Meta Cuts 600 Staff in Core AI Unit
Meta Cuts 600 Staff in Core AI Unit
※ Meta Platforms, parent of Facebook, has laid off about 600 employees from a core artificial intelligence (AI) organization.The move is part of an internal restructuring of the Superintelligence division and is seen as a strategic step to trim duplication while preserving AI research capability.
Meta stressed that this is not a downsizing of AI, but a rebuild to simplify decision-making and raise efficiency. 😅
🏢 Why the Restructuring?
Over the past three years, Meta aggressively hired talent in the AI race.
Key researchers from OpenAI, Google DeepMind, Microsoft, and Scale AI joined Meta with multi–hundred million–dollar compensation packages, rapidly bulking up the organization.
That expansion, however, increased internal complexity. To return to a “fast experiment → fast execution” model, Meta appears to have opted for selective reductions.
The cuts centered on Superintelligence Labs, which oversees Meta’s AI projects.
Some employees were notified of departure, while others are being considered for reassignment to different teams.
👨💻 Core Talent Remains
Artificial Superintelligence (ASI) remains CEO Mark Zuckerberg’s top priority.
New Chief AI Officer Alexander Wang, cofounder of Scale AI, is a key hire tied to Meta’s $14.3 billion investment announced in June.
In a staff memo, he wrote:
“As the team gets smaller, decision-making becomes faster,
and each person’s responsibility and impact grow.”
In short, this is not mere trimming but a shift to a “lean and agile AI organization.”
⚙️ Where Meta Stands in the AI Race
Meta once drew attention with its open-source model LLaMA,
but productization lagged. As OpenAI’s ChatGPT, Google’s Gemini, and Microsoft Copilot rolled out rapidly, Meta came to be seen as a late mover in AI despite its massive platforms (Facebook, Instagram, WhatsApp).
The LLaMA 4 model, unveiled in April 2025, fell short of expectations.
Zuckerberg then ordered a reboot of AI strategy and staffing, culminating in this restructuring.
💬 Meta’s Official Line
A Meta spokesperson said the changes are a reorganization around core talent, not an AI pullback.
AI and superintelligence development remain central to the company’s long-term growth strategy, with substantial R&D investment to continue.
📊 Industry View: Risks & Opportunities for Meta
| Intensifying competition | Tech gap vs. OpenAI/Google/MS | Long-horizon vision for ASI |
| Workforce changes | Short-term uncertainty; morale concerns | Faster decisions; less bureaucracy |
| Cost control | R&D delay risk | Financial stability and potential re-rating |
As AI shifts from a “scale race” to a “speed race,” Meta’s move looks like a “focus and select” strategy.
💰 Investment View: U.S. & Korea Stock Ideas
🇺🇸 U.S. Stocks
| META (Meta Platforms) | Near-term uncertainty, but efficiency in AI and ad recovery support a longer-term rebound |
| MSFT (Microsoft) | Leads commercialization pace via OpenAI partnership |
| NVDA (NVIDIA) | Prime beneficiary of AI server/data center capex |
| GOOGL (Alphabet) | Gemini commercialization expands AI ecosystem |
🇰🇷 Korea Stocks
| Samsung Electronics (005930) | Surging demand for AI server memory & HPC semis |
| SK hynix (000660) | HBM3E ramp; accelerating NVIDIA-linked revenue |
| Hanmi Semiconductor (042700) | Ongoing demand for AI chip packaging equipment |
| NAVER (035420) | Growth potential from HyperCLOVA X–based AI services |
📈 Investment Insight
The layoffs signal cost control and efficiency gains aimed at improving profitability.
With AI competition shifting from “how big” to “how focused,” this marks a strategic inflection point.
- The ecosystem continues to consolidate around Big Tech (Meta, Google, MS, OpenAI)
- After any short-term turbulence, another AI supercomputing investment wave is likely
❓ FAQ
Q1. Will the layoffs slow Meta’s AI development?
→ If anything, removing bottlenecks should accelerate decision-making.
Q2. Is META stock worth buying now?
→ Expect near-term volatility, but AI efficiency + ad recovery can underpin a medium- to long-term rebound.
Q3. How does this affect Korean chipmakers?
→ AI server expansion by Meta, Google, and Microsoft supports HBM, DDR5, SSD demand—medium- to long-term positive for Korean memory names.