Deep Tech Brief #10

Every Saturday, we break down deep tech companies building genuinely hard things, in plain language. No jargon, no hype.

Three companies this week: a forgotten Dutch tuber, Finnish grid hardware made from sawdust, and a Swiss implant the size of a grain of rice.

Let’s dive in.

There are roughly 30,000 edible plant species on Earth. More than half of humanity's calories come from three of them, and 95% come from thirty.

That narrowness is expensive, especially for protein. Protein is the costliest macronutrient to grow because plants need nitrogen to make it, and synthetic nitrogen fertiliser is energy-hungry, volatile in price and damaging where it runs off. Legumes solve this by pulling nitrogen from the air, which is why peas and beans are protein-rich, but they yield far less per hectare than a root crop. Potatoes and cassava yield enormously and contain almost no protein. Europe, which imports most of the plant protein it feeds to animals, sits between the two.

Aardaia, based in Wageningen, thinks one plant does both, and that it has been sitting in European hedgerows the whole time. The aardaker (Lathyrus tuberosus) is a legume that forms edible tubers. It fixes its own nitrogen and grows like a root crop. It was eaten across Europe for centuries, cultivated in Zeeland in the 1700s and sold by weight in Amsterdam markets in the 1800s, prized for a flavour close to sweet chestnut. Then agriculture consolidated around a handful of species and it was left behind, never actually bred into a modern crop.

That is the work. Aardaia collects seed from across the plant's wild range, from the Netherlands to Siberia, sequences it, and pairs that with detailed field measurement to pick winners fast. No genetic modification and no gene editing, which matters commercially: because the aardaker was documented as European food before 1997, it avoids the EU's novel food classification and the years of approval that follow. The company reports around 10 grams of protein per 100 grams of fresh tuber and is screening 750,000 genotypes this year, two million next.

New staple crops essentially do not happen. The last one commercialised at scale was triticale, a wheat and rye hybrid, and that took most of a century. As CEO Pádraic Flood puts it, inventing a crop historically took millennia.

Where they are: founded in 2023 by Flood, a Wageningen plant geneticist, and Mike Henske, previously at Indigo and Solynta, with a team of twelve. A €5 million seed round in July led by Point Nine, with Astanor, Grey Silo and returning backer FoodLabs, follows a pre-seed of about €770,000. The tubers already sell: De Nieuwe Winkel, a two-Michelin-star restaurant in Nijmegen, took the entire harvest last year and nearly all of it again this year.

Worth watching, because the regulatory shortcut is a real head start and chef demand proves people will eat it. The milestone ahead is agronomic: moving from Michelin plates to farm-scale fields, and turning protein-per-hectare potential into a yield a farmer can bank on.

Learn more: aardaia.com

Electricity grids used to be steady. Now they are full of wind and solar arriving in gusts, and factories and chargers that pull hard without warning. Keeping voltage and frequency inside their limits has become an active job, and UBS expects global energy storage demand to grow around 40% this year.

Batteries are the obvious answer, but they are the wrong tool for part of the problem. A battery holds a lot of energy and releases it over hours. Much of what a grid actually needs is the opposite: a very fast, very short push, delivered in seconds, over and over, for years. That is a supercapacitor's job, and supercapacitors have existed for decades.

The reason they are not everywhere is what they are made of. The working part is a carbon electrode with an enormous internal surface area, usually activated carbon produced from coconut shell or petroleum coke, largely imported into Europe and expensive to manufacture.

Granarium, spun out of VTT, Finland's national research centre, makes that carbon from wood. Its platform binds biocarbon structures using nanocellulose, drawing on forest industry side streams and agricultural residues, which in Finland arrive by the truckload as a waste product. The company says this cuts the capital cost of production by up to 80% and produces a device that is fully renewable and can be made locally, which also sidesteps Europe's dependence on imported critical raw materials.

Where they are: early. VTT has transferred the technology and patents to the company, which raised over €1 million in a pre-seed round in June led by BSV Ventures with Beamline and angel networks from Finland, Estonia and Latvia. Chief executive Paula Viinamäki is a former Nokia and Microsoft director who joined VTT specifically to bring this out of the lab. The plan is first pilots with industrial customers, then manufacturing of up to 50 units a year.

Worth watching, because turning a local waste stream into grid hardware is exactly the kind of substitution Europe keeps saying it wants. The milestone ahead is straightforward and near: the technology sits at the stage where it has been validated in a lab, and the pilots now have to show the same numbers running on a real industrial site.

Learn more: granariumtech.com

Some birth defects can now be treated before birth. Surgeons pass a thin scope through the mother's abdomen, through the wall of the uterus and into the amniotic sac, and operate on the fetus in the womb. It is used for spina bifida, twin-to-twin transfusion syndrome and congenital diaphragmatic hernia. It works, and it saves lives.

It also leaves a hole. The instruments have to puncture the amniotic sac, and that membrane does not heal the way skin does, so the puncture is simply left open. Fluid can leak, the membrane can rupture, and the baby arrives early. By founder Yannick Devaud's account this is the main thing holding the field back, because a surgeon weighing an operation has to weigh preterm birth against it, and that limits both how often these procedures happen and how ambitious they get.

KOVE medical, a spin-off from the University of Zurich, has built a small implantable device that closes the defect at the end of the operation, sealing the membrane the instruments went through. It is a deliberately unglamorous product: not a new surgery, not a new drug, just the missing last step of a procedure that already exists.

The evidence so far is preclinical. The company has demonstrated the seal outside the body and in a pregnant sheep model, where it reported the closure holding long term. That is the standard path for an implantable device, and the reason a six-year-old company is only now approaching patients.

Where they are: founded in 2020 by Devaud, a bioengineer trained at EPFL and Zurich, and based at Technopark Zurich with a team of about seven. A CHF 1.6 million seed round in August, backed by a syndicate of angel investors, is earmarked specifically for reaching a first-in-human study. Earlier support came from Venture Kick and a UZH entrepreneur fellowship.

One to watch, carefully. The problem is unusually well defined, the target is a single named complication rather than a vague improvement, and the sheep data suggests the seal holds. The open questions are the ones every implant faces in order: the first-in-human study this round is funding, then a trial large enough to show the rupture rate actually falls.

Learn more: kovemedical.com

Know a deep tech company we should break down next? Hit reply and tell us. The best editions come from real tips.

See you next Saturday.