OpenAI and Anthropic are promoting a new metric for measuring the cost of AI, according to Bloomberg Technology. The report says the companies want customers to rethink the price of using their models. The provided report summary does not name the metric or provide pricing figures, customer examples, or technical details. That limits what can be concluded today: the core development is not a new model release, but a push by two leading AI firms to frame model costs differently for buyers. For enterprise customers, model pricing is often evaluated through the cost of usage. Bloomberg’s framing suggests OpenAI and Anthropic are trying to shift that conversation toward a broader or different measure of cost, though the available summary does not specify how the proposed metric works. That distinction matters because buyers comparing AI systems need a way to translate model performance and usage into budget impact. If vendors can define the preferred cost metric, they can also influence how procurement teams compare competing systems. For now, the story should be treated as developing. Bloomberg identifies the companies and the direction of the pitch, but the public summary does not yet establish the formula, adoption status, or whether customers are accepting the new framing. Who benefits: OpenAI and Anthropic may benefit if customers adopt a pricing framework that makes their models look more favorable. Customers could benefit if the metric makes AI usage costs easier to compare, but the evidence provided does not establish that yet. Who's exposed: Model buyers are exposed if the new framing obscures rather than clarifies real usage costs. Rival AI providers could also be affected if procurement comparisons begin using a metric promoted by OpenAI and Anthropic.