Tailwind CSS reported an 80% revenue drop, sparking fears that AI is killing open source. But that panic misses the real story. While documentation-driven projects struggle, cloud service providers like Databricks and Vercel are raising billions. The difference? AI scrapes documentation without driving sales, breaking one specific business model while leaving others untouched.

The failing model relies on traffic from docs converting to paid UI kits. AI simply caches that information, cutting off the funnel. Successful companies sell what can't be cached: hosted infrastructure, data processing, and enterprise contracts. These services force developers to connect to real backends, ensuring revenue keeps flowing even as AI changes how code is consumed.

A second pressure point is quality control. AI has flooded repositories with low-effort contributions, forcing projects to tighten guidelines and require Contributor License Agreements. The solution isn't abandoning open source, but choosing the right license. Many founders are shifting toward AGPL-3.0 to protect against proprietary forks while keeping code accessible.