Coding agents drive revenue growth at AI labs faster than infrastructure can scale, creating a cash flow crisis despite record demand

Anthropic's revenue is up 3x since the start of the year, and Claude Code commits on GitHub (in publicly accessible repos) went from 2% to 4% in January alone. Dylan Patel forecasts Claude hitting 20%+ by end of year. OpenAI and Anthropic are posting rapid growth ahead of potential IPOs.

Coding agents are the first AI product people are paying for at volume and regularly because they directly speed up work. A $100 subscription can deliver 10-30x ROI for a developer earning $350-500 a day if it automates just the simplest 10% of their work, making the economics compelling for early adopters.

Anthropic rents GPUs from hyperscalers (Google, Amazon, Microsoft, Meta) but cannot freely request more capacity because datacenter owners are also exposed to bubble risk and book capacity in advance. For Anthropic to make $30B a year, someone had to spend $80B on infrastructure, creating a structural mismatch between demand growth and available compute.