The AI Pricing Conundrum: It Started as a Nightmare, Now It's Worse
Computerworld, Tuesday, June 2nd, 2026
Enterprise IT struggles to find viable AI pricing models that balance vendor profitability with provable ROI.
Enterprise IT leaders face an unsolvable pricing dilemma with AI vendors, as current token-based or per-task models fail to align vendor interests with business outcomes.
Buyers want pricing tied to realized business value, while AI vendors want pricing tied to resource consumption.
The core problem is that pricing must be locked in before projects begin, yet the benefits of emerging technology remain unclear and hard to quantify. Outcome-based pricing creates perverse incentives, potentially pushing AI systems to optimize metrics rather than serve genuine business needs.
The article recommends forming AI committees with technical experts to evaluate projects, quantify realistic benefits and failure scenarios, and tie executive bonuses to results before negotiating fair deals.