OpenAI, Anthropic, and the strange math of AI’s “returns”

OpenAI, Anthropic, and the strange math of AI’s “returns”

A community post asks the blunt question: can something be transformative and disruptive and still deliver disappointing economics? The piece says it’s a sequel to a Jan 2026 look at value capture in AI inference, and this time it turns to Anthropic’s revenue surge, “tokenomics,” and publicly available API pricing numbers. It also contrasts OpenAI’s huge visibility and user base with the fact that most users are free, while Anthropic is framed as going after enterprise customers, especially with Claude code. The post says a Wall Street Journal look at financial documents Anthropic gave investors came on May 20, 2026.

Gobble's Take: The hype is loud; the unit economics are the part that whisper back. Source: Perplexity Search

In Case You Missed It

Yesterday's top stories:

• SpaceX, passive investors, and the market mood