Writing
Thinking on AI, automation, and what it takes to build B2B software in Europe.
The first decade built the proof points. The second decade is about compounding — and the founders who understand the difference are already ahead.
Being a former founder helps in diligence. It can hurt in portfolio dynamics. The distinction matters more than most people in VC will admit.
Three years of watching supply chain AI investments suggests most of the value accrues in one specific part of the stack — and it's not where most founders build.
After 18 months of evaluating LLM infrastructure plays across our portfolio, a clearer picture has emerged of what actually matters at the infrastructure layer.
The pattern is consistent enough that we now look for it in diligence: extraordinary technical depth, genuine market insight, and a complete misread of how enterprise decisions actually get made.
Procurement doesn't generate press releases. But it's where AI-native software is compounding value fastest — because the manual process cost is enormous and the switch cost is low.
Fund II has a different emphasis than Fund I. Here is what changed in our thinking and why we are now backing horizontal platform plays in addition to vertical automation.
Every time a US fund asks us how we handle the regulatory environment in Europe, they are asking the wrong question.
The DACH technical talent market looks materially different than it did four years ago — and it matters for how we evaluate founding teams.
The investor deck says "support on go-to-market." Here is what that actually means in practice — the introductions, the first-meeting coaching, and the uncomfortable conversations.
Fund I taught us specific lessons about where we add value and where we don't. Fund II is built around that honest accounting.
The first wave of enterprise AI created a thousand point solutions. The second wave is about which of those solutions become the connective tissue of enterprise operations.
Conflating LLM capability with workflow automation creates bad investment theses and bad product roadmaps. Here is why they are different — and why both matter.
The board dynamic at seed sets the tone for everything that follows. Most of the mistakes we have seen start early and compound in ways that are hard to unwind later.
American SaaS moat frameworks don't transfer cleanly to European B2B. Here is a more accurate map of where European software defensibility actually comes from.
Document processing automation has finally reached the quality threshold where enterprise buyers can trust it on live workflows. The window for building defensible positions is open.
Our technical diligence framework at seed is not about finding bugs. It is about understanding whether the technical decisions reflect genuine understanding of the problem.
The cultural and structural dimensions of enterprise sales in Germany that are rarely written about — and that trip up even experienced B2B founders building for the first time in this market.
One year into building a portfolio of AI-native B2B companies, the same structural patterns appear in how automation compounds value. Here are five of them.
Having built a company doesn't automatically make you a better investor. But it changes which questions you ask — and which answers you trust.
Looking at the first four companies we backed reveals a pattern we didn't fully articulate at the time — but one that now defines how we evaluate every new deal.
The term is being used for everything from a linear regression model to a fully adaptive enterprise platform. Here is what we mean when we use it — and why the distinction matters for capital allocation.
The DACH mid-market is not the obvious choice for building AI B2B software. That is precisely why it is interesting — and why the competitive dynamics are better than they appear.
While US funds pour capital into consumer AI and developer tools, European enterprise automation sits underpriced and underloved. This is the first piece in a series on why we think that's about to change.