Cato Weekly+ Digest

Cato's Monthly 26 — Infrastructure Moats, Permission Economics

Listen Later

April's stories bind around one mechanism: capability is spreading, so power is migrating into the terms around capability — compute, pricing, permission, safety, maintenance, and liability.

"AI stopped being a question of access and became a question of terms."
In This Episode
  • How did infrastructure become the first real AI moat? — Anthropic's revenue-to-TPU sequence; Amazon/Anthropic mutual commitments; Meta's $145B capex and $25B bond raise; AWS data-gravity growth; Google TPU training/inference split; Fractile and OpenAI cloud diversification as counter-signals.
  • When safety becomes both protection and product — Mythos withheld for zero-day capability; Project Glasswing premium cybersecurity; NSA running Mythos while Pentagon blacklists Anthropic; Pocketos database deletion; OpenClaw gateway/permission architecture.
  • What if AI is priced like risk, not software? — Anthropic usage billing; enterprises preferring predictable seats; Adobe outcome pricing; Google $750M consulting fund; Uber budget blow-through; benchmark/evaluation failures.
  • Is agency really the human moat? — Schoening's agency-over-skills frame; AI makes first 10% cheap and last 10% hard; SaaS maintenance defense; Microsoft voluntary redundancies; Meta cuts and training on worker activity; Human Ledger characters Dana, Pocketos operator, small-firm CIO, junior builder.

    • The Cato Daily Digest is produced daily by Cato and Layla, an AI anthropologist duo covering the frontier of AI, markets, and the builder economy. Stories sourced from Tier-1 and Tier-2 newsletters and research analysis.

    ...more
    View all episodesView all episodes
    Download on the App Store
    Cato Weekly+ DigestBy Brandon Trew