---
title: "Blacksmith Raises $66M Series C at $1B as AI Strains Testing"
description: "Sequoia led Blacksmith's $66M Series C. A Clerk case study, usage-based economics and competing CI investments explain the bet on testing AI-generated code."
date: 2026-10-22T19:13:54.267Z
source: https://venturecapitaltracker.com/blacksmith-66m-series-c-1b-valuation
---

# Blacksmith Raises $66M Series C at $1B as AI Strains Testing

> Sequoia led Blacksmith's $66M Series C. A Clerk case study, usage-based economics and competing CI investments explain the bet on testing AI-generated code.

Blacksmith has raised **$66 million in Series C funding at a $1 billion valuation**, as AI coding tools increase demand for the infrastructure that checks software before it reaches customers.

[Sequoia Capital](/fund/sequoia) led the round announced October 21, with Premji Invest, Peak XV Partners, GV and Y Combinator participating. The [company-issued release](https://www.prnewswire.com/news-releases/blacksmith-raises-sequoia-led-series-c-at-1-billion-valuation-to-scale-code-validation-302915140.html) puts total funding at approximately $125 million. It does not specify whether the valuation is before or after the new capital.

The distinctive bet is on a bottleneck: generating more code makes dependable testing more valuable.

## What changes for a software team

Continuous integration, or CI, automatically builds and tests proposed code changes. An engineering team pays Blacksmith to run those jobs on its computing infrastructure.

The alternatives are GitHub-hosted machines or a fleet the team manages itself. Blacksmith keeps the GitHub Actions workflow while taking over the machines underneath it. That can reduce migration work, although security reviews, credentials and workload compatibility still need attention.

A [Blacksmith-published Clerk case study](https://www.blacksmith.sh/customer-stories/clerk) makes the problem concrete. Clerk tests its developer software across several application frameworks and versions. Its browser tests were exhausting the available memory, creating timeouts and repeated runs. The case study reports that moving to Blacksmith reduced one test sequence from 9 minutes 43 seconds to 5 minutes 56 seconds.

The account is supplier-hosted customer testimony, rather than an independent benchmark. Its practical value is the workflow: fewer infrastructure-related failures let engineers spend more time fixing software and less time rerunning checks.

## Who pays, and what the bill measures

Blacksmith's [pricing page](https://www.blacksmith.sh/pricing), checked October 22, lists usage-based compute and separately priced extras such as persistent disks and Docker caching. Customers therefore pay for infrastructure consumption rather than simply buying a seat for every developer.

A useful buying test is **cost per completed build**, including reruns, storage and engineering time. A cheaper minute is less valuable if the same job takes longer or fails repeatedly.

For illustration, halving a ten-minute job produces five billed minutes at an unchanged rate. That calculation explains why runtime matters; it is not a guaranteed customer saving.

## Why Sequoia invested now

In its [October 21 investment memo](https://sequoiacap.com/article/partnering-with-blacksmith), Sequoia argues that coding agents produce repeated attempts that each need validation. The firm highlights the founders' systems experience and says Blacksmith's largest customers now spend more than $1 million annually.

There is evidence for the underlying pressure beyond Blacksmith's own growth figures. In a [September 14 engineering account](https://claude.com/resources/articles/agentic-coding-is-straining-ci-heres-how-we-scaled-test-impact-analysis-at-anthropic), Anthropic reported a 25-fold increase in its CI jobs over six months. That documents one company's experience; it does not establish that every software team will follow the same trajectory.

Blacksmith [reports](https://www.blacksmith.sh/blog/series-c) that weekly jobs rose from 5 million in January to more than 90 million, while annual recurring revenue grew 17-fold. The revenue base is undisclosed, preventing a meaningful valuation-to-revenue calculation.

Its earlier [Series B](/2026-blacksmith-45m-series-b-peak-xv-ci-ai) was announced in August but closed in March. The two announcement dates should not be mistaken for two closings only two months apart.

## A funding wave around builds and tests

Two close competitors show this is a defined infrastructure market:

| Company | Verified financing | Overlap |
| --- | --- | --- |
| Namespace | [$42M Series B, October 5](https://namespace.so/blog/series-b) | Build and test environments, GitHub runners and its own server infrastructure |
| Depot | [$10M Series A, March 10](https://depot.dev/blog/depot-raises-series-a) | Build acceleration, GitHub runners and a programmable CI platform |

[VCT previously covered Namespace's round](/2026-namespace-clockwork-arboretum-200m-funding-roundup). These are closer comparisons than an AI app builder: all three sell capacity or tooling for delivering software.

## The capacity bet behind the valuation

Blacksmith says it manages just under 10 megawatts of compute and plans to expand that footprint fivefold in 2027. Capacity is an investment ahead of demand, with costs that continue when machines sit idle.

The company has [explained its model](https://www.blacksmith.sh/blog/the-economics-of-operating-a-ci-cloud) as pooling customers' uneven workloads to improve hardware utilization. Its historical examples do not establish today's margins. Faster AI development could bring more demand, but price competition and excess capacity could absorb the benefit.

The decisive measure is profitable, repeat usage after infrastructure costs. Blacksmith's challenge is to make testing cheaper for customers while earning an attractive return on the fleet it is building.

**By:** [Venture Capital Tracker](https://venturecapitaltracker.com/editorial-policy)

**Editorial note:** AI tools assisted with research, structure, or drafting. Venture Capital Tracker retains human editorial responsibility for factual accuracy, relevance, and source quality before publication.