Blacksmith Hits $550M Valuation on AI Code Cleanup

Blacksmith closed a $45M Series B led by Peak XV, with GV and Y Combinator joining, pushing its valuation to $550M, per TechCrunch. That's up from $60M at its Series A last September, a 10x jump in eleven months. Total funding now sits at $58.5M.

The growth numbers back up the valuation. Customers went from 700 to over 5,000 in under a year, including Mercury, Supabase, Clerk, Ashby, and Expensify, with some spending more than $1M annually. Revenue climbed from $10M ARR to "tens of millions," and the team grew from 10 employees to about 30.

CEO Aditya Jayaprakash's pitch is that AI agents now write more code than humans do, so the constraint shifted from writing to checking. Blacksmith's answer is Codesmith, an agent that auto-fixes failed CI checks instead of just flagging them.

That puts Blacksmith in a crowded lane. GitHub Actions and Cursor Automations both touch this territory, and Codex and Claude Code already ship their own validation logic. Blacksmith is betting that fixing broken checks is different enough from writing code, or from running CI infrastructure, that a standalone company can own it.

The open question

The TechCrunch piece doesn't say whether Codesmith's fixes get reviewed by a human before merge, or what happens when its auto-fix is wrong. For a company selling into teams already spending $1M a year, that gap matters more as usage scales. A 5,000-customer base moving that fast is also a base a platform vendor could route around, if GitHub or Anthropic decide validation belongs in the base product rather than a $45M-funded layer on top of it.