Infrastructure investing at Seed, by operators who built the systems that needed it.
AI applications are proliferating. But the models themselves are commoditizing. The durable value in the AI stack will compound at the infrastructure layer — the databases, orchestration systems, feature stores, and model-serving frameworks that sit beneath every application.
Flintrock invests in that layer. At Seed. Before the winner is obvious.
The data infrastructure companies that defined the last decade — Kafka, Spark, Airflow, PostgreSQL-as-cloud — were investable at inception, when they were open-source projects with small communities and no enterprise sales motion. The AI infrastructure layer is in that moment now.
Seed is when the technical conviction matters more than the revenue model. We can evaluate technical architecture, community trajectory, and ecosystem fit. We write checks when others wait for traction metrics.
A model is only as good as its context window — and its context window is only as good as what the retrieval layer surfaces. Qdrant and Weaviate are building that retrieval layer. The features that go into a model's input are only as good as the feature store's freshness and reliability — Chalk solves that. The predictions that come out are only as useful as the serving infrastructure that routes them to production — BentoML makes that tractable. The pipelines that connect all of it need orchestration — Prefect keeps them observable and recoverable. And the analytical queries that close the loop — DuckDB makes them run anywhere.
Our portfolio is not a collection of bets. It is a map of the stack.
James spent nine years at hyperscale building distributed query systems. Yuki built ML platform infrastructure and then led engineering at a data observability startup. We evaluate infrastructure investments the way a senior engineer evaluates a technical design: architecture, failure modes, adoption trajectory, API ergonomics, ecosystem positioning.
We do not outsource that judgment to market signals. We form it from first principles.
Seattle is where major cloud platforms, incumbent hyperscale infrastructure, and a generation of systems-engineering talent were built. The engineering culture here is systems-first: performance, correctness, and operational reliability before product decoration. Flintrock is native to that culture and sources from it — from open-source communities, engineering blogs, and conversations with principal and staff engineers at the hyperscale companies that call Seattle home.
We are not the only infrastructure investors. But we are operators who spent our careers here, in the systems that define this landscape.
We write Seed checks. Reach us at [email protected] or call +1 (206) 555-0684.
[email protected]