WorkThe story, in brief

Model routing is a fix for AI overspending. That's a problem for OpenAI and Anthropic

Model routing just became the cost-killer that threatens OpenAI and Anthropic's unit economics.

Illustration of a transparent lens revealing connected networks across layers of paper.
Exploring the next frontier of AI research.AI illustration by KeyNews
The KeyNews take

Why it matters

Companies are adopting model routing—matching tasks to appropriately-sized models rather than defaulting to premium ones—creating a structural shift in AI spend that pressures frontier labs' revenue models and forces a reckoning on pricing strategy.

The key facts

4 to know
  1. Model routing practice enables cost optimization by matching tasks to right-sized models

  2. Trend threatens revenue models of OpenAI and Anthropic (heavy reliance on premium model spend)

  3. Represents shift from 'run everything on best model' to 'match task to appropriate capability tier'

  4. Published June 2026 — reflects emerging market behavior with strategic implications for business model sustainability

Go to the source

CNBC Technologycnbc.com

Publisher excerpt: Companies are shifting from running everything on the most powerful AI model to matching each task to the right one, a practice called model routing.
Read original report
Back to today's editionMore work news

Keep reading

Related stories

More from Work