Meta Is Cutting Thousands of Jobs to Fund a $135 Billion AI Bet. Here's What's Really Happening.

On March 25, 2026, Meta began notifying hundreds of employees across at least five divisions that their positions were being eliminated [1][6]. The cuts hit Reality Labs, Facebook, global operations, recruiting, and sales — the latest in a series of reductions that have already claimed roughly 1,500 jobs from Meta's virtual reality unit earlier this year [6][7]. But these layoffs are not the end. Reuters reported that Meta's senior executives have been directed to plan workforce reductions of approximately 20%, which could eliminate roughly 15,000 positions from a workforce of about 79,000 [1][2].

The backdrop: Meta has committed to spending between $115 billion and $135 billion on AI-related capital expenditure in 2026 — nearly double the $72.2 billion it spent in 2025 [3][4]. The company is simultaneously slashing headcount and doubling down on artificial intelligence infrastructure at a scale that dwarfs most national budgets.

The Scale of the Cuts

The March 25 layoffs affect fewer than 1,000 employees across five divisions [1][6]. But they follow a January reduction that cut roughly 10% of Reality Labs — about 1,500 positions [7]. Combined with the broader 20% reduction plan, Meta's 2026 workforce restructuring could rival or exceed the 21,000 jobs eliminated during the 2022-2023 "year of efficiency," when CEO Mark Zuckerberg acknowledged the company had overhired during the pandemic [1][2].

Zuckerberg confirmed in an internal letter that layoffs would proceed in multiple waves through March, April, and May [8]. Fox Business reported the total at roughly 700 for the March 25 round alone [9].

The affected divisions tell a story about what Meta considers expendable. Reality Labs, which has accumulated approximately $83.6 billion in operating losses since 2020 and lost $19.2 billion in 2025 alone, is bearing significant cuts [10][11]. The division's Quest headsets and Ray-Ban smart glasses have shown modest commercial traction, but its core VR/AR hardware development appears to be slowing as resources shift toward AI software [7].

Where the Money Is Going

Meta's $115-$135 billion capex guidance for 2026 is staggering in context [3][4]. The spending is directed primarily at data centers, servers, and networking equipment needed to train and deploy increasingly large AI models [4]. Meta has also forecast total expenses of $162 billion to $169 billion for 2026, including rising compensation costs as the company recruits top AI researchers with seven-figure packages [1].

The AI investments span several categories:

Foundation models: Meta launched Llama 4 in spring 2025, with multimodal variants including Llama 4 Scout and Llama 4 Maverick [12]. The company is developing next-generation models internally code-named "Mango" (image and video) and "Avocado" (text-based LLM), targeted for release in the first half of 2026 [13][14].

Advertising: Meta's Advantage+ automated ad system already uses AI to create, target, and optimize ads with minimal human input, and is driving significant revenue growth [15]. Llama models are deployed internally for ad creative generation and ad quality evaluation [15].

Content moderation: Meta developed Llama Guard 3 for automated content moderation across eight languages, though the company also phased out its third-party fact-checking program in January in favor of a user-driven "Community Notes" model [16].

Infrastructure: The majority of the $115-$135 billion capex goes to physical infrastructure — data centers, GPUs, and networking gear [4]. Meta outlined a $600 billion U.S. infrastructure plan through 2028 [17].

This is not a case of layoffs directly funding AI. Meta is both cutting costs and increasing total spending. Bank of America analysts estimated the restructuring could yield approximately $7-8 billion in annualized savings [5] — a fraction of the $45 billion in projected GAAP expense growth for 2026. The layoffs reduce the growth rate of costs; they do not pay for the AI buildout.

Source: Meta Platforms Earnings Reports / DataCenter Dynamics

Data as of Mar 26, 2026CSV

What Laid-Off Workers Are Getting

Meta's severance package includes 16 weeks of base pay plus two additional weeks for every year of employment, remaining PTO paid out, stock vested through the next pay period, six months of healthcare coverage, and three months of career support [8].

Some affected employees are being offered transfers to other roles or locations within the company [8]. Advertising teams, which continued to grow after previous layoffs, have been directed to prioritize internal transfers over external hires [8]. However, roughly 600 employees are expected to leave Meta's AI division itself — a sign that even within the company's top priority, not all roles are safe [8].

The tech job market these workers face has deteriorated. U.S. computer systems design employment (NAICS 5412) has fallen from a peak of 415,900 jobs in mid-2023 to 344,100 in February 2026 — a decline of over 70,000 positions, or 17%, in less than three years [18]. While AI-related positions are surging with a 92% increase in hiring and a 56% wage premium [19], these roles require specialized skills that many displaced workers in operations, recruiting, and sales do not have.

Source: Bureau of Labor Statistics (CES5051200001)

Data as of Mar 26, 2026CSV

How Competitors Are Playing It

Meta is not an outlier. The pattern of cutting non-AI staff while hiring AI specialists is industry-wide in 2026.

Amazon has cut roughly 16,000 corporate employees in 2026, framing the reductions as flattening management layers, even while reporting record revenue of $716.9 billion in 2025 [19].

