Industry

Meta to Lay Off 10% of Workforce for AI Focus

Meta plans to reduce its workforce by 10 percent, affecting around 8,000 employees, and eliminate 6,000 open positions to prioritize artificial intelligence development. The company, which had over 78,000 staff at the end of 2025, faces high costs from AI investments exceeding $70 billion. CEO Mark Zuckerberg aims to build AI products like personal superintelligence amid competition from OpenAI, Google, and Anthropic.

Neura News

Neura News

Neura Market Editorial

April 23, 20263 min read

Originally reported by nytimes.com

Meta to Lay Off 10% of Workforce for AI Focus

Meta to Lay Off 10% of Workforce for AI Focus

Meta will reduce its workforce by 10 percent, impacting approximately 8,000 employees. The company also intends to eliminate 6,000 open positions. These moves come as Meta shifts resources toward artificial intelligence efforts, according to an internal memo sent to staff on Thursday.

Meta owns Facebook, Instagram, and WhatsApp. The social media giant had more than 78,000 employees at the close of 2025. CEO Mark Zuckerberg has predicted that AI systems will handle a large portion of tech industry tasks. He points to tools like coding assistants that aid software engineers.

Details of the Layoffs

Janelle Gale, Meta's chief people officer, explained the decision in the memo. "We're doing this as part of our continued effort to run the company more efficiently and to allow us to offset the other investments we're making," she wrote. She added, "This is not an easy trade-off and it will mean letting go of people who have made meaningful contributions to Meta during their time here."

A Meta spokesman verified the layoffs but offered no additional details.

Reorganization Around AI

Zuckerberg is restructuring the company to center on AI products. He competes in an intense battle with companies such as OpenAI, Google, and Anthropic to dominate the field. Meta has advanced unevenly but trails rivals in creating core AI models.

The #1 Newsletter in AI

Stay ahead of the AI curve

The most important updates, news, and content — delivered weekly.

No spam. Unsubscribe anytime.

To close the gap, Zuckerberg approved over $70 billion in AI-related spending. This covers data centers, semiconductors, and real estate needed for advanced AI work. During a January investor call, he forecasted expenditures of $115 billion to $135 billion for the current year. That marks almost double the $72 billion from the prior year, with a large share directed at AI.

Zuckerberg envisions AI-driven social media features as a form of "personal superintelligence." He wants users to integrate these into everyday routines.

In a July video on his Facebook page, he stated, "At Meta, we have the resources to build the massive infrastructure required and the ability to deliver new technology to billions of people."

Financial Rationale and Revenue Growth

Zuckerberg has supported the substantial outlays on AI. In the fourth quarter of 2025, Meta's revenue increased 24 percent compared to the year before. He credits AI enhancements that refine ad targeting and suggest videos and posts to users.

Meta, founded in 2004 by Zuckerberg and others as Facebook, has grown into a major tech player. It manages platforms used by billions worldwide. The push into AI reflects broader industry trends where companies invest heavily to stay competitive. Rivals like OpenAI develop models such as GPT series, while Google integrates AI into search and services. Anthropic focuses on safe AI systems.

This story is developing. More updates will follow.

Related on Neura Market

More from Neura News

Research

PlanFlip Attacks Target Multi-Agent LLM Planning Phase

New research from Yuhang Wang introduces PlanFlip, a framework of four prompt injection attacks targeting the planning phase of multi-agent LLM systems. The attacks exploit a single injection into the Planner agent's context to corrupt all downstream sub-tasks. Testing on nine frontier LLMs across 3,479 episodes revealed that stronger models like GPT-5 are more vulnerable, while reasoning-augmented models like DeepSeek-R1 show full resistance.

Jul 21·3 min read