AI’s Brokenomics

Source: Where’s Your Ed At Date: 2026-06-15 URL: https://www.wheresyoured.at/brokenomics/

Summary

Ed Zitron argues that OpenAI and Anthropic lack viable business models: token-based subscription pricing allows heavy users to consume 20–70x their subscription cost in value (a $200/month plan can represent $14,000 in token spend), making most customers unprofitable at any plausible margin. After enterprise customers saw unexplained AI bills with no measurable ROI following a shift to transparent token billing in Q1 2026, both companies were reportedly planning significant price cuts — which Zitron frames as a death spiral, since cuts worsen already-negative unit economics. The piece also cites dependence on hyperscaler infrastructure (Microsoft, Google, Amazon) as evidence that neither company can operate as an independent business.

Implications

• AI ecosystem economics: this is the most pointed public articulation to date of the subscription-vs-actual-cost gap. The Q1 2026 enterprise billing backlash (Uber, Meta, banks) is a real signal that enterprise adoption is stalling on cost-clarity problems, not just capability gaps.
• Work AI-adoption timing: enterprise customers pulling back after seeing transparent token costs supports a slower adoption curve than vendor projections suggest. ROI proof gaps at scale remain unresolved.
• OpenAI/Anthropic distribution strategy: the impending price cuts signal a race-to-commoditization dynamic that benefits hyperscalers and vertical integrators over standalone model companies.