Fixing AI Coding's Bottleneck
In a Nutshell
Blacksmith built a specialized CI infrastructure using gaming CPUs and gaming-optimized data centers to eliminate the validation bottleneck created by AI-assisted coding, achieving 20x growth and a $1B valuation. Their key insight is that CI workloads need single-thread performance like video games, not the multi-threaded cloud infrastructure offered by AWS/GCP. With agentic tools generating 100x more code, the company is positioning itself as the infrastructure layer that enables developers to ship at the speed of thought through autonomous QA and instant compute access.
These notes were generated by AI and may contain inaccuracies.
Every time a developer writes code, they have to run a bunch of checks to make sure that nothing is broken. That process is called CI and Blacksmith writes infrastructure to run those really, really fast. The company has grown 20x in the last year and recently raised their Series C at a billion dollar valuation led by Sequoia, raising about $66 million. The founders were outsiders who were not platform engineers and did not work in DevOps, and none of them had written a CI pipeline before.
When thinking about why they were one of the only ones to approach the problem this way, the answer includes being naive enough in the right way to not kill a good idea before it meets the market. Today with agentic coding tools, developers are writing more and more code than they ever were. But it's not very useful to write 100 times the code if you can't ship it at that speed. Now the bottleneck has become not writing the software but validating it, which usually means waiting on your CI - making sure tests are passing and having confidence that what you're shipping actually works.
The founder always wanted to come to Silicon Valley and start a company. In high school, their favorite thing to do was surf Wikipedia pages and read about technology companies and how they were founded. They went to Waterloo because they had a pretty great startup program and a lot of founders came from Waterloo, which was a very technical school. After starting work at Faire, they began thinking a lot about ideas and how to move, and Blacksmith was their second application to YC. The three co-founders were friends from undergrad, all in Toronto at the time, and all wanted to start a company. They were riffing ideas around what they thought they should work on, and all of them were engineers.
Pretty much at every company they worked at, CI universally sucked because it was super slow, but managing this piece of infrastructure was a pain across the board, and they felt they had a better way. When they initially talked about the idea for what became Blacksmith, they were initially kind of skeptical whether this could work. But over time, as they did more research, their conviction continued to grow. Their co-founders were database engineers and the founder was working on search infrastructure. From the outside, they saw that the reason managing CI info was so hard was because the shape of the workload is very different from what hyperscalers like AWS, GCP, and Azure are actually really good at. It almost felt like a mismatch that caused the pain, and they felt the real way out is building a cloud that is extremely good at running CI workloads.
It comes down to using the best tool for the job. What they started to realize was that the characteristics of CI workloads are a lot similar to desktop applications, in particular video games that you would run on your typical gaming PC, in that they don't care as much about multiple threads of execution but they care about that one single thread of execution being really, really fast. When they started off, they would rent these servers from a German data center services company that was known for hosting Minecraft servers and Call of Duty game servers.
Over time, as their scale has grown, they've started to own more and more of the stack. Now they operate at much larger levels of agreements with data center providers where they're effectively buying megawatts of power rather than individual machines.
An investor actually invested in Blacksmith twice, leading their initial round out of Y Combinator which was on a SAFE, and then a few months later led their Series A and joined the board. It was super interesting that they were using gaming CPUs because the investor had never seen anyone else try that solution for CI. It was even kind of a shock that somebody would even put a bunch of these things in one place. Usually you would have it at home for your single big box under your desk for playing games, but being like "No, we need to buy a hundred of these things and use them to compile" was where it became clear this was different.
A large part of making CI fast is skipping redundant work that has already been done before. At a high level, this is what developers refer to as caching. A big part of Blacksmith's special sauce is moving cache artifacts close to the compute nodes.
The first 200 customers were extremely hard. Initially, they actually asked their friends who were working at companies to give them a shot. They asked them for logo rights and were able to put these logos on their site and build a bit of credibility. Using that, the founder would personally call outbound CTOs and VP of Engineering personas and platform engineers and try to get them to use the product. The earliest customers they had were small companies like themselves. They cared about moving really fast and so they wanted really fast CI, and there was also very little red tape. The only real way to win over engineers is to really perform and solve a problem in a way that they probably wouldn't be able to do themselves. Engineers are very discerning and like to try a bunch of new tools, but sometimes they aren't the stickiest customers.
Without any sales team, they were able to get large five to six figure contracts because of just how easy Blacksmith made it to optimize and run your CI as fast as possible. They make it really easy - in one click, you can switch over all of your CI to them.
One of the customers that onboarded onto Blacksmith recently was Langchain. Langchain was previously running their CI on GCP. On GCP, their developers would routinely wait 15 to 20 minutes to even acquire compute to run the CI job. On Blacksmith, because they effectively expose an infinite pool of compute, this wait time went down from 15 to 20 minutes to anywhere from 10 to 15 seconds, significantly improving the velocity at which their developers could push code.
A big push for Blacksmith now is to get the word out that when any engineering team thinks about CI velocity or gaining trust in the code that they're generating, they immediately think of Blacksmith. To this effect, they are rolling out a massive marketing campaign across both San Francisco and New York. They're taking over 15 subway stations in New York and wrapping a subway train in their bright yellow color.
When the team member joined, the team was about five people running a small cluster of a couple hundred servers. Today they've seen many orders of magnitude more scale. Now they're a team of 30 people with 12 engineers. It boils down to a few key decisions they've made and continue to make as they go along. They have a lot of bespoke software, one to just manage the growth of their fleet. They have a lot of bespoke observability at every layer of the stack. Secondly, they are ruthless about keeping the surface area of the product small and manageable and only scaling it with the size of the team rather than giving into bespoke customer requirements. A big payoff of that focus is that they're able to scale to a very large number of machines and a very large number of customers with just a very small team.
One of the largest challenges is keeping up with the demand. Every time new LLM drops, they see a massive spike in CI usage as people rush to test it out and ship more code. Because it takes time to order and install machines, they have to get creative technically to support that load. This means better algorithms for placing jobs, getting more clever around how they can pack jobs more efficiently, and more efficient network topologies.
Right after their Series B announcement early in the summer, a power surge in Phoenix took out both their primary and backup cooling systems in one of the biggest data centers that they operate in Phoenix. Their customers were affected with longer startup times for their jobs during that window of outage. That underscored the importance of being geographically distributed for Blacksmith.
They want to help give developers confidence in their code changes. As part of that, one of the products they're working on is to do autonomous QA where they gather artifacts like screenshots or video snippets that induce confidence in a change. With the amount of code that's being produced, developers won't have time to go and try to understand if things are working. They're going to look for evidence and Blacksmith is going to give them confidence to merge their code faster. They want to remove that bottleneck and allow developers to basically bring shipping software as close as they can to their speed of thought.
Keep Y Combinator in your library
Save the videos and channels worth coming back to, and find them again in one place.
