Blacksmith Software Inc. has raised $45 million in Series B funding to expand its continuous integration service that tests code in the cloud instead of on developers' computers. The investment comes as agentic AI tools like Claude Code and Codex generate increasing amounts of code that requires validation before it can ship.
Peak XV Partners led the round, with Y Combinator and GV participating. The company is now valued at $550 million.
Cloud-based testing for agentic code
Blacksmith recently launched codesmith, a cloud coding agent that developers can delegate tasks to. The tool works inside the validation loop - diagnosing integration failures, fixing them, and healing submissions to the codebase without pulling developers off their work.
Entire teams can submit code changes collaboratively. Blacksmith tests submissions as they arrive, commits changes that pass, rejects failing submissions, and sends notes explaining what went wrong.
"Writing code has gotten dramatically easier. Validating it hasn't," said co-founder and CEO Aditya "JP" Jayaprakash. "Every piece of code an agent writes still has to be built, tested and reviewed before it can ship. That validation layer is going to become increasingly important as more software is written by agents."
The company manages hundreds of thousands of compute cores and plans to expand capacity by an order of magnitude in the coming months. Jobs running on its service have grown between 5 percent on 10 percent week over week since January. More than 6,000 companies now use Blacksmith, including Supabase, Clerk, Ashby, and Mercury Systems. That's up from approximately about 800 users when the company announced its Series A last September.
What the funding goes toward
Blacksmith will direct the new funding into compute infrastructure to handle anticipated demand. As agentic development increases, the volume of code requiring automated validation testing before deployment is expected to scale accordingly.
Why this matters for IT and development teams
The core problem Blacksmith addresses is that AI tools can write code faster than humans can test it. For development teams adopting AI for IT & Development, the validation layer - not the code generation - is becoming the bottleneck. When teams delegate coding tasks to agentic tools, testing and staging infrastructure must scale to match. That shift will require developers to understand not just how to prompt AI, but how to build systems that validate AI-generated code in production. For teams exploring AI Agents & Automation, Blacksmith's approach shows one model for how those pieces can fit together.
Your membership also unlocks:
