代码测试初创公司 Blacksmith 完成 6600 万美元 C 轮融资,估值达 10 亿美元
Code testing startup Blacksmith becomes a unicorn after raising $66M in Series C funding - SiliconANGLE
持续集成云服务商 Blacksmith 完成 6600 万美元 C 轮融资,由 Sequoia Capital 领投,Peak XV Partners、Y Combinator 和 GV 参投,公司估值达 10 亿美元,累计融资 1.25 亿美元。该公司自建服务器提供 CI 云服务,目前有超过 1 万家企业使用,每周代码检查量从 2026 年初的约 500 万次增至约 9000 万次,年度经常性收入同期增长 17 倍。CEO Aditya Jayaprakash 表示,AI 编码工具让写代码变得容易,瓶颈已转向代码验证,本轮资金将主要用于购买服务器扩建全球 CI 云,并将团队从约 40 人扩至年底 75 人。
正文
Blacksmith Software Inc., a continuous integration cloud startup that helps companies to test and verify new code before they ship it into production, says it’s now valued at $1 billion after closing on a $66 million Series C round of funding today.
The round was led by Sequoia Capital and saw participation from Peak XV Partners, Y Combinator and GV, and brings Blacksmith’s total amount raised so far to $125 million. It comes just two months after it raised $45 million in a Series B round led by Peak XV. Prior to that, it raised $10 million in September 2025.
Blacksmith came into being in 2024. It’s led by its co-founder and Chief Executive Aditya Jayaprakash and it’s offering a fairly novel service in the shape of a “CI cloud,” with the CI standing for “continuous integration. That’s the process of automating builds and testing software before it’s merged into an existing codebase and deployed in production.
Before Blacksmith was a thing, most companies would handle CI themselves, running those jobs in the cloud on servers hosted by Amazon Web Services Inc., Microsoft Azure, Google Cloud or someone else. But nowadays, they can go with Blacksmith instead, which has bought its own servers and built a dedicated CI cloud that spans multiple global locations.
The startup writes its own software to run those CI jobs. One of its main advantages is that it can automate much of the once-manual set up work required before each new piece of code can be tested.
In a blog post announcing today’s round, Jayaprakash explained that before the advent of artificial intelligence coding assistants, most developers used to spend approximately half their time writing code and the other half waiting for it to build and pass various tests. But the emergence of AI coding tools like Cursor, Anthropic PBC’s Claude and OpenAI Group PBC’s Codex, the process of creating new code has been made “extremely easy,” he said.
“Now the bottleneck has shifted to the second piece, which is: how do I actually validate all this code?” Jayaprakash told Fortune in an interview.
This phenomenon is best evidenced by Anthropic, which said last month that its code validation workload has grown more than 25-fold in the last six months as a result of Claude taking over most of the coding duties from its human developers. Blacksmith hasn’t quite managed to fix this bottleneck, but its CI cloud does ease the pressure somewhat. “Companies bring us their CI workloads and we run them faster than they could anywhere else, with zero operational overhead,” Jayaprakash explained.
Part of the problem with CI work is its unpredictable nature, with demand jumping from zero to tens of thousands of central processing units and then dropping to nothing again when the testing gets done and there’s nothing waiting in the pipeline. Jayaprakash said his company has focused on building its cloud infrastructure in a way that absorbs this unpredictability.
Because the company owns its own hardware rather than renting it, it has more control over its performance, enabling it to run through customer’s CI workloads faster. However, Jayaprakash admitted that this has caused a problem for the startup, as it’s now exposed to the global shortage of components such as memory.
Capacity planning has become one of the most challenging aspects of running the company, he said, as demand often jumps whenever a new frontier model is launched. For instance, when Anthropic debuted Claude Fable 5 in July, its CI workload volume shot up more than 40% in a single day.
Though Blacksmith is helping companies to deal with the deluge of AI-generated code, there is a risk that AI model vendors could become rivals, assuming code validation duties themselves.
That said, there can be no denying Blacksmith’s traction so far. The startup says more than 10,000 companies now use its platform for their CI work, up from around 5,000 in August. Some of its best known customers include LangChain, Supabase and Granola.
In early 2026, it was running about 5 million code checks per week, and that number has now grown to about 90 million, with annual recurring revenue up 17 times over the same period. Jayaprakash said the company’s revenue is now in the “tens of millions of dollars,” and some of its largest clients spend more than $1 million annually.
To cater to its growing customer base, Blacksmith said it will use most of the money from today’s round on buying more servers to build out its global CI cloud. It also wants to expand its team of about 40 employees to 75 by the end of the year.
Image: SiliconANGLE/Gemini