Deep Tech Brief #05

Welcome to this week's issue of Deep Tech Brief.

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

Three companies this week, one thing in common: each wins by deleting the expensive part. A Greek software team is pulling more capacity out of the grid without laying new cable. A German engineer built an industrial mixer with no blade in it. A UK defence startup is flying strike drones with no engine. In each case, the component everyone assumed was essential turned out to be the thing holding the system back.

Europe is adding renewables faster than its grid can absorb them, and Greece is the preview. In 2025 the country curtailed 1,867 GWh of wind and solar, 6.6% of everything its renewables produced and more than double the year before. Researchers at Aristotle University expect 3.3 to 3.7 TWh in 2026. One Greek developer estimated curtailment wiped out roughly 20% of PV producers' revenue last year.

The other failure mode is louder. On 28 April 2025, mainland Spain and Portugal went dark for most of a day. Spain lost about 15 GW, close to 60% of its supply, in five seconds, in what ENTSO-E's expert panel later called the worst power system failure in Europe in over twenty years. Around 40% of Europe's distribution grids are already older than 40 years.

Optimems builds the software that manages what is already connected, treating every distributed asset (solar, wind, batteries, EVs, biogas) as a single dispatchable plant. That is what microgrids and virtual power plants were always meant to be. As co-founder and CEO Angelina Bintoudi puts it, academia proved these concepts 15 to 20 years ago; nobody built the intelligence layer that makes them deliver.

The edge is provenance plus product choice. Optimems spun out of CERTH, Greece's largest research centre, in 2022, carrying a decade of peer-reviewed work on microgrid scheduling and control. In a sector Bintoudi describes as running on "trust me, bro" from the big names, a publication trail is a sales asset. The product is deliberately unglamorous: a manufacturer-agnostic controller that lets a solar park meet grid operator specifications and respond to dynamic curtailment orders. Exactly the wound Greek producers are nursing.

By Bintoudi's account, Optimems became number one in sales in Greece within months, took over 25% market share, booked more than €2.5 million in revenue and just under €1 million in EBITDA, and raised funding to replicate that across eight EU countries including Germany. Those figures are the company's own. The awards are verifiable: first place at StartSmart SEE in 2023, a top-10 slot at Web Summit Vancouver in 2025, and Greece's NBG Seeds award in 2025.

Worth watching, because they are selling into the bottleneck rather than around it. The risk is the expansion: Germany's virtual power plant market is crowded with well-funded incumbents, and dominating a small home market is a different sport from running eight foreign ones at once.

Learn more: https://optimems.gr

Mixing is one of the most universal steps in manufacturing and one of the least re-examined. Nearly every process industry does it the same way: put a stirrer in the vessel and spin it. The costs are known. The blade shears delicate product, only a fraction of the mass passes through the mixing zone at any moment, dead zones persist where the tool cannot reach, and it takes hours. Battery makers feel it hardest: conventional planetary mixers run electrode slurry batches for two to six hours, upstream of everything else in a cell plant.

hs-tumbler removes the stirrer entirely. Its trajectory mixers seal the product in a container and move the container itself along programmed two-dimensional curves, driven by two motors swinging on X and Y axes and overlapped into circles, figure eights and effectively unlimited other paths. Because the whole mass moves, all of it is mixing all of the time. Different curves behave like different tools: a tight one applies high shear like a cream whipper, a wider one acts like a kneading hook, and the operator can switch profiles mid-run. Founder Bernhard Hukelmann adds that the sound inside the vessel is itself a quality signal.

The headline numbers are Hukelmann's: slurry that takes five and a half hours in a conventional stirred mixer done in roughly 100 seconds, with more than 90% less energy. There is independent validation too. The company won the International FoodTec Award in Gold at Anuga FoodTec 2024 and holds patents on the method, and by its own count runs around 20 systems from the US to Japan.

The honest comparison is twin-screw extrusion, which the battery industry already uses to turn batch mixing into continuous mixing at residence times of 30 to 120 seconds. hs-tumbler is not the only escape from the multi-hour batch. What it adds is programmability, a sealed vessel with no implement to clean, and near-instant changeover, which matter most where recipes change often and hygiene is tightly regulated.

One to watch, carefully. The physics is sound and the food traction is real, but the vessels are small, roughly 4.5 to 50 litres, and the throughput case rests on near-continuous operation rather than batch size. Battery and pharma buyers will want that proven at plant scale before replacing equipment that already works.

Learn more: https://www.hs-tumbler.com

Europe has a deep-strike problem it has been discussing openly for two years. Long-range precision weapons are extraordinarily expensive: a Storm Shadow costs around $2.5 million a unit, Taurus roughly the same, the American JASSM about $2 million. Buying thousands is a budget event, not a procurement decision. Which is why in February 2026 France, Germany, Italy, Poland, Sweden and the UK signed a letter of intent to jointly develop a 500km "one-way effector" priced in the five figures. That programme is still pre-procurement. The gap is real and unfilled.

Planetfall, a London-registered startup founded by aerospace engineer and German army reserve officer Victor Süsskind with co-founder Jóhannes Karlsson, attacks it from an unusual angle. Its system, Valkyrie, is a one-way drone carried into near space by a balloon, which then glides to a target hundreds of kilometres away. No engine means no heat signature, no noise and no fuel, and leaves a vehicle that is mostly airframe and guidance. Hence the cost argument.

The hard part is not the gliding. It is surviving where the balloon puts you: Süsskind describes drones that must keep working after up to 12 hours at minus 50 degrees in near vacuum, a satellite-grade environmental problem on a disposable-munition budget. He says it took three years, that the drone now works, and that the company holds an R&D contract with the British military. We could not independently verify that contract.

For a sense of how unforgiving the stratosphere is, Alphabet ran Loon for nine years, set a 312-day endurance record, then shut it down in January 2021 because it could not get costs low enough. Planetfall's escape from that trap is that its balloons are single-use and need to survive hours, not months.

One to watch, carefully, with the usual caveats for pre-revenue defence hardware. The physics is credible and the cost logic compelling, but a very small team is competing for procurement attention against national champions and fast movers like Cambridge Aerospace, which took an interceptor from first flight to a UK MoD production contract in about fourteen months. Wind-dependent delivery also imposes timing and targeting limits nobody has publicly quantified.

Learn more: https://planetfall.uk

The cheapest capability is usually the one you do not have to build. A controller instead of new cable. A moving vessel instead of a stirrer. A balloon instead of an engine.

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.