AI Is Eating Its Own: Meta's Infrastructure Bet Means 16,000 Jobs Lost
Meta is planning its largest layoffs since the brutal "year of efficiency" cuts in 2022-2023. This time: over 16,000 employees, more than 20% of the workforce.
The reason? To pay for $600 billion in AI infrastructure by 2028.
Here's the part that should make everyone uncomfortable: Meta is investing billions in AI to make workers more efficient, then firing the workers.
The Brutal Math
Mark Zuckerberg said it plainly in January: "projects that used to require big teams can now be accomplished by a single very talented person."
Translation: we need fewer humans.
Meta isn't alone. Amazon cut 16,000 jobs (10% of corporate staff). Block eliminated nearly half its workforce. Across Big Tech, the pattern is identical. Build AI tools that make teams more productive. Realize you need smaller teams. Cut.
But Meta's situation is particularly dark. These aren't just any workers. These are the people who build the AI systems that are replacing them.
AI is eating its own.
The $600B Question
Meta's AI ambitions are staggering. $600 billion in data centers and infrastructure by 2028. That's not a typo. That's more than the GDP of most countries.
Where does that money come from? Not thin air. It comes from cost savings. And in a company where 70% of expenses are people, cost savings means layoffs.
The irony is suffocating. Meta invests billions to build systems that eliminate the need for the very engineers building those systems. The snake eating its tail.
Recent AI setbacks make it worse. Llama 4 faced criticism. Project Behemoth got abandoned. Avocado is lagging competitors. Meta is pouring money into AI infrastructure while its AI products struggle to justify the investment.
So they cut people to fund the machines. Even when the machines aren't quite working yet.
What This Means For Everyone Else
If you think this is just a Big Tech problem, you're not paying attention.
The playbook is being written right now:
- Invest in AI tools to boost productivity
- Measure efficiency gains
- Realize you can do the same work with fewer people
- Cut headcount to fund more AI investment
- Repeat
Zuckerberg's "small teams" vision isn't aspirational. It's a roadmap. And every company with a spreadsheet and a CFO is watching.
The productivity paradox reaches its darkest conclusion here. We were told AI would augment human workers, make us more capable, free us from drudgery. Instead, it's becoming justification for eliminating positions entirely.
Not "you'll work alongside AI." Just "we don't need you anymore."
The Questions No One Wants to Ask
What happens when the most advanced AI companies decide they need 20% fewer humans?
What happens when "AI fluency" becomes the reason companies demand more output from smaller teams?
What happens when efficiency gains don't lead to better work-life balance, but to unemployment?
Meta's $600B bet isn't just on AI infrastructure. It's a bet that machines can replace the majority of knowledge workers. And they're willing to prove it by starting with their own employees.
The people building the future are being told they're not part of it.
The Uncomfortable Truth
This isn't about Meta being evil or short-sighted. This is about economics. If AI can genuinely do the work of multiple people, companies will choose AI. Every time.
The transition everyone's been talking about in abstract terms? It's here. It's not gradual. It's not gentle. And it's starting with the people who understood it best.
If the engineers who build AI can't protect their own jobs from AI, what chance does everyone else have?
That's the question keeping people up at night. And Meta just made it a lot harder to answer with optimism.
What do you think? Are we building our own replacement? Reply and let us know how you're thinking about AI and job security.
Share this article