The 2020–2023 funding cycle in European InsurTech produced a large number of companies at the seed and Series A stage, many of which are now at or past the point where their original runway assumptions need to be revisited. We have spent the past several months tracking the market's response: which companies have found paths to extension rounds, which have been acquired, and which are in a quieter process of winding down or being absorbed into carrier innovation programmes. The patterns are instructive.
This is not an obituary for the cohort. A consolidation phase in a sector that attracted substantial capital for the first time is a normal part of the cycle, and the companies that survive it are typically those that built something structurally durable rather than something that depended on favourable financing conditions. What we are trying to understand — as investors who have backed a number of these companies — is whether the consolidation patterns map to what we would have predicted from first principles, and where we need to update our model.
Where Acquirers Are Coming From
The most active acquirers of European InsurTech companies over the past eighteen months have not been the large international carriers, as many in the market expected. They have been three other categories: mid-market brokers seeking to accelerate their digital distribution capability; reinsurers looking to bring underwriting model capability in-house rather than continuing to buy it as a service; and, in a handful of cases, larger InsurTech companies that have reached scale and are acquiring complementary functionality.
The broker-driven acquisitions make economic sense. A regional commercial broker with €300–500M in annual gross written premium does not have the engineering capacity to build a competitive digital distribution stack from scratch, and the regulatory and carrier relationship assets that make a broker valuable take years to accumulate. Acquiring a technology company that has already built the stack — and whose founders will stay on to run the product — is substantially faster than the alternative. The challenge for founders in this scenario is that broker acquirers are often buying the capability without buying the growth option. The valuation multiple reflects the cost of internal development avoided, not the sum-of-discounted-cash-flows of an independent growth company.
Reinsurer acquisitions are more varied. Several of the ML underwriting model companies that built individual risk assessment capabilities — particularly in commercial property and cargo — have been acquired by reinsurers who concluded that having the model capability in-house changed their treaty pricing authority in ways that were worth the acquisition cost. We are not saying that reinsurer acquirers overpaid in these transactions; many of the deals we are aware of were done at disciplined multiples. But the strategic logic is clear: a reinsurer that can price individual risks rather than pooled portfolios has a structural cost advantage over one that cannot, and the fastest path to that capability is acquisition of a team that has already built it.
What Determines Who Gets Acquired Versus Who Closes
The observable difference between companies that have attracted acquisition interest and those that have not comes down to a single variable more clearly than we expected: whether the company built a defensible data asset or sold access to a methodology. Companies that accumulated proprietary claims data, telematic event data, or structured loss history over multiple underwriting cycles have something that an acquirer values independently of the team. Companies that built a machine learning methodology on top of carrier data that remains the carrier's property have a much harder time making the case for a strategic acquisition — because the carrier could, in principle, replicate the methodology with a different team.
This is consistent with the framework we have used in evaluation since Fund I, but the consolidation cycle has produced a cleaner empirical test than we have had before. The companies we backed on the thesis that proprietary data accumulation was the core asset have, with one exception, either found acquirers or are in productive extension round conversations. The companies where the data question was less clearly answered at the time of investment have had harder outcomes.
The exception is worth noting: one company with a strong proprietary data asset has struggled because the asset is in a line of business — cyber liability — where the underwriting cycle turned sharply negative in 2024 and carrier appetite for new distribution relationships contracted. A good data asset in a temporarily distressed market is not sufficient to sustain a company if the distribution channel closes. This is a reminder that data asset defensibility and distribution access are both necessary conditions, not substitutes for each other.
The MGA Structure as Consolidation Vehicle
One pattern we have observed more than we anticipated is the use of the MGA structure as a consolidation vehicle. Several InsurTech companies that could not raise growth capital on standalone terms have agreed to be absorbed into larger MGA platforms, effectively becoming product lines within a broader insurance distribution entity. The founding team frequently remains, operating the product they built within a more capitalised structure that provides carrier access and compliance infrastructure.
From a venture perspective, this outcome is typically below the return profile that justified the original investment. From the perspective of the market as a whole, it may be productive: the technology and domain knowledge built by these teams does not disappear, and the MGA platform model provides a route to continued operation and eventual profitability that pure software company economics would not have supported. The insurance distribution chain has always been structured around relationships and capacity access in ways that make pure software margin profiles difficult to achieve; some of the companies that are being absorbed into MGA structures are finding a natural home for their products rather than experiencing a failure.
Implications for Seed Investment Criteria
The consolidation cycle has caused us to revise one element of our evaluation framework in a specific way. We now weight more heavily the question of whether the company's data asset is portable — that is, whether it accumulates in a form that the company controls, rather than residing in a carrier or broker system that the company accesses under contract. The distinction was always present in our framework; the consolidation patterns have made it the first question rather than a secondary one.
We have also updated our view on team composition for data-intensive insurance businesses. Domain expertise in the specific line of business — motor, property, cyber, cargo — matters more than we initially believed relative to general ML engineering competence. The teams that have navigated the consolidation period best are those where at least one founder has operated inside the underwriting process for the specific risk class they are modelling. The tacit knowledge of what the data actually means — what a claims adjuster's note in a motor file tells you that is not in the structured data fields, or which cargo risk variables are truly predictive versus spuriously correlated with loss — is not something that can be acquired quickly from outside the business.
We are not saying that general ML expertise is unimportant. It is a baseline requirement. But the differentiated performance in the current cycle has come from domain depth, not from algorithmic sophistication that any well-resourced team could replicate. This is the most important lesson the consolidation phase has produced for us, and it will affect how we evaluate the next cohort of seed-stage companies coming to market in the coming months.