Meta Cuts 8,000 Jobs, Cancels 6,000 Hires While Budgeting Up to $135 Billion for AI

Meta eliminated roughly 8,000 roles in May — about 10% of its workforce — and scrapped plans to fill 6,000 open positions, while simultaneously earmarking $115 billion to $135 billion in capital expenditures for 2026, nearly double the $72 billion spent in 2025 [1][3][4]. The company also cut approximately 600 jobs from its newly formed AI Superintelligence Labs on October 22, just months after spending $14.3 billion to acquire a stake in Scale AI and recruit its founder, Alexandr Wang, as chief AI officer [1].

The cuts

The May reductions landed hardest on non-AI product and support teams, including parts of integrity and cybersecurity staff [1]. An earlier round in March removed about 700 people from Reality Labs, social media and recruiting teams [3][5]. Meta also reportedly set a late-May exit date for roughly 200 Bay Area workers [3]. The canceled hires equal three-quarters of the actual dismissals, shrinking both the payroll Meta carries today and the one it once meant to build [3].

About 7,000 existing employees were reassigned to AI-related work, a figure the company stresses should not be added to the layoff total [4]. In France, a local voluntary-departure plan could affect up to 100 of 300 staff, mainly in sales [4].

The AI spending

Meta has told investors it is preparing a data-center build-out that could approach $600 billion through 2028 [1][7]. The 2026 capex range, disclosed during the year and later revised, covers servers, compute capacity, networking gear and power for data centers [4]. CEO Mark Zuckerberg has said he wants smaller teams producing more with AI tools, and in January claimed projects that once needed big teams were being done by single talented people [7].

Yet the company’s own AI products have stumbled. The Llama 4 release last year drew criticism for misleading benchmark results; the largest version, Behemoth, was abandoned [7]. The Superintelligence Labs’ new model, Avocado, has also lagged expectations [7]. Zuckerberg recently told employees the 2026 spending surge has not yet delivered the payoff he promised [1].

Investor skepticism

Semiconductor stocks sold off sharply in November, with AMD and Arm each down more than 20% on the month [1]. KeyBanc Capital Markets analysts flagged mounting investor questions over Meta’s spending levels against unclear returns [1]. Unlike Microsoft, Amazon and Google, Meta generates no meaningful revenue from AI products; its cloud peers can point to AI compute already billed to customers [1]. The company’s advertising business remains strong — Q3 2026 revenue reached about $63.06 billion, up 26% year over year — but that growth has not prevented the cuts [4].

Internal tensions

A Reuters investigation reported in late summer that Meta studied scenarios cutting certain teams by up to 60% as part of a push toward an “AI native” organization [4]. That hypothesis was not a final plan and did not apply company-wide. Zuckerberg reportedly abandoned a second global wave of layoffs considered for November after earlier saying he did not expect further company-wide cuts in 2026 [4].

In July, 26 U.S. employees filed a legal action alleging Meta used internal AI tools to evaluate performance and help identify workers for layoff, claiming the criteria penalized people who had taken protected leave [4]. Meta disputes the allegations and says staffing decisions are made by people, not AI [4]. The claims are not judicially established.

What remains unclear

How much of the $115–135 billion capex has actually been committed versus budgeted is not broken out in the available reporting. The extent to which reassigned staff are filling AI roles versus backfilling other gaps is also unspecified. Meta has not publicly commented on the French restructuring reports [4].