Model routing is a fix for AI overspending. That’s a problem for OpenAI and Anthropic Companies are implementing “model routing” to optimize AI spending, matching tasks to the most cost-effective model rather than defaulting to expensive frontier models for every query. This strategic shift is driven by escalating AI bills and aims to improve cost efficiency, potentially reshaping the AI market. Major AI providers like OpenAI and Anthropic may face pressure as their premium pricing models are challenged by this new approach.

• Companies are adopting “model routing” to control rising AI expenses.
• Model routing assigns tasks to the appropriate AI model based on complexity, using cheaper alternatives for simpler jobs.
• Currently, most enterprise AI usage runs on expensive frontier models, even for basic tasks.
• High AI costs, such as $10,000 per employee annually, are prompting this spending discipline.
• AI companies are offering guarantees to demonstrate return on investment, focusing on actual value delivered rather than token usage.
• This shift could impact the business models and valuations of leading AI companies like OpenAI and Anthropic.
• Pricing power is moving from AI vendors to AI buyers.