Publishers are encountering a contradiction in how they are paid for content used by AI. Licensing deals with AI companies are reportedly slowing down, despite increasing demand for content from AI. Publishers are openly discussing this slowdown, with some market observers indicating the deal market is effectively closed. Simultaneously, AI bots are attempting to access vast amounts of publisher content daily, sometimes requesting millions of articles. This situation has led publishers to block many bots, as scraping incurs costs. The industry is shifting towards metered pricing models to address this. Initially, this focused on "pay per crawl" but is evolving to "pay per use," where publishers are compensated only when their content contributes to an AI-generated answer. Data suggests AI models only use a fraction of the content they access for any given query. However, uncited content still shapes AI responses and has value in wider indexes. This raises questions about who should pay: AI applications seeking specific answers, or those building the comprehensive indexes AI draws upon. A third party, AI data brokers, also plays a role, crawling the web and selling access to content. While some brokers are starting to partner with publishers for payment, their model of acquiring content for free and reselling it is debated. The value of content for inclusion in a broad AI index, beyond its direct use in answering a specific query, is also significant. A flat "index fee" could incentivize publishers to make their content available for this broader purpose. Some AI platforms already define "use" to include content indexing, suggesting payment for the library itself is happening. The distinct value of smaller, long-tail publishers for specific local queries is also being recognized. Analogies like Spotify are drawn, suggesting a need for a more structured payment system with potential advances and guaranteed minimums for publishers. However, major players like OpenAI and Google have their own established methods, with Google's News AI pilot payment structure being opaque. Ultimately, a pricing system that compensates for both the "ingredients" and the "pantry" is needed for publishers to sustainably create new content for AI markets.