Let’s be honest. December is usually the month where productivity goes to die. You are mentally checking out, arguing with your uncle about politics, and wondering if it’s socially acceptable to eat chocolate for breakfast.

But apparently, nobody told the European Deep Tech scene that it was time to rest.

While the rest of the world was putting up Christmas trees, Venture Capitalists were busy shoving roughly +$1.5 billion into startups that are trying to do things that are, quite frankly, absurdly difficult. We are talking about companies trying to rewrite the laws of quantum mechanics, grow meat in vats, and cure cancer by stressing it to death.

If SaaS is “selling shovels during a gold rush,” Deep Tech is “inventing a laser that teleports the gold directly out of the ground.” It’s capital intensive, high risk, and if it works, it changes civilization. If it doesn’t, well, at least we got some cool patents out of it.

Here is what happened in December 2025.

🚀 Market Overview

I tracked 87 deals this month. The money flowed heavily into three main arteries: Energy Transition (hydrogen, tidal, batteries), Quantum/Semiconductors (the plumbing of the future), and Biotech (specifically AI-driven discovery and novel delivery mechanisms).

Geography: The UK, France, and Germany are still the heavyweights (almost half of the market), but the Nordics (Finland, Denmark, Norway) are punching way above their weight class, particularly in space and quantum. The Netherlands continues to be the undisputed king of “we build the machines that build the chips.”

The Trend: We are seeing a massive shift from “Bits” to “Atoms.” Investors are bored of chatbots. They want hardware. They want sovereignty. They want things that you can drop on your foot.

🟢 My Top10 Picks

These are the deals that actually matter. They aren’t just buzzwords; they represent genuine engineering breakthroughs or scientific pivots. This is where the “Science Risk” is high, but the payoff is astronomical. Maybe they won’t give the best returns but still more interesting than another SaaS, 

1. Sparrow Quantum (Denmark)

• Round: $7M Series A
• Investors: Scale Capital, LIFTT, North Ventures
• The Tech: Deterministic Single-Photon Sources (SPS). Sparrow utilizes self-assembled semiconductor quantum dots embedded within photonic crystal waveguides. This architecture couples the light-matter interaction with near-unity efficiency, ensuring that exactly one photon is emitted upon excitation.
• The Problem: Stochastic Emission & Multi-Photon Events. Current quantum light sources are largely probabilistic (attenuated lasers). They suffer from “multi-photon events” (emitting two or zero photons instead of one), which drastically reduces the fidelity of quantum gates and ruins entanglement protocols required for optical quantum computing.

2. Delft Circuits (Netherlands)

• Round: $9.4M Series A
• Investors: High-Tech Gründerfonds, DeeptechXL
• The Tech: Cri/oFlex® Technology. These are multi-channel microwave transmission lines lithographically printed on flexible, multilayer polymer substrates designed specifically for cryogenic environments.
• The Problem: Thermal Conductivity & Signal Integrity. Traditional semi-rigid coaxial cables act as “thermal bridges,” conducting heat from room temperature down to the milli-Kelvin mixing chamber of a dilution refrigerator. This heat load limits the cooling power available, capping the number of qubits a system can support before it thermally destabilizes.

3. Relation Therapeutics (UK)

• Round: $55M Late VC
• Investors: Novartis
• The Tech: Active Learning Framework on Single-Cell Genomics. Their platform integrates wet-lab robotics with computational models that utilize “Active Learning”—meaning the AI doesn’t just analyze data; it autonomously designs the next experiment to fill gaps in its knowledge graph, specifically using high-resolution single-cell RNA sequencing (scRNA-seq).
• The Problem: Data Sparsity & Cellular Heterogeneity. Traditional drug discovery relies on bulk sequencing, which averages out the gene expression of millions of cells, obscuring the specific mechanisms of disease in complex microenvironments (like bone tissue). Public datasets are often too “noisy” to train reliable predictive models.

4. Spark Cleantech (France)

• Round: $35.1M Series A
• Investors: Innovacom, 360 Capital
• The Tech: Cold Plasma Methane Pyrolysis. The company utilizes Nanosecond Repetitively Pulsed (NRP) discharges to generate a non-equilibrium plasma. This state of matter efficiently breaks the C-H bonds in methane (CH4) to produce hydrogen (H2) and solid carbon (C).
• The Problem: Thermodynamic Efficiency & Carbon Intensity. Steam Methane Reforming (SMR) is cheap but releases massive amounts of CO2. Water electrolysis (Green Hydrogen) is clean but energetically expensive due to the high bond dissociation energy of water.

