Special Edition: The pricing story this week

The AI vendor pricing model just stopped being a finance problem. It is a strategy problem now.

TechCrunch coined the term Tokenpocalypse this week. The broader story is what matters. Capital events accelerating toward public market scrutiny. Token-based metering replacing flat-rate seat licenses. Unit economics that have been hiding behind venture subsidy, about to be priced in. OpenAI's S one trajectory. Anthropic's continuing capital build. The actual cost of frontier inference is about to become legible to every CFO who reads a ten K. That is going to change what AI vendor contracts look like by Q four.

Here is the harder part. Your AI-enabled product was probably priced under one set of vendor cost assumptions. Your customer contracts probably lock you into another. The gap between what you pay your vendors, and what your customers pay you for AI-enhanced features, is the variable that will compound or destroy your gross margin in the coming year or two.

Companies who built pricing models with elasticity to vendor repricing, who packaged AI capability into the right value metric for their customer's job-to-be-done, will compound. Companies who treated AI vendor cost as a flat operating expense, bolted onto seat-based pricing, are going to spend the next year defending margin to their board.

The question your board should be asking next quarter is not whether you have an AI strategy. It is whether your pricing model is structurally compatible with the AI cost structure you are about to live with. Three questions to bring to the next board meeting. Do your customer contracts let you reprice when vendor costs reprice? Is the value metric you charge on actually correlated with your AI vendor spend? Have you stress-tested gross margin against a two times token cost increase, and a fifty percent decrease? Most companies cannot answer those three questions with conviction today. The ones who can will be the ones whose PE owners are still happy with them in twenty twenty-seven.

Pricing is no longer something you set once at launch and revisit when the board complains. AI made pricing a continuous discipline. The companies who treat it that way will earn back their exit multiple. The ones who do not will explain to their board why margin compressed.

See you next week. Just Keen A.I. dot com.