Fund II closed in September 2023 at €63 million, and we have now deployed approximately half of that capital across a set of seed and early Series A investments in European InsurTech. The halfway point feels like a reasonable moment to write something honest about what the deployment pattern has revealed — where the thesis has held, where we have had to update it, and what the second half of the fund is likely to look like given what we know now that we did not know then.
I want to be specific about what "the thesis has held" means. We wrote the Fund II thesis in mid-2023, when we had five years of Fund I data to draw on and a view of where European insurance infrastructure was heading over the following decade. The Fund II thesis centred on three claims: that ML-driven individual risk assessment would displace pooled loss modelling across the major commercial insurance lines by the end of the decade; that the highest-value InsurTech companies would be those that accumulated proprietary risk data rather than those that built algorithms on top of shared carrier data; and that regulatory complexity in European insurance was a moat for specialised operators rather than a barrier to entry. The Fund I data supported all three claims. The Fund II investments are beginning to provide a second test.
Where the Thesis Has Held
The individual risk assessment thesis has held most clearly in the commercial property and cargo lines. The companies we have backed in these sectors are now at the point where they have enough live underwriting data to demonstrate that their individual risk models are outperforming the pooled pricing benchmarks on combined ratio. The performance gap is not dramatic — three to five percentage points on loss ratio in a normal underwriting year — but it is consistent and, more importantly, it is widening as the proprietary data sets accumulate. The models trained on eighteen months of live data outperform the models trained on six months of live data by a measurable margin. This is the compounding data asset dynamic that the thesis predicted.
The regulatory moat thesis has also held, though not in the way we initially described it. We expected that regulatory complexity would advantage specialists who had built compliance infrastructure over generalists who had not. What we observed instead is that regulatory complexity has advantaged specialists who have built trusted relationships with supervisory staff — relationships that are built over multiple engagements with the same national competent authority, not simply by building compliant systems. The carriers that are moving fastest through the procurement process for InsurTech solutions are choosing vendors they have already had a supervisory conversation about, not merely the most compliant vendor they have not yet spoken to a regulator about. Regulatory trust is relational, not just procedural.
Where We Have Had to Update
The area where we have updated our thinking most substantially is on the commercial model structure for ML underwriting infrastructure companies. We expected the dominant commercial model to be a software-as-a-service arrangement with carriers: the InsurTech company builds and maintains the model, the carrier uses it for underwriting decisions and pays a per-decision or licence fee. This model exists, and several portfolio companies operate it successfully. But a significant portion of the value creation in the Fund II cohort is coming from companies that have chosen to take underwriting risk themselves — operating as fronting carriers, capacity providers, or MGAs with binding authority — rather than selling model access to incumbents.
This was predictable in retrospect. A company that has built a demonstrably superior individual risk assessment model and has enough capital to take underwriting risk can capture the full economic value of its accuracy advantage rather than sharing it with a carrier that uses the model under licence. The SaaS model is capital-efficient but value-sharing; the capacity model is capital-intensive but value-capturing. The best-performing companies in the Fund II cohort are those that started as SaaS and migrated to capacity as they accumulated proof of model performance — using the SaaS arrangement as a subsidised data accumulation period before assuming risk.
We are not saying every InsurTech company should aspire to be a carrier. The capital requirements for taking underwriting risk are real, and the regulatory authorisation process for insurance is substantially longer and more demanding than for a software company. But the trajectory of the highest-performing companies in the cohort is instructive: they treated the SaaS phase as a means to an end, not as a terminal business model.
On Deployment Pace and the Remaining Capital
We have deployed at a somewhat slower pace than the original fund model projected for the first half of the deployment period. This is not a reflection of deal flow shortage — we have reviewed more seed-stage InsurTech opportunities in 2024 and 2025 than in the comparable period of Fund I. It is a reflection of a deliberate decision to be more patient on valuation than we were in 2022 and early 2023, when seed valuations in European InsurTech were elevated by the broader venture market environment.
The current market has normalised. Seed round valuations for European InsurTech companies with credible domain expertise and a clear data strategy are back to the range we consider structurally sound for the risk profile. Pre-revenue seed companies are pricing at pre-money valuations in the €4–8M range for the founding team and early architecture, compared with the €10–15M range that was common in 2022. This normalisation is good for the fund's entry economics and, more importantly, it means that the companies raising at current valuations have had to sharpen their proposition — the environment has filtered out teams that were relying on narrative rather than evidence.
For the remaining capital, we expect the deployment pattern to shift modestly toward companies that are at a slightly later point in their product development than the typical seed we have backed — companies that have a live pilot with at least one carrier or MGA, have accumulated some proprietary claims or underwriting data, and are looking for the first institutional investor to anchor a seed extension or Series A. The market correction has created a cohort of companies that raised seed rounds in 2022–2023 on optimistic assumptions, have survived the adjustment period, and are now at the point of needing growth capital at a valuation that is realistic relative to their current position. Several of these companies are among the most interesting conversations we are having heading into 2026.
A Note on the Portfolio Support Model
One learning from the first half of deployment that will inform how we work with the second cohort is about the nature of the support that matters most at the seed stage in insurance. We have historically provided introductions to carrier and broker contacts and general strategic advice on product and market direction. What we have found is more specifically useful is regulatory navigation support — introductions to supervisory staff, support in understanding how a specific regulatory question will likely be interpreted by a specific national competent authority, and access to legal counsel with deep knowledge of European insurance regulation rather than general financial services regulation.
The regulatory navigation advantage of an insurance-specialist investor over a generalist is larger than we had fully appreciated when we were raising Fund II. We have now formalised this as a structured offering for portfolio companies — a regulatory access programme that gives seed-stage companies access to the supervisory relationships we have built over six years of investing in the sector. The feedback from portfolio companies that have used it suggests it materially accelerates the carrier procurement timeline, which is the primary constraint on revenue growth for most companies at this stage. We plan to deepen this programme for the second half of deployment.