Almost everyone knows Tailwind CSS. For many developers, it feels like the default choice when styling a modern web application. When we start a new project, we usually consider many technical decisions, but styling often feels already decided: Tailwind seems like enough.
By January 2026, Tailwind’s reported usage metrics had reached record highs, with around 75 million monthly NPM downloads and appearances in more than 7.8 million GitHub projects. Those numbers naturally lead to an assumption: the company behind Tailwind must be making a great deal of money.
The reported business picture is more complicated. According to the figures and statements discussed in this article, revenue had fallen by around 80%, while traffic to the Tailwind website had dropped by approximately 40% since early 2023.
How can usage keep growing while revenue and website traffic decline? The answer may be connected to one of the biggest changes in software development today: AI.
On January 6, Adam Wathan, the creator and CEO of Tailwind CSS, rejected a community pull request proposing an /llms.txt endpoint for the Tailwind documentation. The endpoint would have provided a text-only version of the documentation, formatted for large language models (LLMs). In practical terms, it would have made the documentation easier for AI systems to read and use without sending users through the Tailwind website.
The community questioned that decision, so Wathan explained why the pull request was closed. He said that making the documentation easier for AI systems to read could reduce traffic to the documentation and, as a result, reduce the number of people discovering Tailwind’s paid products.
He also stated that 75% of the company’s engineering team had lost their jobs because of the severe effect AI had had on the business. These statements should be understood as the company’s own account of the situation, and the financial figures should be checked against their original sources before being treated as independently verified facts.
Tailwind’s business model has traditionally depended on a simple trade: the core framework and its documentation are free, while the company earns revenue from paid products such as Tailwind UI, a collection of pre-built interface components.
The model works when developers discover Tailwind through the documentation and later decide that buying polished, ready-made components is worth the time saved. A developer may begin by building a component from scratch, struggle with the details, and eventually think: “I would rather buy a finished component.” That conversion is part of how the free framework supports the business.
AI changes that path. If an AI system can read the full documentation and generate a component from a short prompt, a developer may never visit the Tailwind website. They may get the implementation immediately, without discovering Tailwind UI or any of the company’s paid products.
This does not mean Tailwind CSS is failing as a technology. It may mean that popularity, downloads, and business revenue no longer move together. AI can increase the use of a free tool while reducing the human attention and website traffic that once helped convert that use into paid sales.
The larger lesson is about the economics of open-source software. Documentation is not only educational material; for some companies, it is also the main path to distribution, brand discovery, and revenue. When AI systems consume that knowledge and return the result without sending users to the original source, that path can be disrupted.