The week's thread — 1 November

Four rounds. Approximately €33 million. Four countries. The week of 26 October–1 November was the quietest of the month, closing on Toussaint weekend with a tight cluster of announcements that arrived over two days — Monday 26 October and Thursday 29 October — and nothing before or after.

The four rounds are: Volare (Pori, €7.1M, insect protein scale-up), Sonexos (Lausanne, €5.4M, plasma acoustics), Ravical (Ghent, $17M / €15.6M, accounting firm AI), and Bleu Robotics (Paris, €5M pre-seed, industrial humanoid learning). Median round: €6.25M. Stage spread: one seed (Ravical), one Pre-Series A (Sonexos), one Pre-Seed (Bleu Robotics), one unstated (Volare).

Research meets factory economics

Three of the four companies have the same origin story in different sectors: technology developed inside a European public research institute, spun out, proved in controlled conditions, and now raising capital specifically to answer whether the underlying science holds at industrial scale.

Volare, spun out of VTT in 2021, is the furthest along. It has built the facility — Volare 01, 10,000 m² in Pori, capacity 5,000 tonnes per year — on €15 million of prior capital, and this week's €7.1M from Springvest and Maki.vc is a ramp-phase raise, not a proof-of-concept. The insect protein sector has spent years promising factory economics without delivering them at verified industrial scale. Volare's claim is that the promise is now a plant, and the next eighteen months will say whether the per-tonne figures match the design projections.

Sonexos, an exclusive licensee of EPFL's Plasmacoustic Metalayer, is one stage earlier. The CHF 5M Pre-Series A from Valquest Partners and Robert Bosch Venture Capital funds the translation from working prototypes to customer-integrated hardware: prototype programmes with OEM leads in HVAC, automotive, and data-centre cooling, followed by regulatory qualification and IP expansion. The open question is not whether the plasma layer absorbs low-frequency noise in the lab — it does — but whether it survives OEM manufacturing tolerances, repeated production cycles, and the unit-economics scrutiny that hardware faces before a volume contract.

Bleu Robotics (Paris) spun out of INRIA in 2025. Its €5M pre-seed, co-led by Contrarian Ventures and OVNI Capital, funds the same kind of translation: moving abductive learning — which lets industrial humanoids acquire new tasks from as few as ten operator demonstrations — from a controlled research environment to automotive factory floors and WEPA Group production sites. The ten-demonstration claim is plausible for a bounded task in a stable environment; whether it holds across the full breadth of tasks on a live production line, where conditions and operators vary, is the question the next pilot cycle must answer.

Expert knowledge as a deployable asset

Ravical (Ghent) does not have a research-institute origin, but it shares the week's structural logic. Its $17M seed, led by MMC Ventures with Lakestar, encodes the working methods of a firm's best accountants into AI-executed processes that run across every client engagement — whether or not the best practitioner is assigned. The validated-in-controlled-conditions performance is the platform itself, already adopted by more than 100 European accounting firms. The deployment question is different in kind from Volare's: not whether the technology survives factory physics, but whether firms will accept the organisational change that firm-level standardisation requires. Individual AI tools are easier to sell because they improve the employee without asking the firm to change. Ravical's product asks the firm to change, and the change-management challenge will take at least as long to resolve as any technical one.

The week's shape

Four rounds in four countries. No investor appeared twice. No sector repeated. The aggregate is approximately €33.1M at a range of €5M (Bleu Robotics) to €15.6M (Ravical) — well below the prior week's pace. What the four share is not a sector or a geography but a moment: each is raising capital to resolve a specific open question that the proof-of-concept phase deliberately left unanswered. That is early-stage Europe's most reliable profile of fundable risk — not "can this work?" but "will it work here, at this scale, with these customers?". The next few months will say which of the four has asked the question that capital was right to back.