Blacksmith, the San Francisco infrastructure startup that runs builds and tests for other companies' code, announced on October 21, 2026 that it has raised a $66 million Series C led by Sequoia Capital at a $1 billion valuation. The round, which the company disclosed in a press release, included new investor Premji Invest alongside existing backers Peak XV Partners, GV and Y Combinator, and brings Blacksmith's total funding to $125 million.
The valuation makes Blacksmith one of the clearest bets yet on a thesis that has been building all year: as AI coding agents write more of the world's software, the bottleneck in shipping it has moved from writing code to checking it.
The bottleneck moved
Every change to a software project has to be built and tested before it can be merged - a step the industry calls continuous integration, or CI. For years that was a background chore: developers pushed a handful of changes a day, and general-purpose cloud servers chewed through the queue. Blacksmith's pitch is that this balance has broken. In the company's telling, developers who used to open three or four pull requests a day now open orders of magnitude more, because AI coding assistants generate the code for them - and the CI infrastructure underneath was never designed for that volume.
The company points to a striking data point from a customer-adjacent giant: Anthropic reported that its own CI volume grew 25-fold in six months as Claude took over most of its code output. "Since Claude Code took off in March, almost every conversation we've had with a large engineering team has started the same way," Blacksmith co-founder and CEO Aditya "JP" Jayaprakash said in the announcement. "Developers who used to push three or four pull requests a day are now pushing orders of magnitude more and their CI infrastructure can't keep up."
The numbers behind the round
Blacksmith says its weekly CI job volume has grown from 5 million in January to more than 90 million now, and that annual recurring revenue has grown 17x over the same period. More than 10,000 startups and enterprise companies - the company names Ashby, Granola, Handshake and LangChain - run their builds and tests on the platform. As with all private-company figures, these are the company's own numbers, offered without audited detail, but they were substantial enough to bring Sequoia in as lead investor at a billion-dollar mark.
One named customer put a figure on the pitch. "Blacksmith has significantly reduced the burden of managing our CI runners while cutting our runner costs by roughly 40%, and that doesn't even account for the engineering time we've saved," said Joaquin Borggio, an infrastructure engineer at LangChain, in the company's release. Blacksmith claims its customers see CI jobs complete twice as fast at 60% lower cost on average, with a one-line change to a workflow file as the entire migration.
Purpose-built for spiky demand
CI is an awkward workload for conventional cloud setups: demand swings from zero to tens of thousands of CPUs and back, sometimes within the hour, so teams constantly overpay for idle capacity or under-provision and wait. Blacksmith says it designed every layer of its stack - down to newer-generation CPUs and its own compute footprint - to absorb those spikes routinely. The company says it now manages nearly 10 megawatts of compute across data centers in the United States and Europe and plans to grow that footprint at least five-fold in 2027.
Blacksmith was founded in 2024 by Jayaprakash, Aayush Shah and Aditya Maru, who met as undergraduates at the University of Waterloo. Shah and Maru were early engineers at Cockroach Labs, and Jayaprakash was an engineer at Faire. The company says the new capital will go toward growing the team - it is hiring in New York and San Francisco - and expanding its compute footprint.
Why it matters
Two years into the AI-coding wave, the picks-and-shovels layer of the boom is starting to look as investable as the models themselves. The first wave of infrastructure spending went to training and serving models; the second is going to the unglamorous plumbing that an agent-heavy development workflow stresses - and CI is near the top of that list. If AI agents keep multiplying the volume of code that needs building and testing, the companies selling validation capacity become a toll booth on every software team's release pipeline.
Sequoia's check is also a signal about where top-tier firms see durable revenue in the agent era. Model providers compete on capability that converges quickly; a CI platform's economics - usage-based, deeply embedded in a customer's daily workflow, painful to rip out - look more like the cloud infrastructure franchises venture firms have historically prized. Premji Invest joining the cap table adds a large growth-stage backer to that bet.
The open questions are the usual ones for a newly minted unicorn: whether 17x ARR growth survives contact with a normalized market, whether the big clouds compete harder on price, and whether AI-generated code volumes keep climbing or plateau as teams find their rhythm. Blacksmith's answer, for now, is a bigger compute footprint and a billion-dollar label - and a very long queue of builds to run.