The Mother Of All Meta Layoffs?

What It Covers

Meta reportedly plans layoffs affecting 20% of its 79,000-person workforce to fund a $600 billion AI infrastructure buildout by 2028. The episode also covers Apple's contrarian AI spending strategy, San Francisco's housing rebound, OpenAI's delayed adult mode, and new AirPods Max 2 at $549.

Key Questions Answered

• •Meta's labor-to-CapEx pivot: Meta may cut roughly 15,800 employees — redirecting payroll savings directly into AI chip infrastructure. This follows its 2022–2023 "year of efficiency" cuts of 21,000 jobs. Executives are explicitly framing reduced headcount as a funding mechanism for data center expansion, signaling that AI infrastructure costs now compete directly with human labor budgets.
• •Apple's asset-light AI strategy: Rather than matching hyperscaler CapEx, Apple licenses Google Gemini for approximately $1 billion annually and embeds AI processing into hardware — the M5 chip runs 70-billion-parameter models locally. This approach lets Apple swap AI vendors if better models emerge, avoiding the $650 billion infrastructure commitments draining competitors' free cash flow by up to 90%.
• •Hyperscaler debt risk: Amazon, Alphabet, Meta, and peers collectively raised $121 billion in bonds in 2025 alone, with Morgan Stanley projecting $1.5 trillion in tech debt ahead. Amazon is projected to go negative free cash flow by $28 billion this year. AI services currently generate only $35 billion in revenue against $650 billion in infrastructure spending — a 5% return ratio.
• •San Francisco housing as AI economic indicator: SF rents rose 14% year-over-year in February 2026, the fastest in the US. Condo prices jumped 12% and single-family home prices surged 23%, with medians at $1.23M and $1.96M respectively. Anticipated IPOs from Anthropic and others are expected to intensify competition further, making SF real estate a real-time proxy for AI sector confidence.
• •OpenAI's adult mode age-verification failure: OpenAI's age-prediction system, designed to block minors from explicit ChatGPT conversations, misclassified minors as adults approximately 12% of the time. With roughly 100 million under-18 users weekly, that error rate could expose millions of minors. The company delayed the feature's Q1 launch while continuing to develop text-only erotic conversations with image, voice, and video restrictions intact.