5. Alva (Norway)

• Round: Early VC
• Investors: Samsung Ventures
• The Tech: Coreless Stator Topology (FiberPrinting™). Alva manufactures electric motors using a proprietary process that weaves conductive copper directly into a fiber-reinforced composite structure, eliminating the heavy steel laminate core found in traditional motors.
• The Problem: Hysteresis Losses & Cogging Torque. Iron cores in standard motors are heavy and suffer from magnetic hysteresis (energy lost as heat during magnetization cycles) and “cogging” (torque ripple), which introduces vibration and limits precision in robotics.

6. FYLA (Spain)

• Round: Late VC
• Investors: Indraventures
• The Tech: Supercontinuum Fiber Lasers. FYLA develops light sources based on nonlinear fiber optics that broaden a narrowband laser pulse into a massive spectral bandwidth (from UV to IR) while maintaining spatial coherence—effectively a “white light laser” with femtosecond pulse duration.
• The Problem: Spectral Limitations in Metrology. Detailed material analysis often requires hyperspectral imaging. Currently, this necessitates complex, unstable arrays of multiple laser sources or slow tunable filters, making real-time industrial inspection difficult.

7. Sofant Technologies (UK)

• Round: $8.4M Late VC
• Investors: EMV Capital, Scottish Enterprise
• The Tech: RF MEMS Beamforming. Sofant utilizes Radio Frequency Micro-Electro-Mechanical Systems (microscopic physical switches) to control the phase of radio waves, allowing antennas to electronically steer beams without moving parts.
• The Problem: Power Consumption in Phased Arrays. Current Active Electronically Scanned Arrays (AESA) use solid-state semiconductor phase shifters. These are “lossy” (signal strength drops) and inefficient, generating significant heat and draining batteries rapidly in mobile applications.

8. BiPER Therapeutics (France)

• Round: $1.2M Early VC
• Investors: Bpifrance
• The Tech: First-in-Class BiP Inhibitors. The company develops small molecules that target the Binding immunoglobulin Protein (BiP), a master regulator chaperone located in the Endoplasmic Reticulum (ER).
• The Problem: The Unfolded Protein Response (UPR). Cancer cells have high metabolic rates, leading to an accumulation of misfolded proteins. To survive this stress, they hijack the UPR pathway. Current chemotherapies often fail because the cancer cell simply upregulates this survival mechanism to repair the damage.

9. Prior Labs (Germany)

• Round: Seed
• Investors: Sequoia
• The Tech: Structured data Model. Prior Labs builds TabPFN, a pre-trained transformer for structured (tabular) data. Trained on billions of synthetic tables, it learns general “how-to-learn” priors and makes classification/regression predictions via in-context learning—often a single forward pass, no per-dataset training.
• The Problem: Hallucination & Lack of Reasoning. Most business and scientific data lives in tables, but using it usually means slow, expert-heavy ML pipelines (feature engineering, tuning, retraining) for every new dataset. Prior Labs aims to deliver strong predictions in seconds with minimal setup, so teams can ship models as easily as they run a query.

10. UMA Robotics (France)

• Round: Seed
• Investors: Yann LeCun, Thomas Wolf, Greycroft
• The Tech: Embodied AI & Universal Navigation. UMA utilizes vision-based foundation models and end-to-end learning. This allows their mobile and humanoid robots (Universal Mechanical Assistants) to “see” and understand the semantics of the physical world, demonstrating human-level dexterity and adapting to dynamic changes in real-time without needing a pre-mapped “digital twin.”

   

• The Problem: The Rigidity of Automation. Traditional industrial robots are “blind” and brittle. If you move a shelf or a person walks in the way, the robot stops or fails. This limits automation to perfectly controlled factories, keeping it out of dynamic logistics hubs or construction sites.

🟡 Honorable Mentions

These are solid companies, massive infrastructure plays, or technologies that are promising but aren’t as interesting for me as TOP10. But still great technology and a solid piece of deep tech. 

Amplifold (Germany)

They are building a “Generative AI” specifically for biology. Think of it like ChatGPT, but instead of writing essays, it writes the genetic code to create new proteins that don’t exist in nature. These custom-designed proteins can then be used to develop new medicines and therapeutics.

ICEYE (Finland)

They operate the world’s largest fleet of radar satellites (SAR). Unlike normal cameras, these satellites send out microwave pulses that can “see” through clouds, smoke, and total darkness, taking high-resolution photos of the Earth’s surface 24/7.

