Woofun AI reports that the AI industry is transitioning from a high-consumption "tokenmaxxing" model to a cost-efficiency framework, constraining growth expectations for major developers like OpenAI and Anthropic. Companies are optimizing model call costs; for instance, Lindy CEO stated switching 100% of traffic from Anthropic's Claude to DeepSeek to save millions, while Uber implemented monthly internal AI spending caps.

Analysts note that as enterprises move toward "refined calling", the previous high-growth model faces challenges despite Anthropic's $47 billion annualized revenue and OpenAI's $25 billion run rate. Emerging "model routing" technologies utilize lower-cost models for simple tasks to optimize computational expenses.

Meanwhile, Microsoft, Amazon, and Google are accelerating low-cost model launches, compressing price margins and increasing IPO window pressure for major AI firms.