Meta spent $1.18B firing 8,000 people. Then its AI replacement plan imploded.

Meta paid $1.18 billion in severance to cut 8,000 workers in May.

The plan was bigger than the layoffs.

Reuters revealed Project OT — an internal initiative to make Meta "AI native" — contemplated reducing some teams by up to 60%.

Zuckerberg killed it.

Employee resistance. Weak AI agent performance. Internal systems that couldn't meet expectations.

CTO Andrew Bosworth called the communication of the vision "atrocious."

Then the numbers arrived.

40% more major site-reliability emergencies than a year earlier.

Employees spending 70% more time fixing failures.

AI-generated code and unpredictable agent behavior driving the spike.

The efficiency argument collapsed.

Meta spent $31B in a single quarter on infrastructure. Free cash flow dropped 91% to $784 million.

Revenue grew 28%. Net income fell 14%.

The company just launched Muse — an AI agent that accesses email, calendars, and payments — while its internal AI systems were still breaking production.

October earnings will tell you whether the AI investment is working or whether Meta is funding its own disruption.

Audit your AI deployment assumptions. Internal capability is not the same as production readiness.

SOURCE: https://tradesprophet.com/2026/09/11/metas-meta-ai-layoff-controversy-returns-weeks-before/
VERIFIED: Trades Prophet (Sep 11 2026), Reuters (Sep 3 2026), Associated Press
SIGNAL: Meta's Project OT failure is the most expensive proof point yet that AI workforce replacement doesn't work the way the pitch deck promised — and the October earnings call will force every executive to confront that.