Blacksmith Raises Funding for AI Code Validation Layer

Blacksmith, a San Francisco startup that runs continuous integration workloads on purpose-built hardware, has raised a $45 million Series B led by Peak XV Partners at a $550 million valuation, the company announced on August 12, 2026. Existing investors Y Combinator and GV also participated, which co-founder and CEO Aditya “JP” Jayaprakash described as both firms “tripling down.”

The company’s wager is that the bottleneck in software development is moving. AI coding tools have made producing code cheap and fast; every line an agent writes still has to be built, tested, and reviewed before it ships. Blacksmith sells the infrastructure for that second half of the pipeline, and the usage numbers in its announcement suggest the shift is already measurable: weekly CI jobs on the platform have grown between 5% and 10% week over week since the start of 2026, and its customer count has climbed from roughly 800 companies to more than 6,000, including Supabase, Clerk, Ashby, and Mercury.

“Writing code has gotten dramatically easier. Validating it hasn’t,” Jayaprakash said in the company’s funding announcement. “We’re seeing teams adopt coding agents, generate several times more pull requests, and suddenly CI becomes a bottleneck.”

What Blacksmith Actually Sells

Blacksmith’s core product is a CI cloud for teams running GitHub Actions. Rather than renting general-purpose instances from hyperscalers, it runs workloads on bare-metal, gaming-grade CPUs with caching and storage tuned for CI, and migration takes a one-line change to a workflow file. The company claims the setup runs jobs at up to twice the speed of GitHub’s own runners while costing about 60% less, with cache downloads four times faster and Docker builds up to 40 times faster.

The funding extends that foundation into what the company calls a validation platform. It recently launched codesmith, a cloud coding agent that developers can delegate tasks to, and that works inside the validation loop by diagnosing CI failures, autofixing them, and keeping pull requests green. Codesmith is also the base for codesmith QA, a planned product that would autonomously test changes before they merge. The direction puts Blacksmith in the position of selling picks to both sides of the AI-coding trade: the teams generating more code with agents, and the agents themselves as they take on more of the merge process.

The Round and the Run-Up

The company was founded in 2024 by Jayaprakash, Aayush Shah, and Aditya Maru, who met at the University of Waterloo and went on to build large-scale distributed systems at Faire and Cockroach Labs. Blacksmith launched out of Y Combinator’s Winter 2024 batch, raised a $3.5 million seed led by GV and Y Combinator in May 2025, and followed it with a $10 million Series A led by GV announced on September 18, 2025. That round, which the company said closed in 14 days, came as Blacksmith passed $1 million in annual recurring revenue with about 800 customers.

One detail in the company’s own post stands out for readers tracking venture timing: Blacksmith actually raised the Series B in March 2026 and is only now announcing it, five months later. The growth figures it is disclosing therefore describe a business that already had this capital in the bank. Peak XV, the firm formerly known as Sequoia Capital India and SEA, manages more than $10 billion across 16 funds, and Jayaprakash noted the firm’s developer-tools portfolio includes Supabase and PostHog. Supabase is itself one of Blacksmith’s named customers, and its own growth has tracked the same AI-coding wave, as Unite.AI covered when Supabase raised $500 million at a $10.5 billion valuation in June 2026.

Where the Money Goes

Most of the new capital is earmarked for compute. Blacksmith says it manages on the order of hundreds of thousands of cores and plans to grow that footprint by an order of magnitude in the coming months to stay ahead of demand. The company is also hiring engineers in New York and San Francisco.

The allocation is a straightforward read on the business model: Blacksmith owns its performance advantage because it owns its hardware stack, and every multiple of AI-generated code volume translates directly into demand for cores that can validate it. The next observable milestone is the launch of codesmith QA, which the company says will autonomously test changes before they merge.