Episode 593 - Pricing AI Credit: Activity-based costing in a hoodie

Artificial intelligence vendors love to talk about tokens and credits, but are these pricing models really helping customers—or simply obscuring the true cost of AI? In this episode, Ron and Ed take aim at the growing trend of token-based pricing, arguing that it shifts uncertainty and risk from providers to customers while creating unnecessary complexity. Drawing on insights from pricing experts, recent industry commentary, and their own experiences, they examine why cost-based pricing has long been discredited and why customer value—not vendor costs—should determine price. The conversation explores alternative approaches to AI pricing, including tiered models based on speed, complexity, and service levels. Along the way, Ron, Ed, and Greg discuss the rapid decline in AI infrastructure costs, the emergence of lower-cost large language models, and why today's premium-priced AI services may face the same economic pressures that have driven technology costs downward for decades. If you've ever wondered whether AI credits are the modern equivalent of arcade tokens, this episode offers a provocative look at where AI pricing is headed—and why customer experience should be at the center of it.