The Case for Building, Not Just Betting

There is a version of the venture capital business that is purely financial: you study markets, identify promising teams, deploy capital, and wait for outcomes. This model has produced extraordinary returns across multiple technology cycles. We don't dismiss it. But we don't think it is the right model for investing in deeptech at the current moment, and it is not how we operate at Nexum.

The argument for being a builder as well as a backer starts with information. If you only invest in a domain, you develop pattern recognition over time — you learn to read founding teams, you develop instincts about market timing, and you refine your model for evaluating technical claims. That is genuine expertise, and it has real value. But if you also build in the domain, you develop something different: an understanding of what it actually costs to ship, where the technical dead ends are, how long different classes of problem take to solve, and what a team looks like under pressure. This is not pattern recognition. It is operational knowledge, and it changes how you invest.

At Nexum, our build studio exists for two reasons. The first is straightforward: we believe there are valuable products to be built in applied AI, developer tooling, and data infrastructure, and we want to build them. We have the engineering capacity, the domain knowledge, and the appetite for long-horizon product development. Some of these products will become standalone businesses. Others will become tools we use internally and eventually open-source. The returns here are part financial, part strategic.

The second reason is less obvious: building makes us better investors. When we sit across the table from a founding team that is developing a data pipeline product, we are not relying on market maps and second-hand due diligence. We have written similar pipelines. We know the edge cases. We know which parts of the architecture tend to become expensive to maintain, which vendor relationships matter, and where the performance cliffs are. This doesn't mean we always agree with the founder's approach — often the founder knows things we don't. But it means the conversation is more honest and more useful.

The model also creates a virtuous loop within our portfolio. When a founder in our network runs into a technical challenge that our build team has already solved, we can help directly. Not with introductions and advice, but with code, architecture reviews, and shared infrastructure. This is a different category of support than most early-stage investors can offer, and it is one of the primary reasons founders choose to work with us over larger funds with more capital but less operational depth.

Being a builder also keeps us humble. We know, from direct experience, how hard it is to ship good software, to recruit strong engineers in a competitive market, to make the right technical bets early before the architectural decisions have locked in. That knowledge makes us less likely to impose unrealistic expectations on our portfolio companies, and more likely to recognise genuine progress when we see it. Building is the best antidote to the hubris that can creep into investor thinking when your only reference point is a pitch deck.