Blacksmith Raises $45M Series B to Scale AI Code Validation

Blacksmith has raised a $45M Series B led by Peak XV Partners, with existing investors Y Combinator and GV participating again. The company announced the financing on August 12, 2026, said it had raised the round in March, and reported a $550M valuation.

The round is a bet on a specific consequence of AI-assisted development: producing code is getting cheaper while validating, testing, and merging it is getting more demanding. Blacksmith says more than 6,000 companies now run continuous integration workloads on its platform, and it plans to use the new capital primarily to expand its compute footprint.

What Happened

Blacksmith's $45M Series B brings its disclosed funding to about $58.5M. The company previously announced a $10M Series A led by GV in September 2025 and a $3.5M seed round backed by GV and Y Combinator in May 2025.

Peak XV led the Series B, with Y Combinator and GV increasing their support. Peak XV Managing Director Shailendra Singh and Partner Arnav Sahu worked with Blacksmith on the deal. That combination gives Blacksmith a new lead investor with developer-tools exposure and continued backing from the firms that financed its earlier growth.

Why This Matters

AI coding tools have increased the volume and speed of software production, but generated code still has to be tested against real systems, reviewed by humans, and cleared for production. Blacksmith says CI jobs on its platform have grown 5-10% week over week since the start of 2026. If that pace reflects a durable shift, the validation layer becomes more important precisely because code generation keeps improving.

That changes the economics of continuous integration. Slow tests, flaky checks, overloaded queues, and limited compute are no longer background annoyances when agents can generate multiple changes in the time a conventional pipeline handles one. A faster code factory without a faster quality-control line is still a factory with inventory piling up at the door.

The investment thesis is therefore broader than selling alternative GitHub Actions runners. Blacksmith is positioning itself as infrastructure for the full path between a proposed change and a trusted merge. The company is betting that engineering teams will spend more on validation, observability, and automated repair as AI expands the amount of code they must evaluate.

The Infrastructure Under the Story

Blacksmith was founded in 2024 by University of Waterloo graduates Aditya Jayaprakash, CEO, and Aditya Maru and Aayush Shah, co-CTOs. Their experience at Faire, Cockroach Labs, and Superblocks gave the team a useful starting point for a product built around distributed systems, reliability, and developer infrastructure.

The company's core product is a purpose-built CI cloud for GitHub Actions. Blacksmith operates bare-metal hardware, runs jobs in ephemeral virtual machines, and combines compute with caching, analytics, observability, and security features. Its current pricing claims up to 2x faster runtimes than GitHub-hosted runners and lower per-minute pricing for common configurations, though those performance and savings figures are company-reported.

Blacksmith says more than 6,000 companies use its CI infrastructure, including Supabase, Clerk, Ashby, and Mercury, while its company page reports more than 30,000 users and a team of 29. Those company-reported figures help explain why the financing is aimed at physical capacity rather than a broad marketing campaign.

From Faster CI to Automated Validation

The next layer is [code]smith, Blacksmith's cloud software-engineering agent. The company's product documentation describes [code]smith as capable of building features, fixing bugs, working across repositories, and diagnosing or automatically addressing CI failures. It can be used through Slack, GitHub, or Blacksmith's web interface, keeping the coding agent close to the infrastructure that evaluates its work.

Blacksmith is also developing [code]smith QA, which it says will autonomously test changes before they are merged. The distinction matters because AI coding products are increasingly competing on how much work they can generate, while engineering leaders still carry the risk of what reaches production. Blacksmith's opportunity is to make validation an integrated system instead of a trail of disconnected checks and late-night judgment calls.

Where the $45M Goes

Blacksmith says the primary use of proceeds is compute. The company reports that it already manages hundreds of thousands of cores and plans to expand that footprint tenfold in the coming months. It is also hiring across New York City and San Francisco for engineering, product, operations, support, and go-to-market roles.

That capital plan is unusually direct for a software company because Blacksmith owns more of the physical stack than a conventional SaaS vendor. More customers and more CI jobs require actual machines, storage, networking, and regional capacity. The financing gives the company room to build ahead of demand rather than wait for queues and incidents to dictate the next expansion.

What This Signals

Blacksmith's Series B suggests that investors see code validation as a distinct AI infrastructure category, not merely a feature inside an existing developer platform. Peak XV's lead investment, combined with renewed support from GV and Y Combinator, places a sizable wager on the idea that faster code generation increases the value of testing and merge infrastructure.

The harder question is whether Blacksmith can turn a strong CI product into the broader control layer it describes. The company has credible infrastructure roots, fast reported growth, and a product path that connects agents to the systems judging their output. If Blacksmith executes, the most valuable infrastructure in the AI coding economy may not be the system producing more code, but the one proving which code is ready to ship.