When we raised Fund I, our thesis fit on an index card: back technical founders building the infrastructure other companies quietly depend on. No consumer apps, no marketplaces, no hardware.
What held up
Infrastructure companies compound. Once a build cache, a ledger or a credentials service is wired into a customer’s stack, it rarely leaves. Our best outcomes came from products that customers stopped thinking about.
What we would change
We underestimated how quickly AI would turn into an infrastructure problem. Evaluation, observability and identity for models now look exactly like the categories we already knew how to underwrite. Fund II reflects that.