AI Hasn't Taken Jobs at Meta, Microsoft, Oracle, Yet

Since 2020, nearly 900,000 tech workers have been laid off  globally, according to the tracking site Layoffs.fyi. 

More recently, on April 23, 2026, Meta announced it was cutting 8,000 jobs (10 percent of staff) and cancelling 6,000 open roles effective May 20, while Microsoft said it will offer voluntary retirement benefits for up to 8,750 US employees whose age plus years of service equals 70 (about seven percent of its U.S. workforce).

That will be interpreted by some as more “evidence” that artificial intelligence is displacing humans at work.

But companies are, for the moment,  reallocating capital from labor costs (payroll, benefits, and related overhead) to massive capital expenditures on AI infrastructure. 

The moves do not reflect an actual displacement of humans by AI workflows. 

Company	Job Impact (2026)	Stated Rationale	AI Infrastructure Spending (Key Figures)	Sources
Meta	~8,000 layoffs (10% workforce) + 6,000 roles unfilled (effective May 2026)	Efficiency to offset AI investments; become "AI native"	2025: $72.2B capex; 2026: $115–135B (data centers, GPUs, Llama support; nearly double prior year)	NYT, Forbes, PYMNTS, BBC
Microsoft	Voluntary buyouts for ~7% of U.S. workforce (~8,750 eligible)	Cost management/reshaping amid AI shift; first broad buyout program	~$100–120B estimated relevant spend; recent deals incl. $18B Australia, prior Japan commitments	CNBC, Inc., Guardian
Oracle	20,000–30,000 cuts (12–18% global workforce)	Fund AI data center buildout amid cash/debt pressures	Capex ramp to ~$50B+ for FY2026 (data centers for AI workloads, OpenAI-related contracts)	CNBC, Forbes
Broader Big Tech (e.g., Amazon, Google/Alphabet)	Thousands in ongoing/prior rounds (e.g., Amazon corporate cuts); part of industry-wide ~50K–90K+ tech layoffs in early 2026	Efficiency, flattening, offset heavy AI buildouts	Combined Meta/MSFT/GOOG/AMZN: ~$650–700B capex in 2026, majority AI infrastructure (data centers, compute)	CNBC, QZ

In a nutshell, this is less about "AI took my job" in a narrow automation sense and more a balance-sheet shift. 

Firms are trading human capital costs for compute capital. 

Actual AI productivity gains could reduce headcount in the future, but current layoffs are driven more by funding the infrastructure foundation.

AI has not “taken your job,” yet.