Businesses Overhaul AI Budgets as Pricing Shifts to Usage

Every generative AI inference consumes GPU capacity, electricity and tokens, upending the SaaS budgeting model of fixed annual per-seat fees. PYMNTS said businesses must spend another $5–$10 on integration, compliance and monitoring for every $1 paid in model fees. That is prompting CFOs to demand measurable returns on investment and move toward usage- or outcome-based pricing.

A KPMG survey in June 2026 found that businesses planned to invest an average of $202 million in AI over the next 12 months, but only 26% could track operating costs in real time. Fortune reported on June 18 that Uber had exhausted its entire 2026 AI budget in four months, while another company was spending $500 million a month. The costs are pushing finance executives to accelerate the adoption of token caps, dynamic budgets and outcome-based pricing.