Blitzy usage-based pricing model
Blitzy
Usage based pricing pushes Blitzy out of the normal developer seat budget and into the budget for software work completed. A customer is not mainly paying for how many engineers log in. It is paying for how much codebase gets mapped, how many code changes get generated, and how many repositories move through the system. That fits a product that works asynchronously on large legacy systems and expands by taking on more backlog, not by adding more editor users.
-
A seat model works best for editor tools used one developer at a time. Blitzy starts with codebase ingestion, builds a technical spec and action plan, then delivers pull requests, so the natural meter is work output. That is why pricing is tied to lines of code onboarded and generated, alongside enterprise contracts that start at $500K and can reach $10M per year.
-
This also changes expansion math. Cursor sells team plans at $40 per user per month, with extra usage layered on top, and GitHub Copilot Business and Enterprise include pooled per user AI credits before overage billing. Blitzy can grow inside an account even if the number of active engineers stays flat, as more repositories and modernization projects move onto the platform.
-
The tradeoff is that revenue and cost both track output. Blitzy publishes a $0.20 rate per generated line and says it uses model routing to protect margins, sending simpler subtasks to cheaper models and harder reasoning work to more expensive ones. That makes gross margin discipline a core product capability, not just a finance target.
The category is moving toward hybrid pricing, but the strategic split will remain clear. Seat pricing fits copilots that help humans code faster inside the editor. Work based pricing fits systems that absorb whole chunks of backlog and ship production changes with limited human touch. If Blitzy keeps proving quality on large codebases, more enterprise spend will shift from software seats to program level delivery budgets.