Capacity pricing gains ground as 80% of AI software vendors choose fixed commitments: Bain

Capacity-based pricing is gaining traction among AI software vendors, with 80% of them favoring fixed commitments over pure consumption-based pricing, according to a Bain & Company report. The report, which analyzed pricing structures of approximately 200 B2B software-as-a-service (SaaS) companies, found that capacity models offer customers more predictable budgets while ensuring vendors secure stable revenue.

Under this model, customers commit to a fixed amount of AI usage over a set period, typically without refunds or carryover of unused capacity.

Bain highlighted that capacity-based pricing maintains some of the economic benefits of traditional seat-based software pricing. Vendors continue to receive revenue from committed entitlements, even when customers fail to fully utilize the allocated capacity. For customers, this model simplifies budget planning, allowing procurement teams and CFOs to approve defined commitments instead of facing potentially infinite variable costs.

The report pointed out that while AI pricing is shifting away from traditional per-seat licensing, seats are not entirely disappearing. Around one in five AI-native software companies still rely primarily on per-seat licensing, often combined with usage entitlements. Most companies adopting hybrid AI pricing models are adding new pricing tiers rather than replacing existing seats.

Among companies introducing hybrid AI pricing approaches, output-based models account for 55%, followed by effort-based models at 35%, and outcome-based models, which account for only 10%. Bain noted that outcome-based pricing, where vendors are paid based on business results rather than the work produced, has a more limited role.

It is primarily suitable when outcomes are observable, uniquely attributable to AI, and clearly agreed upon between buyer and seller. Outcome-based pricing is expected to remain valuable but remain relatively narrow in scope over the coming years.

The report concluded that direct usage pricing will persist, especially in infrastructure software and products sold to technical buyers. However, capacity-based pricing is expected to remain the favored approach for most enterprise applications. Bain emphasized that outcome-based and usage-based pricing models are currently the headline strategies in AI pricing, with the specific meter being the key differentiating factor in the pricing strategy.