Meta reportedly plans 20%+ layoffs as AI costs surge

Meta is reportedly preparing to cut 20% or more of its workforce, according to a Reuters report citing three people familiar with the matter. With nearly 79,000 employees as of late December, a cut of that scale would wipe out roughly 15,000 or more roles.

The timing and exact headcount targets have not been finalized yet. However, executives have already told senior leaders to start sketching out plans to trim their teams. When reached for comment, a Meta spokesperson called the story speculative reporting about theoretical approaches.

Paying for servers by cutting staff

If confirmed, this would mark Meta's deepest round of job cuts since its late 2022 and early 2023 "year of efficiency," when it eliminated over 21,000 positions across two separate waves.

The primary driver this time is the staggering bill for generative AI. Meta has outlined plans to pour $600B into data center buildouts through 2028. It has also handed out four-year compensation packages worth hundreds of millions of dollars to recruit top researchers for a new superintelligence team.

On top of that, Meta just snapped up Moltbook, a social network built for AI agents, and reportedly spent at least $2B to acquire Chinese AI startup Manus. Zuckerberg hinted at the workforce shift in January, noting that projects once requiring massive teams can now be tackled by a single talented person using AI.

Efficiency gains or AI-washing?

Meta is far from alone in using this playbook. Tech giants across the board are restructuring around AI promises.

In January, Amazon confirmed it was laying off around 16,000 workers, or nearly 10% of its workforce. Last month, fintech player Block wiped out nearly half of its staff, with CEO Jack Dorsey directly crediting AI productivity gains.

Still, industry skepticism is running high. Critics and even tech figures like OpenAI chief Sam Altman have suggested many of these moves are "AI-washing" using the tech as air cover to trim pandemic-era bloat.

Meta's push also arrives amid rocky AI progress. The company ran into benchmark controversies with Llama 4 last year, ultimately scrapping its summer release of the massive Behemoth model. Its new Avocado model, built by the superintelligence unit, has also reportedly lagged internal expectations.

Why it matters

We think this planned bloodbath shows the real toll of the generative AI infrastructure race. Meta cannot spend $600B on chips and concrete without pulling cash from somewhere else, and payroll is the quickest lever to pull. It signals a shift where human headcount is directly converted into compute budget.

For tech workers, the takeaway is stark. Whether AI tools are actually doubling developer output or executives are just using AI as convenient cover to slash headcounts, the result is identical: corporate giants are betting they can run core revenue engines with vastly leaner benches.

The strategy looks bold, but it carries immense risk. Meta is spending fortunes courting elite talent and buying up startups while trimming the rank-and-file who run its lucrative ads engine. But with Llama 4 Behemoth abandoned and Avocado reportedly falling short, pouring everything into unproven frontier models while gutting operational teams could backfire if the next generation of AI still fails to deliver.