Y Combinator9 min
Fixing AI Coding's Bottleneck
Conference Presentation, Fireside Chat
- Blacksmith recently secured a $66 million Series C funding round at a $1 billion valuation, led by Sequoia.
- The company has achieved a 20x performance increase in CI (Continuous Integration) speed over the last year.
- The three co-founders are software engineers without prior DevOps or CI pipeline experience.
- The company targets the bottleneck of modern software development where CI validation speed lags behind the volume of code generated by agentic AI tools.
- Blacksmith identifies that CI workloads resemble desktop gaming applications rather than standard cloud workloads, prioritizing single-thread performance over multi-threaded parallelism.
- The team originally leveraged high-performance gaming CPUs from a German data center provider known for hosting game servers.
- The company has evolved from renting individual machines to purchasing megawatts of power directly from data center providers.
- Blacksmith's technical differentiator relies on caching strategies that move artifacts close to compute nodes to minimize redundant work.
- A specific case study involving customer Langchain reduced compute wait times from 15–20 minutes on GCP to 10–15 seconds on Blacksmith.
- The platform offers a one-click migration process for existing CI pipelines.
- Go-to-market strategy relied on zero initial sales teams, utilizing founder-led outreach to CTOs and platform engineers at early-stage startups.
- The first 200 customers provided logo rights for credibility but were primarily small, agile companies.
- The company secured large five-to-six-figure contracts without a dedicated sales force by demonstrating superior performance.
- Operational metrics and infrastructure resilience challenges include:
- Headcount grew from five employees (running a few hundred servers) to 30 employees (12 engineers) managing orders of magnitude more scale.
- A summer power surge in Phoenix disabled primary and backup cooling systems, causing customer job startup delays and highlighting the need for geographic distribution.
- New LLM releases trigger massive spikes in CI usage, requiring advanced job packing algorithms and network topology optimizations to handle demand.
- Future product and strategic initiatives focus on:
- Launching a massive marketing campaign across San Francisco and New York, including 15 subway stations and a wrapped train in the brand's yellow color.
- Developing "autonomous QA" features that generate screenshots or video snippets to validate code changes and increase developer confidence.
- Maintaining a small surface area of bespoke software to manage fleet growth and avoid one-off customer requirements.
- The strategic goal is to become the primary association for developers regarding CI velocity and code trust.