Board Case Study · Technology (Platform)
Every AI feature runs on one external model. Overnight the provider changed terms, raised prices 5×, and restricted a key capability. There is no fallback.
The situation
A company built its entire AI product line on a single third-party foundation model. It was the fast path to market and it worked, until the provider changed commercial terms overnight: a roughly 5× price increase, a usage restriction on a capability central to the flagship product, and new data terms.
There is no abstraction layer and no tested fallback. Switching models would mean re-engineering and re-validating every feature, and quality would vary. Margins built on the old pricing evaporate; the roadmap assumes capabilities that may no longer be available.
The board realizes the company's strategy rested on a dependency no one had named as a board-level risk. The questions are concentration, resilience, and who owned this risk while it grew.
“One vendor just repriced our entire product line and we have no plan B. How did single-model dependency never reach this table as a risk?”
Chair, Technology Committee
The decision on the table
The full case, scenario architecture, board materials, and facilitation notes, is shared with boards and partners on request.