European defence money went late-stage: $4.7B of a record $8.7B in 2025

European defence funding has stopped behaving like an early-stage market. Late-stage investment tripled to $4.7 billion in 2025, more than half of the record $8.7 billion raised by European defence, security and resilience startups over the year ^[1][2][3], which is about 54 percent of the total and leaves roughly $4.0 billion for everything earlier in the stack ^[1][2].

That split matters more than the headline total. A market where the majority of dollars sit in late-stage rounds is a market being priced by growth funds and strategics against contracts and delivery schedules, not by seed investors against theses. It also means the record year is not evidence that it got easier to start a defence company in Europe. It got easier to finance one that already ships.

The early-stage numbers are smaller and reported in a different currency, which is worth flagging rather than smoothing over. Pre-Series B startups have raised 556 million euros so far in 2026, nearly matching the 635 million euros raised across all of 2025 ^[4], about 88 percent of the prior year in a fraction of the time ^[4]. Set 635 million euros against $8.7 billion and, at any exchange rate between 1.0 and 1.2 dollars per euro, the pre-Series B cohort accounted for roughly 7 to 9 percent of the 2025 total ^[5]. The bottom of the funnel is active. It is not where the money is.

Within that bottom, one segment moved violently. Funding for munitions-adjacent strike, air defence and counter-UAS reached just under $250 million through the third quarter of 2025, up from $5.1 million in 2024, the fastest-growing slice of Europe's defence tech market ^[5]. That is close to a 49-fold increase in a year ^[3]. Growth of that shape is a procurement signal being repriced by capital in real time, and it is also a crowding warning: the segment went from unfunded to contested inside four quarters.

The company picks that Tech Funding News collected from European investors ^[6] show what that crowding buys. Agon, nominated by Speedinvest's Julie Forel, was founded in 2026 and raised a $30 million seed led by Lakestar and XYZ ^[9], money raised inside its first calendar year of existence ^[6]. Its product is a synthetic battle arena for training and testing autonomous defence systems against adaptive opponents ^[10], with founders out of Anduril, Applied Intuition, Cambridge Aerospace and Unity, and Unity's David Helgason as board chair ^[11]. Forel's stated reason is the shortage of scalable, high-quality training and testing data for autonomous systems ^[12].

Compare the older cohort. 3YOURMIND, founded in 2014 and nominated by UVC Partners' Johannes von Borries, has raised $35.8 million across eleven years and two pivots, and sold to the US Department of Defense before defence tech was fashionable, against its investors' advice ^[7][8]. Agon's single seed round is roughly the size of what 3YOURMIND assembled over a decade ^[8]. Hypersonica, founded 2023 and nominated by Antler's Alan Poensgen, has raised $28.4 million total ^[13], including a 23.3 million euro Series A led by Plural with SPRIND, General Catalyst and 201 Ventures ^[14]. In February 2026 it flew an unarmed prototype above Mach 6, over 7,400 kilometres per hour, for more than 300 kilometres at Andoya Space in Norway ^[15], with PitchBook putting headcount at about 50 to 60 ^[16].

Watch three things. Whether the 2026 pre-Series B pace holds past the 635 million euro mark ^[4], or whether it stalls as late-stage absorbs allocation. Whether the counter-UAS jump repeats in 2026 or reads as a single-year pull-forward ^[5]. And whether the seed-stage prices being paid in simulation and strike survive contact with actual procurement, because at 54 percent late-stage ^[1] the buyers of those companies are already the ones setting terms.