AI is quietly repricing your software stack

What happened

L.E.K. Consulting reviewed how AI is changing SaaS pricing at the end of 2025. Its main points:

• Pricing tied to usage, such as tokens, data processed or API calls, is gaining on seats and flat fees.
• Compute and data costs are rising, which exposes the weakness of flat pricing.
• Vendors are raising prices. Atlassian raised cloud prices by up to 10% in October 2025, citing compute demand and new AI features. Microsoft ended volume-based enterprise cloud discounts from November 1, 2025, moving many large customers closer to list price.
• AI features are being split out of core products as add-ons, usage-based extras and premium tiers.
• Finance teams now watch cohort revenue, margin by cohort, and the mix of committed and uncommitted spend, because usage revenue is harder to forecast.

L.E.K. expected 2026 to settle around hybrid models, with buyers demanding proof of value instead of novelty.

So what for your business

If you buy software, the price on the renewal quote isn't the whole story. AI features may arrive as a separate line, and discounts you assumed would continue may not. Each of those is small. Together they can add up to a meaningful rise in your software spend.

If you build software, the same pressures are on you. AI features cost real money to run, and a flat price can quietly turn a profitable customer into an unprofitable one. Decide how you'll package AI before your costs decide for you.

What to do this month

1. Put renewals on one calendar. Look ahead six months and note the contract terms and notice periods.
2. Ask for the AI line items. Request a breakdown of what's core and what's an add-on, and what each costs.
3. Measure value per active user. If you're paying for seats that nobody uses, now is the time to find out.
4. If you build, find your cost per AI action. You can't price what you haven't measured.
5. Choose your packaging. Included, add-on or usage-based, and write down why.