Model exchanges are becoming the AI market's quiet distribution layer. OpenRouter's BusinessWire release says it raised $113 million in a CapitalG-led Series B as weekly volume reached 25 trillion tokens, a scale claim that only makes sense if routing has become more than a developer convenience [2].

The model-lab race gets the attention because names and benchmarks are easy to argue about. The exchange layer may prove more durable because it solves a different problem: developers and companies want access to many models without rebuilding procurement, billing, fallback, and traffic decisions for each one [2].

Dentro's AI timeline shows why that layer is valuable. In one week it records launches, subscriptions, image products, funding, and benchmarks. A market moving that fast creates demand for brokers, dashboards, and routing systems that reduce switching friction [3].

Lionsgate's release provides the useful non-AI contrast. A media catalog earns value by repeated licensing [1]. A model exchange earns value by repeated access decisions. The business question is whether OpenRouter can make itself the place where those decisions happen often enough to justify venture-scale expectations. Distribution power starts as convenience and ends as leverage.

OpenRouter is the distribution source; the broader market claim stays there. [2]

-- THEO KAPLAN, San Francisco