Thanks, really fair feedback. Honestly we're still landing on exact tiers as we work with our early customers. The model is per GPU under management, lower at scale. Happy to chat if you want to learn more — [email protected]