Peak XV Partners led a $45 million Series B into Blacksmith, a startup that helps engineering teams build, test, and verify code before deployment, valuing the company at $550 million. That values Blacksmith at close to nine times what it was worth less than a year ago, when a $10 million Series A priced it at $60 million. GV and Y Combinator, both prior backers, returned for the new round, bringing Blacksmith’s total funding to $58.5 million.
The repricing is a bet on a specific idea: checking AI-generated code has become the harder engineering problem, not producing it. Coding tools such as Cursor, OpenAI’s Codex, and Anthropic’s Claude Code have made writing software dramatically faster. Aditya Jayaprakash, Blacksmith’s co-founder and chief executive, said the volume of output has outrun teams’ capacity to review it. “Validating code is still a bottleneck, and it’s an even bigger bottleneck because people are writing even more,” he said.
Founded in 2024, Blacksmith started by selling cloud infrastructure for continuous integration, the builds and automated tests that run before code ships. It later added Codesmith, an AI agent that can automatically repair code that fails those checks, moving the product from flagging problems toward fixing them.
Customer growth has followed that expansion. Blacksmith counts Mercury, Supabase, Clerk, Ashby, and Expensify among its paying accounts, and puts its total customer count above 5,000, roughly seven times the level it reported a year earlier.
Revenue has scaled too, though the specifics are fuzzy. Jayaprakash said Blacksmith reached a $10 million annualized run rate on a ten-person team, and has since tripled headcount to roughly 30 while pushing revenue into the “tens of millions of dollars.” He would not share a current annualized figure, saying only that several of Blacksmith’s biggest accounts now pay upward of $1 million a year. Those figures come from Jayaprakash rather than any audited disclosure, which is typical for a funding announcement but worth flagging given how much of the new valuation rests on growth claims rather than public financials.
The category Blacksmith is selling into is not empty. GitHub Actions and Cursor Automations compete directly, as do the validation checks now built into OpenAI’s Codex and Anthropic’s Claude Code themselves, plus code-testing services from Amazon Web Services, Microsoft Azure, and Google Cloud. Jayaprakash said Blacksmith is competing on speed and price rather than trying to out-feature the cloud giants.
That competitive field is the real test of whether $550 million holds up. The multiple works only if Blacksmith keeps converting new accounts faster than GitHub, the major clouds, and the coding agents fold validation into their existing products for free. It also depends on Codesmith becoming a genuine revenue driver rather than a retention feature bundled into the CI business. Frontier labs are actively working to make agent-written code more reliable on the first pass. If they succeed quickly, the exact bottleneck Jayaprakash describes narrows before Blacksmith’s next round arrives.
For engineering leaders shopping for CI and code-validation vendors this quarter, the valuation itself is not the useful data point. What matters is that a company built specifically to catch bad AI-generated code is growing fast enough to raise at this price, a market signal that the gap between how much code agents produce and how much gets properly checked is wider than most budgets currently account for.
TechCrunch first reported Blacksmith’s funding round and valuation on August 12, 2026.