Google/Alphabet cut hundreds from its Platform and Devices unit after extending buyout offers in January 2026. The company is restructuring to prioritize AI investments across search, cloud, and its Gemini model family [20].

Microsoft has not announced comparable mass layoffs in 2026 but has been steadily shifting hiring toward AI roles, with its $80+ billion annual AI infrastructure commitment rivaling Meta's [3].

Block (Square, Cash App) went further than any of them, slashing 4,000 roles — nearly 40% of its workforce — in early 2026 [19].

Across the five largest U.S. cloud and AI infrastructure providers — Microsoft, Alphabet, Amazon, Meta, and Oracle — collective capital expenditure commitments for 2026 total between $660 billion and $690 billion, nearly double 2025 levels [3]. The industry-wide bet is that AI infrastructure spending now will generate returns that justify displacing existing workers.

So far in 2026, 171 tech companies have conducted layoffs affecting 55,911 workers — roughly 736 per day. Of those, 20.4% were explicitly linked to AI and automation, up from fewer than 8% in 2025 [19].

AI Transformation or Rebranded Cost-Cutting?

The question of whether "AI transformation" is genuine or a convenient label for standard cost reduction is central to understanding these layoffs.

Evidence supporting genuine transformation: Meta is spending $115-$135 billion on capex — money that would be difficult to justify without concrete AI product plans [3][4]. The Llama model family is actively shipping and generating revenue through advertising applications [12][15]. The next-generation models (Mango and Avocado) have public development timelines [13][14]. Meta AI, the company's consumer chatbot, is deployed across its 3.3 billion-user ecosystem [15].

Evidence that raises questions: Reality Labs, despite its $83.6 billion in losses, was Meta's most visible "future bet" just two years ago — and is now being gutted [10][11]. The company briefly announced it would shut down Horizon Worlds, its flagship VR social platform, before reversing the decision days later [11]. Approximately 600 AI team members are themselves being laid off, which complicates the narrative that cuts serve the AI mission [8]. And the 20% overall reduction target — roughly 15,000 people — goes well beyond what would be needed to staff up AI teams [2].

Wall Street has largely accepted the AI-transformation framing. Meta stock climbed nearly 3% on the initial Reuters report of planned layoffs [5]. The company holds a consensus "Strong Buy" rating, with 46 of 56 analysts rating it the top category [5]. Jefferies analysts wrote that "if Meta is willing to reduce headcount at this scale while ramping AI investment, we think it signals a broader shift: AI is increasingly driving productivity" [5].

The Competitive Stakes

Meta has been racing to close a perceived gap with rivals. OpenAI, Anthropic, and Google all had more advanced commercial AI products when Meta began its AI pivot in earnest. The Llama model family — particularly its open-source strategy — has been Meta's primary differentiator, allowing external developers to build on Meta's models and creating an ecosystem that competes with closed-source alternatives [12].

Analysts who support the restructuring argue that maintaining the current workforce while competitors invest aggressively in AI would put Meta at a structural disadvantage. If AI-powered advertising, content recommendations, and user engagement tools become the primary drivers of social media revenue — a transition already underway — then a company that falls behind in AI capabilities risks losing its core business [5][15].

Critics counter that the scale and speed of cuts create institutional knowledge loss that cannot easily be replaced, and that the affected employees in operations, recruiting, and sales perform functions that AI tools are not yet capable of replicating at quality [21].

What Gets Left Behind

The products and services that laid-off teams maintained do not vanish overnight, but they do lose momentum. Reality Labs' VR hardware roadmap will likely slow as engineering headcount drops [7]. The brief Horizon Worlds shutdown announcement — later reversed — suggests internal uncertainty about which products justify continued investment [11].

Facebook's app teams, which serve roughly 3 billion monthly active users, are also affected by these cuts [6]. While Meta's core social media products generate massive revenue — the family of apps division has earned more than $352 billion since 2021 [7] — reduced staffing in areas like global operations and sales could affect advertiser relationships and platform maintenance.

The clearest signal is in recruiting: cutting the team that hires people tells the market that Meta does not plan to backfill most of the eliminated positions. The company is getting smaller in headcount while getting larger in capital spending — a configuration that only works if AI tools genuinely replace the productivity of displaced workers.

The Broader Pattern

Meta's layoffs are part of a structural shift across the technology industry. The 245,000 tech jobs cut globally in 2025 have been followed by an accelerating pace in 2026 [19]. The companies executing these cuts are simultaneously posting record revenues and investing unprecedented sums in AI infrastructure.

For the roughly 15,000 Meta employees who may lose their jobs this year, the distinction between "AI transformation" and "cost optimization" is largely academic. The outcome is the same: fewer roles in traditional tech functions, a shrinking market for those skills, and a growing premium on AI expertise that most displaced workers do not possess.

Whether Meta's $135 billion bet produces returns commensurate with its human cost will take years to determine. In the meantime, the company is making a wager that the technology industry's largest employers have collectively endorsed: that artificial intelligence will generate more value with fewer people than the organizational structures it replaces.