Automata (UK)

They build a robotic system called LINQ that physically connects all the different machines in a science lab. It uses robotic arms and conveyor belts to move test tubes and samples from one instrument to another, allowing experiments to run entirely on autopilot without a human needing to be there.

Neural Concept (Switzerland)

They created an AI software that understands 3D shapes and physics. It helps engineers design cars and planes by instantly predicting how air will flow over a design (aerodynamics), replacing the need for slow and expensive supercomputer simulations.

Gradium (France)

They are building an AI model that natively understands and generates audio. Instead of turning speech into text first (like Siri does), it processes sound directly, allowing it to have instant, natural conversations with human-like emotions and timing.

NanoXplore SAS (France)

They manufacture special programmable computer chips (FPGAs) designed specifically for space. These chips are built with a unique physical structure that prevents them from breaking or glitching when hit by cosmic radiation outside Earth’s atmosphere.

Nu Quantum (UK)

They build the networking equipment for quantum computers. Their technology uses light to connect multiple small quantum processors together, allowing them to communicate and act as one giant, powerful supercomputer.

Mosa Meat (Netherlands)

They grow real beef burgers directly from cow cells in a bioreactor, without raising or slaughtering animals. They focus specifically on cultivating real fat tissue, which is the secret ingredient that gives meat its authentic flavor and texture.

Water Horizon (France)

They make “thermal batteries” that capture wasted heat from factories. Using a reversible chemical reaction, they store this energy in a tank and transport it by truck to heat other buildings (like swimming pools or district heating networks), releasing the heat only when needed.

Orbital Marine Power (UK)

They build the world’s most powerful floating tidal turbine. It looks like a ship floating on the sea surface with giant rotors submerged underwater that generate electricity from the predictable movement of the tides.

Xeltis (Netherlands)

They create heart valves and blood vessels from a special porous polymer. Once implanted, the patient’s own body grows new natural tissue over the scaffold. As the new tissue forms, the plastic slowly dissolves, leaving behind a brand-new, living heart valve.

Q5D Technologies (UK)

They use 5-axis robots to print electrical wiring directly onto the surface of objects, like the inside of a car door or an airplane seat. This eliminates the need for heavy, tangled bundles of cables and automates the manufacturing process.

Nuage Therapeutics (Spain)

They are developing drugs for prostate cancer that target “shapeshifting” proteins (IDPs) which are usually impossible to block. Their medicine works by attacking the liquid droplets inside cells where these harmful proteins hide.

.lumen (Romania)

They build glasses for the visually impaired that replicate the main features of a guide dog. Using self-driving car technology (sensors and AI), the headset maps the environment and guides the user through haptic impulses (vibrations) on the forehead, helping them navigate obstacles safely without needing a biological animal.

Liom (Switzerland)

They are developing a wearable sensor that measures blood sugar levels using light. This technology allows people with diabetes to continuously monitor their glucose levels without ever having to prick their finger with a needle.

ORiS (Italy)

They are developing a system to send electricity wirelessly through space using laser beams. They can beam power from a satellite in the sunlight to a rover exploring the dark side of the moon, keeping it charged without heavy batteries.

Damae Medical (France)

They built a device that looks like a microscope but sees deep into the skin’s layers in 3D. It allows dermatologists to diagnose skin cancer instantly by simply scanning the skin, without needing to cut a piece out for laboratory testing.

🕵️‍♂️ Investor Activity: Following the “Smart Money”

In Deep Tech, the source of capital is often more important than the amount. Looking at December’s cap tables, we see a clear convergence of Silicon Valley royalty, AI legends, and industrial giants all making strategic bets in Europe.

American venture capital are no longer waiting for startups to move to San Francisco; they are actively hunting in Europe.

• Sequoia backed Prior Labs (Germany), betting on specialized AI for science.
• Andreessen Horowitz (a16z) invested in Mirelo AI (Germany), pushing into generative audio.
• Khosla Ventures doubled down on PolyAI (UK), validating enterprise voice agents. 

Strategic Sovereignty & Supply Chains Finally, industrial giants moved to secure critical components. Samsung Ventures invested in Alva (Norway), likely seeking next-gen motors for their future hardware. MBDA (missile systems) backed NanoXplore (France) to ensure a sovereign European supply of radiation-hardened chips. And NVIDIA partnered with pharma giant Novartis to invest in Relation Therapeutics (UK), confirming that the future of drug discovery relies as much on silicon as it does on biology.

Happy New Year. Go build something hard. 🚀