Provider Concentration Ratio

Provider Concentration Ratio measures what share of an agency's AI spend and client deliverables routes through one model provider. Above roughly 60 percent, the agency loses pricing leverage: a single API price change, deprecation, or terms update resets the economics of every retainer built on it. The framework matters because multi-model orchestration is cheap to add early and expensive to retrofit once client code, prompts, and evals are tuned to one vendor. A concrete trigger: Anthropic's IPO prospectus disclosed tens of billions in annual losses alongside a formal risk warning, which is exactly the kind of filing that precedes pricing or availability changes for agencies standardized on Claude. Gateways such as Helicone, Portkey, and OpenRouter let a delivery team route the same request across Anthropic, OpenAI, and Google endpoints without rewriting client integrations, so concentration becomes a routing decision rather than an architecture rebuild. Track the ratio quarterly per retainer, not per agency.