AI Skeptic: This Business Makes No Sense

Ed Zitron, author of the "Where's Your Ed At" newsletter and host of Better Offline, joins Prof G Markets host Ed Elson to critique the AI boom. Amid market gains driven by AI hype—Nasdaq up 27% since late March—Zitron argues public companies show no AI profits, revenues are circular, and massive capital expenditures face severe constraints.

AI Revenue: Circular and Inflated

Public tech giants like Microsoft, Google, and Amazon report AI-related revenue backlogs totaling $748 billion, but over half stems from just two labs: OpenAI and Anthropic. These labs buy compute from the hyperscalers—Microsoft Azure for OpenAI, Google TPUs rented back from Anthropic—creating a loop where venture-backed labs fund their partners' "growth." Outside this, no broad revenue streams exist.

Zitron questions private ARR figures: OpenAI claims jumped from $2 billion in 2023 to $25 billion; Anthropic hit $30 billion. Yet Anthropic's CFO affidavit in a lawsuit pegged lifetime revenue at $5 billion through early 2026, far below summed ARR reports. He suspects upfront token purchases by big clients inflate monthly run-rates, as clients prepay for bulk usage lasting months, booking full value immediately rather than over time.

Smaller AI startups linger at $100-300 million ARR—modest for a supposed revolution—with no other clear winners. Cursor, once prominent, faded without acquisition.