Your AI IDE changed its billing, its price, and its owner — and you didn't get a vote

In one quarter, the AI IDE a lot of people build inside changed three things about itself without asking anyone who uses it. Windsurf retired its credit system and replaced it with daily and weekly quotas on March 19, 2026. Pro went from $15 to $20. A new $200 Max tier appeared. And the whole thing now sits under Cognition — the Devin company — after a 2025 acquisition.

None of those were bugs. They're the normal lifecycle of a tool you rent. The point isn't that Windsurf did anything wrong. The point is that when you build inside a tool, its price, its billing model, and its owner are all variables you absorb rather than choose — and right now all three are moving across the whole category at once.

Three things changed, and you didn't get a vote

Break down what actually shifted, because each one lands differently:

• The billing model changed. Credits — a pool you understood — became daily/weekly quotas, with overages now at API rates instead of credit packs. The mental model you'd built for "how much will this cost me" was reset.
• The price changed. $15 to $20 is a 33% Pro increase, plus a new tier 10× above it. Not catastrophic. But it's a number you don't control, on a tool your workflow now depends on.
• The owner changed. Cognition folding Windsurf in with Devin means roadmap, priorities, and terms now answer to an acquirer's strategy, not the thing you signed up for.

Any one of these is survivable. The pattern behind them is the thing to notice.

This is the category, not one tool

Windsurf is just the clearest recent receipt. Zoom out and the same forces are everywhere in vibe-coding: credit pools and token quotas that get re-cut, prices that drift up, and a consolidation wave pulling the field into a few large platforms. The tools are racing, raising, and merging. That's healthy for the market and genuinely exciting to watch — and it means the terms under your project are the least stable part of your stack.

If your app's foundation lives inside one of these tools, you're not just betting the tool is good. You're betting its price, its billing, and its corporate parent stay convenient for as long as your project lives. That's a lot of bets to have no vote in.

What you own versus what you rent

Here's the line that actually matters, and it's a clean one:

When the rented column changes — and it will keep changing — the owned column doesn't care. Code on your own filesystem builds the same whether Windsurf is $15 or $200, credits or quotas, independent or acquired. An agent reading your repo is swappable; if one tool reprices or gets bought, you point a different agent at the same code and keep going.

The hedge against a volatile tool market isn't picking the one tool that won't change. None of them will promise you that. The hedge is making sure the tool is the replaceable part — that your code, your conventions, and your deploy path live in a repo you control, so swapping tools is a config change, not a rebuild.

The math owning gets you

A subscription is a bet that renews monthly. The tool stays good, the price stays put, the company stays independent — pay again, repeat. The moment any of those breaks, the bet was on the wrong thing.

Buying code you own is a different shape. The OTF kit you bought once still builds if Windsurf tripled its price this morning, because it doesn't run on Windsurf — it runs in your repo, and your agent of choice reads it. The tool can consolidate, reprice, or rebrand. Your codebase doesn't get a memo, because it doesn't work for them.

Tools will keep getting acquired and keep changing their billing — that's the 2026 AI-IDE market, and it's not slowing down. The only part of your stack you can keep stable through all of it is the part you actually own. Build there, and the shakeup is something you read about instead of something that happens to you.