Blacksmith raises $66 million at a $1 billion valuation to scale code testing

Blacksmith has secured a $66 million Series C led by Sequoia Capital, valuing the code-testing startup at $1 billion. The financing, reported October 21, 2026, will largely pay for more computing power—a costly requirement for a business whose chief executive says AI coding tools are shifting developers’ bottleneck from writing software to checking it.

The round follows a $45 million Series B led by Peak XV Partners about two months earlier, at a $550 million valuation. Blacksmith’s total funding now stands at about $124.5 million, Fortune reported. The latest investment puts a substantially higher price on the company after a short interval between rounds.

Premji Invest, Peak XV Partners, Google Ventures and Y Combinator also participated, according to Inc42. Alongside the financing, Blacksmith said it would expand its technical engineering team as well as its computing infrastructure.

The work after the code is written

Founded in 2024, Blacksmith sells continuous integration, or CI: the automated builds and tests that code must pass before joining an existing codebase and being deployed. It supplies the infrastructure that runs those checks, rather than a tool whose main job is to write the code.

Most companies run those jobs on servers rented from Amazon Web Services, Google Cloud or Microsoft Azure. Blacksmith instead buys its own machines and installs them in data centers. CEO and cofounder Aditya Jayaprakash told Fortune that the company owns tens of thousands of machines.

The company also writes the software that runs that hardware. Its systems save repetitive setup work that each test run would otherwise have to do again. That makes the business more than a supplier of server capacity: it also manages how recurring testing work uses those machines.

Revenue rises alongside testing demand

Jayaprakash said annual recurring revenue had grown 17-fold since the start of the year and was now in the tens of millions of dollars. Blacksmith said its largest customers spend more than $1 million annually. Customers include Granola, Handshake, Supabase and LangChain.

The CEO’s explanation for that growth is a shift in where developers spend their time. Tools such as Cursor, OpenAI’s Codex and Anthropic’s Claude Code have made writing code easier, he told Fortune. Validating the resulting output has become the constraint. That is Blacksmith’s argument for selling more testing capacity as coding models improve.

More machines, less predictable demand

Blacksmith says it manages nearly 10 megawatts of computing capacity across data centers in the United States and Europe. It plans to increase capacity fivefold in 2027, Inc42 reported, citing the company’s statement. Separately, Jayaprakash told Fortune that Blacksmith plans to grow from about 40 employees to around 75 by year-end.

Owning hardware gives Blacksmith control over performance, Jayaprakash said, but also exposes it to shortages of computing equipment and memory. He described capacity planning as one of the business’s hardest tasks because demand jumps when stronger models arrive. Buying machines therefore requires planning for workloads that can change abruptly.

There is also competition from companies closer to the code-writing process. Coding agents such as Codex and Claude Code can already run a project’s tests as they work, while Microsoft-owned GitHub sells hosted machines for CI jobs. Jayaprakash argues Blacksmith complements model providers rather than being displaced by them: better models generate more code, and more code means more checks.