Two-thirds of Forbes AI 50 companies adopt usage-based pricing as AI monetization shifts to hybrid models

Two-thirds of the companies on the Forbes AI 50 list now use some form of usage-based pricing, a shift driven by AI’s variable costs and unpredictable demand. At the second annual Monetize conference, Metronome CEO Scott Woody characterized these leading companies as operating in ways that would seem “alien” to traditional SaaS businesses, noting their focus on rapid pricing iteration and market share over ARR.

Hybrid pricing and predictability

Companies are increasingly adopting hybrid models that package variable usage within subscriptions, often using credits to simplify multiple usage units. Fin evolved from a platform subscription to a hybrid model with outcome-based pricing, allowing customers to commit to a pool of credits and draw down against various billable outcomes, such as resolved support conversations or qualified leads. Simultaneously, predictability has become a critical product feature; Snowflake and Stripe highlight the importance of in-product tooling like forecasting dashboards and usage control panes to help customers monitor and cap spending.

Flexible packaging and cost dynamics

As AI inference costs fall due to advances in open-source models, companies are adjusting packaging to maintain margins while offering flexibility. Replit initially charged 25¢ per agent checkpoint but switched to usage-based billing as runtimes grew and costs exceeded the fixed price. To address user concerns about unpredictable bills, Replit introduced a rate-limited Free Mode and adjusts its pricing several times a month to pass savings from lower model costs to customers.

Measuring value and machine-readable pricing

Finance teams are moving away from token-level ROI calculations, which are often misleading due to complex attribution. Clay treats model tokens and third-party data as pass-through inputs, instead measuring AI investments against function-specific targets like productivity and development time. Additionally, pricing information must now be accessible to both humans and AI agents. A study by Growth Unhinged founder Kyle Poyar found that when pricing was not machine-readable, LLMs relied on informal sources, resulting in reported prices ranging from $100 to $2,400 per user per month for a single company.