fundcraft raises €12M to run buyout fund back offices off a single ledger

The grievance that started this was with external service providers, which Julien De Mayer met while running fund administration at Rocket Internet ^[5], and the thing he built to fix it is an external service provider ^[16]. But nothing here is being pulled in-house by general partners; the vertical integration is happening on the vendor's side ^[10]. Net asset value calculations, capital calls, investor reporting and compliance filings ^[8] are the four jobs currently split across providers who reconcile to each other, and fundcraft's claim is that most of the market still runs those as dispersed processes rather than off one platform ^[19].

The licence is the part a rival cannot copy by shipping a new feature. fundcraft holds administrator permissions in Luxembourg and France and treats that as its edge as European reporting requirements tighten ^[16], while Carta, the largest US provider, keeps its own general ledger, leads US venture and private equity administration, and is by this account hard to leave ^[13]. Both facts point the same way: stickiness is fundcraft's asset and its obstacle, since every buyout manager it wants already has an administrator.

The capital side is thin in a way that is worth doing the division on. Of the €40M raised since 2021, €12M is new and €11M was the Series A closed in two stages during 2024, which leaves roughly €17M behind the first three years ^[1]. Spread across the funds now live on the platform, that is about €133,000 of lifetime equity per fund ^[2], at an average of roughly 67 limited partner subscriptions each ^[3], which is the shape of the venture and growth book the company started with ^[6]. Buyout structures are heavier than that, which is the whole point of the round and also the reason the historic ratio will not hold.

One number in the market framing does not check out. A $6.8bn 2025 market reaching $14.2bn by 2034 at 9.2% a year ^[9] compounds to about $15.0bn on those inputs, and the actual walk from 6.8 to 14.2 over nine years is 8.5% ^[4]. That is a sales input rather than a planning input, and it is the kind of figure a buyer of fund administration software should discount before it reaches a board pack.

What the €12M is not buying is somebody else's client list. De Mayer says the plan is to implement the operating model organically across jurisdictions "without stitching together disconnected platforms or service models" ^[7], which means licence applications and onboarding, and it means the French entity preparing ELTIF 2.0 and evergreen launches aimed at more than €1bn of commitments ^[18] competes internally with the AI-native build ^[17]. For scale on the fronts being funded: Hanover Park, founded in 2024, took $27M in a single Series A for AI agents on a dedicated ledger ^[17], and Juniper Square's disclosed total is above $130M ^[14].

This can be read three different ways going forward. If the first-half client surge is retail plumbing rather than institutional demand, the buyout claim is an intention with a licence attached; if the long, resource-intensive implementations the source attributes to Allvue ^[15] are a property of buyout complexity rather than of Allvue, fundcraft inherits them; and if AI-ledger entrants compress administration pricing, the licence stops being what the GP decides on. The client count is the only growth figure disclosed, and if the second half of 2026 matches the first, additions land at twice 2025 ^[5], with slightly more than two-thirds coming from clients new to the platform ^[6]. Until a buyout fund shows up inside that count of almost 300, the institutional half of the story is a French licence and €12M of intent ^[1].