Something feels off about monthly SaaS pricing in the AI era

Over the last decade, SaaS worked incredibly well.

High engineering costs, slow distribution, proprietary infra, and strong lock-in made subscriptions make sense.

But AI seems to be changing some of those assumptions.

In a lot of AI-first products today:
– Entry barriers are much lower
– Core capabilities are exposed via APIs
– Features are easier to replicate
– Switching costs are thinner
– Differentiation feels more fragile

Yet pricing hasn’t really adapted. Most tools still default to $100–$500/month subscriptions.

That makes me wonder: does the traditional SaaS model still make sense for many AI-native tools?

I’m starting to see more builders experiment with:
– Self-hosted software
– One-time purchases
– Bring-your-own-API-key models
– Full data custody
– Ownership-based pricing

Not saying SaaS is dead — but it feels like the default assumption (“everything should be a subscription”) is being challenged.

I’m personally exploring this space by building a self-hosted AI visibility tool (SEO for AI answers) that follows this ownership-first model:
👉 https://mayin.app

But I’m much more interested in the discussion than promotion.

Curious what others here think:

• Do you feel SaaS moats are eroding in AI-first categories?
• Would you personally prefer owning tools instead of renting them?
• In what kinds of products does self-hosted + one-time pricing actually make sense?

Trying to understand whether this is a real shift or just a niche pattern.