Google, and Microsoft have advanced rapidly in consumer AI products. Meta’s Llama 4 models launched in April 2025 received mixed reception. Leadership wanted stronger performance against rival systems.
The company has recruited over 50 researchers from competing labs. Compensation packages stretch into hundreds of millions of dollars. However, OpenAI CEO Sam Altman claimed none of their best people accepted Meta’s offers. This hiring spree preceded the current AI unit reductions.
Impact on the FAIR Research Team
The Fundamental AI Research unit bears significant cuts. FAIR had been Meta’s legacy research division for years. Former FAIR leader Joelle Pineau departed earlier in 2025. The division’s prominence declined as Meta shifted focus to product-driven AI work.
Wang stated Meta would integrate FAIR’s research ideas into TBD Lab. The company wants to merge academic concepts with large-scale model training. This represents a shift from pure research toward commercial applications. Many affected FAIR employees can apply for positions in other divisions.
Year of Efficiency Continues
These cuts extend Zuckerberg’s efficiency initiative from 2023. That year, Meta eliminated over 20,000 positions. The CEO declared 2023 the “year of efficiency” as economic conditions worsened. He made these changes permanent rather than temporary.
Meta’s workforce dropped from 86,482 employees in 2022 to 67,317 in 2023. By June 2025, headcount recovered to 75,945 workers. The AI unit layoffs represent roughly 0.8 percent of total staff. Efficiency measures have boosted revenue per employee significantly.
Llama Models Performance
Meta released Llama 4 models in April 2025. The family includes Llama 4 Scout and Llama 4 Maverick. A third model, Llama 4 Behemoth, remains in development. These systems use mixture-of-experts architecture for efficient processing.
Llama 4 Maverick performs comparably to OpenAI’s GPT-4o and DeepSeek-V3. Llama 4 Scout outperforms Google’s Gemma 3 and Mistral models. However, the lukewarm reception disappointed leadership. Zuckerberg wanted stronger market impact from these releases.
Open Source Strategy
Meta continues releasing AI models as open-source products. This differs from competitors who keep models proprietary. The company claims over one billion people use its AI products monthly. However, Meta trails behind OpenAI and Google in consumer adoption.
The Llama license restricts commercial use by entities with over 700 million users. Critics debate whether this truly qualifies as open-source. Developers have downloaded Llama models over 650 million times by late 2024. The strategy aims to build a broad ecosystem around Meta’s technology.
Employee Reactions and Concerns
Current and former employees have shared mixed reactions. Some expressed understanding about strategic needs. Others voiced disappointment about losing colleagues. The tech industry has seen over 22,000 layoffs across companies in 2025.
Meta’s cuts differ from some competitors in transparency and support. The company provides clear timelines and generous severance terms. Affected workers can access internal job postings immediately. Many companies have used AI as justification for workforce reductions throughout 2025.
Social Media Discussion
The announcement generated significant discussion on social media platforms. Reddit users analyzed the implications for AI research. Some noted that TBD Lab remained untouched by cuts. Others questioned whether Meta could achieve superintelligence with fewer researchers.
LinkedIn and Twitter saw posts from affected employees. Many highlighted their contributions to Meta’s AI efforts. The community offered support and job leads. Industry observers compared Meta’s approach to other tech company layoffs in 2025.
Data Center Expansion Plans
Meta is building multiple gigawatt-scale data centers. The Prometheus facility in Ohio launches in 2026. Hyperion in Louisiana could reach five gigawatts over several years. A new facility in El Paso, Texas was announced in October 2025.
These data centers will power AI model training. Meta plans to deploy over one million GPUs. The infrastructure supports efforts to develop superintelligence. Zuckerberg stated the company would spend hundreds of billions by 2028. This spending continues despite workforce reductions.
TBD Lab Remains Priority
The TBD Lab division escapes all job cuts. This unit develops Meta’s most advanced AI models. Leadership views it as essential to achieving superintelligence. The division receives continued investment and hiring support.
TBD Lab works on next-generation large language models. These systems aim to surpass human cognitive abilities. Wang leads this effort alongside his broader Superintelligence Labs responsibilities. The division includes many of Meta’s highest-paid AI researchers.
Industry Context and Trends
Tech companies increasingly prioritize AI efficiency over headcount growth. Accenture cut 11,000 positions while training remaining staff in AI. Elon Musk’s xAI laid off 500 data annotation workers. Salesforce eliminated 4,000 customer support roles citing AI capabilities.
Critics argue companies use AI as convenient justification for layoffs. Oxford researcher Fabian Stephany calls it “scapegoating” for difficult decisions. Only one percent of firms actually reported AI as the layoff reason. The pattern suggests market correction after pandemic-era overhiring.
Meta’s Revenue and Financial Performance
Meta reported strong financial performance despite efficiency measures. Revenue per employee reached approximately $2.2 million in 2025. Free cash flow hit record levels under Zuckerberg’s efficiency focus. The company generated $52.1 billion in free cash flow in 2024.
Stock performance has rewarded efficiency initiatives. Investors approved of workforce reductions and cost controls. Meta’s advertising business remains healthy with 10 percent year-over-year price growth. This revenue funds ambitious AI infrastructure spending.
Future Workforce Projections
Analysts expect Meta’s total headcount to stabilize around 72,000 to 75,000 employees. The company continues hiring for specialized AI roles. Competition for talent remains intense across the industry. Compensation packages and equity structures influence retention rates.
Meta faces ongoing pressure to balance efficiency with innovation. The company must retain top talent while controlling costs. Leadership believes leaner teams move faster and deliver better results. This philosophy drives current restructuring efforts.
Looking Ahead: Meta Lays Off
Meta lays off employees while maintaining focus on long-term AI goals. The company believes restructuring will accelerate progress toward superintelligence. Substantial infrastructure investments continue despite workforce reductions. Leadership remains confident in their path to developing advanced AI systems.
The AI unit reorganization reflects broader industry trends. Companies prioritize efficiency and specialized talent over large teams. Meta’s approach balances cost control with ambitious technological goals. The coming months will reveal whether smaller teams can deliver breakthrough innovations faster than larger research groups.
