Blacksmith CI Platform Cuts Build Times from Minutes to Seconds

Blacksmith is a company that provides infrastructure to run Continuous Integration (CI) checks at extremely high speeds. CI is the process developers use to run checks and ensure code changes don't break existing functionality. The company has experienced significant growth, expanding 20x in the last year and recently raising a $66 million Series C round at a billion-dollar valuation, led by Sequoia.

The Bottleneck of Modern Development

With the rise of agentic coding tools, developers are generating more code than ever before. However, this increased code generation is only useful if it can be shipped quickly. The bottleneck has shifted from writing software to validating it, which often means waiting for CI processes to complete. This involves ensuring tests pass and that the shipped code functions as intended.

Origins and Unique Approach

The founders of Blacksmith, who were friends from undergrad, were all engineers who experienced universally slow and difficult-to-manage CI infrastructure at previous companies. They recognized a fundamental mismatch: hyperscalers like AWS, GCP, and Azure are optimized for workloads different from CI. CI workloads, they discovered, share characteristics with desktop applications, particularly video games, in that they prioritize single-thread execution speed over multi-threaded performance.

This insight led them to a unique solution: building a cloud specifically optimized for CI workloads. Initially, they rented servers from a German data center company known for hosting Minecraft and Call of Duty game servers. As they scaled, they began to own more of the stack, now operating at much larger agreements with data center providers, effectively buying megawatts of power rather than individual machines. This innovative use of gaming CPUs for CI was a novel approach that surprised even early investors.

Blacksmith's "Special Sauce"

A key factor in Blacksmith's speed is its ability to skip redundant work through efficient caching. A significant part of their "special sauce" involves moving cache artifacts physically close to the compute nodes, minimizing latency.

Customer Acquisition and Growth

Acquiring the first 200 customers was challenging. Initially, the team leveraged personal connections, asking friends at companies to try their product and grant logo rights to build credibility. This allowed them to cold-call CTOs, VPs of engineering, and platform engineers. Their earliest customers were small companies, similar to Blacksmith itself, who prioritized speed and had less bureaucratic red tape. The company found that the only way to win over discerning engineers was through superior performance and by solving problems in ways customers couldn't achieve themselves.

Blacksmith has secured large, five-to-six-figure contracts without a dedicated sales team, primarily due to the ease with which their platform optimizes and accelerates CI. They offer a one-click solution to switch CI operations to their service.

Impact on Customers: The Langchain Example

Langchain, a recent Blacksmith customer, previously ran its CI on GCP. Developers routinely faced 15 to 20-minute wait times just to acquire compute resources for CI jobs. With Blacksmith, which provides an effectively infinite pool of compute, this wait time plummeted to 10 to 15 seconds, drastically improving developer velocity.

Scaling and Future Vision

Blacksmith is now focusing on a massive marketing campaign in San Francisco and New York to ensure that any engineering team considering CI velocity or code trust immediately thinks of them. This includes taking over 15 subway stations in New York and wrapping a subway train in their bright yellow branding.

The team has grown from five people running a small cluster of a few hundred servers to 30 people, including 12 engineers, managing orders of magnitude more scale. This growth is attributed to key decisions: - Bespoke software: Custom solutions for fleet management and observability at every layer of the stack. - Small product surface area: Ruthlessly keeping the product manageable and scaling it only with team size, rather than acceding to every bespoke customer requirement. This focus allows them to scale to a large number of machines and customers with a small team.

Challenges and Future Developments

One of Blacksmith's biggest challenges is keeping up with demand. New Large Language Models (LLMs) often trigger massive spikes in CI usage as developers rush to test and ship code. Since ordering and installing machines takes time, Blacksmith employs technical creativity, including better job placement algorithms, more efficient job packing, and optimized network topologies, to handle these loads.

A power surge in Phoenix that took out both primary and backup cooling systems in one of their largest data centers underscored the importance of geographical distribution.

Blacksmith's future plans include developing autonomous QA products. These tools will gather artifacts like screenshots or video snippets to provide developers with confidence in their code changes. The goal is to remove the CI bottleneck, allowing developers to ship software as quickly as they can think, providing evidence of functionality rather than requiring manual verification.