OpenAI has started testing a pricing model where enterprise customers pay based on successful outcomes, not compute consumed.
The appeal is obvious. You pay when the agent actually finishes the task, not while it burns tokens trying. It reframes agents as a service that delivers results.
The hard part is definition. Someone has to decide when a task counts as complete, and determining whether an agent did the right thing is genuinely difficult.
If this spreads, it changes how agents are sold and how they are evaluated.
Read the full analysis. The Cost of AI Is Finally Falling. The Cost of Using It Is Not.
The AI Infrastructure report covers where the money actually sits, including the shift from token price to the cost of a whole workflow.
Get the next one before it is old news
Independent analysis of cloud-native infrastructure, Kubernetes and data centre economics. No vendor spin.