When Mark Zuckerberg announced 11,000 layoffs in late 2022, he framed it as a mea culpa for pandemic-era overhiring. Three years and 21,000 job cuts later, the tone has shifted. Starting this week, Meta is trimming another 8,000 roles—roughly 10% of its workforce—while abandoning plans to fill 6,000 open positions. There’s no apology this time. The message is straightforward: headcount reduction funds AI infrastructure.

Meta’s capex guidance for 2026 now reaches $145 billion, up $10 billion from prior estimates. Finance chief Susan Li acknowledged on the Q1 earnings call that the company consistently underestimates compute demand. Internally, the mood is grim. Blind ratings show a 25% drop since mid-2024, with culture scores falling 39%. Employees cite the Model Capability Initiative (MCI)—a tool tracking mouse movements and keystrokes to train AI agents—as particularly dystopian. An internal petition calls it nonconsensual data extraction.

The broader pattern is industry-wide. Layoffs.fyi reports nearly 110,000 tech cuts in 2026, approaching 2023's peak of 260,000. Cisco just shed 4,000 roles alongside strong quarterly results. The logic is blunt: investors reward AI investment, not headcount. As Kingsley Gate’s Umesh Ramakrishnan put it, “Jobs are being replaced by machines, and if you’re not doing that, shareholders get upset.”

Meta’s stock is down 7% year-to-date, underperforming peers except Microsoft. Wall Street remains skeptical of a scattered AI strategy. Inside the company, some long-time engineers are eyeing exits. The question isn’t whether AI will reshape Meta—it’s whether the engineering culture can survive the transformation.

Source: CNBC