Meta is preparing a fresh, large-scale reduction in force that could cut as much as 20% of its workforce as the social-media giant reallocates resources toward an expensive bet on artificial intelligence. Company leaders have reportedly asked senior managers to draw up plans to shrink teams, a sign that Meta is weighing broad organisational change to offset soaring AI infrastructure costs and to capture productivity gains promised by advanced models.
Timing and final scale remain unsettled; Meta spokesman Andy Stone dismissed media coverage as speculative, but the reported potential pace of cuts would mark the largest retrenchment since the company’s “year of efficiency” reorganisation in late 2022 and early 2023. Meta employed roughly 79,000 people at the end of last year, meaning a 20% reduction would translate into about 16,000 roles being eliminated.
The move flows from two competing pressures. On one side are enormous capital commitments: Meta has signalled plans to invest heavily in data centres and compute, and has been paying top dollar to recruit leading AI researchers into a so-called “superintelligence” team. On the other side is a strategic conviction inside the company that next‑generation models will materially increase productivity—letting fewer people do more work—and therefore justify trimming headcount.
Meta’s AI push has not been smooth. The company has poured large compensation packages into hiring elite researchers, acquired niche platforms such as Moltbook to bolster agent-style experiences, and disclosed multi‑year data‑centre investments measured in the hundreds of billions of dollars. Its public technical record this year includes controversies around Llama 4 and delays to its next flagship model, Avocado, which has been pushed back after early benchmarks lagged those of Google, OpenAI and Anthropic.
The potential cuts mirror a wider shift across US technology firms. Amazon recently announced reductions of about 16,000 roles, Block pared staff dramatically, and industry leaders publicly argue that rising AI capabilities will let companies streamline operations. That shift is altering hiring calculus for universities, startups and established firms alike and is reshaping expectations for how much human labour large technology projects will require.
For Meta, the trade-offs are stark. A large layoff would shave near-term costs and accelerate the company’s pivot toward model-centric products, but it risks eroding institutional knowledge, damaging morale and creating public-relations and regulatory headaches. It will also intensify competition for displaced engineers and researchers, potentially fuelling rival AI groups and startups.
What to watch next are three variables: whether Meta moves from planning to execution, how deep the cuts go in revenue-generating versus research teams, and whether the company’s AI products and cost-savings materialise fast enough to justify reduced headcount. The decision will influence not only Meta’s trajectory but also broader expectations about labour, productivity and the pace of AI-driven organisational change in the technology sector.