I Read OpenRouter's 36-Page AI Report So You Don't Have To

OpenRouter and A16Z analyzed two years and 100 trillion tokens of real AI usage data—and the findings are not what you'd expect. Programming and role-playing dominate AI use. Unlike usage claims often promoted by the model providers, this isn't marketing hype, it's empirical data from OpenRouter's State of AI report tracking how millions of users worldwide actually interact with AI models. The results might change your opinion on current AI adoption and reveal where the AI bubble actually stands. I break down the most surprising insights: how role-playing is the #2 use case, how programming has been the main driver in AI growth, represented by a 4x increase in input tokens, 3x output tokens, and more. ⏱️ **TIMESTAMPS** 0:00 – Intro: OpenRouter's 100 Trillion Token Study 0:30 – The Reasoning Model Revolution 2:20 – Programming & Role-Playing: AI's Real Use Cases 4:16 – Rise of Agentic Inference & Tool Calling 5:55 – Tool Calling Adoption (MCP Era) 7:27 – Why Programming Dominates AI Usage 8:51 – Role-Playing Breakdown 9:39 – The Glass Slipper Effect Explained 10:36 – Cost vs Usage Dynamics 11:13 – Big Picture: AI Bubble, Scaling, and Code Red 13:28 – Final Thoughts 🔗 **RESOURCES** OpenRouter State of AI Report: https://openrouter.ai/state-of-ai Book a call with me → https://yedatechs.com/#container06 Sponsorship inquiries → hi@yedatechs.com #OpenRouter #StateOfAI #AIUsage #ClaudeCode #AIDevelopment