SaaS Pricing in 2026: What Two Recent Studies Report

Scope of this article

This is not a survey of the whole SaaS market. It reports what two published studies say about pricing, with their sample sizes and the interests of their authors stated, so you can judge how far each number travels. We do not make predictions of our own.

Study one: pricing among 80 AI agent companies

Orb's 2026 study looked at 80 AI agent companies (the study states a 90% confidence level). Its figures, with the 2025 comparison where the study gives one:

• 91.3% use usage-based pricing, up from 83.3% in 2025.
• 95% combine more than one pricing mechanism, up from 92.4% in 2025.
• 71.3% keep a subscription element, up from 62.1% in 2025, and 94.7% of those that charge a subscription also charge for usage.
• 3.8% use outcome-based pricing, down slightly from 4.5% in 2025.

How to read it: Orb sells usage-based billing software, the sample is AI agent companies and not SaaS in general, and the article does not state its publication date.

Study two: the hy and OMR Reviews pricing report

The SaaS & AI Pricing Report 2026 was produced by the consultancy hy with OMR Reviews. It draws on an analysis of OMR Reviews software profiles (4,197 in its main sample, as of 30 June 2025) and on an hy survey about monetizing AI agents and AI features in software companies (n = 180). The report's own executive summary states:

• The dominance of pure subscription models is waning, and the focus is shifting from per-seat metrics toward usage-based metrics such as transactions or API calls.
• "Currently, 37% of companies use usage-based models, and 69% plan to switch within two years."
• Most SaaS companies still use a package structure with three or four pricing tiers.
• Over 50% of SaaS companies adjust their price levels annually.

It also reports that 92% of traditional software is still sold by subscription, against 83% for AI features. How to read it: hy is a pricing consultancy, the survey is small (n = 180), and the data is as of mid-2025, so it is older than the Orb study.

Where the two studies agree and differ

Both point to usage-based and hybrid pricing becoming more common, and both show subscriptions remaining in place alongside it. They measure different groups (AI agent companies in one, software companies and software profiles in the other) and different dates, so their percentages should not be compared directly or averaged.