Meta Is Spending $135 Billion on AI and Cutting 15,000 Jobs. Both Numbers Tell the Same Story.

Here are two numbers that belong in the same sentence: Meta plans to spend up to $135 billion on AI infrastructure in 2026 — nearly double last year’s $72 billion — while simultaneously cutting up to 15,000 jobs, roughly 20% of its workforce.

These aren’t contradictory moves. They’re the same move.

The Playbook

Meta’s strategy has a name internally: becoming “AI-native.” The practical version is blunt. Leaked internal documents from March 2026 reveal that Meta wants 65% of its engineers writing 75% or more of their committed code using AI tools by mid-2026. CEO Mark Zuckerberg has been even more direct: “We’re starting to see projects that used to require big teams now be accomplished by a single very talented person.”

The layoffs reflect this. Data from early 2026 cuts show the roles being eliminated aren’t random — they’re specific:

• Mid-level management — coordination layers that AI workflows bypass
• Quality assurance teams — increasingly automated by AI testing tools
• Customer support staff — replaced by Meta’s internal “Metamate” AI agents
• Internal IT — automated by “DevMate,” Meta’s proprietary agentic coding tool

The company claims these AI tools now handle up to 70% of routine coding and administrative tasks. Whether that number is precise matters less than the direction: every function that can be described as “routine” is being automated.

Meta Is Not Alone

The tech sector has eliminated nearly 60,000 jobs in the first three months of 2026. Of those, at least 20% were explicitly linked to AI and automation by the companies themselves.

• Amazon: 16,000 positions — the single largest layoff of the year
• Block: 4,000 jobs (40% of workforce). CEO Jack Dorsey cited “the growing capability of AI tools to perform a wider range of tasks”
• Oracle: An estimated 20,000–30,000 employees in a sweeping reduction announced via a 6 AM email
• Meta: 1,500 already cut from Reality Labs, with up to 15,000 more planned

The companies doing the cutting are posting record revenues. That’s the part that should get your attention. This isn’t cost-cutting out of desperation. It’s restructuring out of conviction.

What’s Actually Happening

Strip away the corporate language and the pattern is clear: companies are splitting their workforces into two groups.

Group 1: People who direct AI systems — who scope problems, evaluate outputs, make judgment calls, and ship products using AI as their execution engine. These people are getting raises, promotions, and hiring priority. PwC’s 2025 analysis of nearly 1 billion job ads found they earn 56% more than peers without AI skills — and that premium doubled in a single year.

Group 2: People whose work can be described as a series of steps. If your job can be written as a checklist — review this, format that, route this to the right person — it’s being automated. Not in five years. Now.

Meta’s $135 billion isn’t being spent to replace all workers. It’s being spent to make Group 1 dramatically more productive while making Group 2 unnecessary.

The Uncomfortable Question

Which group are you in?

The answer isn’t about your job title. It’s about a specific skill: can you direct AI to produce work you couldn’t produce alone?

That means knowing how to ask the right questions — not “help me with marketing” but “create a 7-day content calendar for a boutique coffee shop targeting 25–34 professionals within 2 miles, with post times and Reel concepts.” The difference between those two prompts is the difference between getting generic advice and getting an executable plan.

Meta’s bet is that a smaller number of people who know how to direct AI will outproduce a larger number who don’t. Every hiring decision, every layoff, every dollar of that $135 billion reflects this belief.

What This Means for You

If Meta’s playbook works — and Wall Street clearly thinks it will, given the stock price reaction — every company in every industry will follow. The World Economic Forum projects AI will create 170 million new jobs by 2030 while displacing 92 million. The net is positive, but only for people on the right side of the skill divide.

The good news: this skill is learnable. It isn’t a computer science degree. It isn’t learning to code from scratch. It’s learning to think in specifics, to scope problems precisely, to evaluate AI output critically, and to iterate until the result is production-ready. That takes months, not years.

The window to learn it is now. Meta’s 2026 is every company’s 2027.