I’ve been thinking about pricing models for AI products and wanted to get some perspectives.
Traditionally, most SaaS products follow a subscription model (monthly/annual). But with AI tools — especially compute-heavy ones — this model feels increasingly misaligned with actual cost.
I’m starting to see more products experimenting with usage-based pricing, hybrid models (base subscription + usage), or even fully pay-as-you-go.
From a builder perspective, this isn’t just a pricing change — it affects: product architecture (tracking usage, cost control), UX (how transparent pricing is to users), growth & conversion (friction vs flexibility)
So I’m curious: Are you sticking with traditional SaaS pricing, or moving toward usage-based? Have you tested hybrid models? Does usage-based pricing help or hurt conversion in your experience? As a user, do you prefer predictability (subscription) or flexibility (pay per use)?
Feels like this shift is especially relevant for anyone building AI or infra-heavy tools.
Would love to hear what others are seeing in practice.