The Hidden Costs Behind Generative AI's Economics

Ed Zitron argues that the economics of generative AI are fundamentally broken: subscription pricing conceals the true per-token costs, data-center capitalism is brutally capital-intensive with slim margins and heavy debt, and OpenAI/Anthropic depend on unsustainable burn rates and venture funding. As token-based billing spreads (e.g., Copilot), the industry faces a reckoning where profitability seems unattainable, media coverage often understates real costs, and investors may rethink AI bets as revenue targets appear impossible to reach.