Enact's robot senses, Sonexos' plasma acoustics and Turiq's banking guardrails
Zurich really seems to be carving out an unmistakably grounded identity in practical deep tech right now. Look at Enact Intelligence stepping directly out of stealth with a thirty million dollar pre-seed financing round. Thirty million dollars for a pre-seed round is massive anywhere, but especially in Europe.
Source notes
This introduction contains no news claims.
Script
Emilia Hi, welcome to Europe Builds. I'm Emilia.
Erik And I'm Erik.
Emilia Zurich really seems to be carving out an unmistakably grounded identity in practical deep tech right now.
Erik It genuinely stands out. Look at Enact Intelligence stepping directly out of stealth with a thirty million dollar pre-seed financing round.
Emilia Thirty million dollars for a pre-seed round is massive anywhere, but especially in Europe. What specific engineering obstacle are they trying to solve?
Erik They frame it as physical AI. Their fundamental thesis is moving beyond modern autonomous robots that rely almost exclusively on computer vision to understand their surroundings.
Emilia Because when an autonomous machine relies solely on cameras, it stays remarkably clumsy whenever it has to touch, grip, or assemble items in unstructured physical environments.
Erik Exactly. If an arm can only look at an object, it has no direct sense of contact pressure until something slips, bends, or shatters.
Emilia Haptic feedback sensors have existed in academic research labs for decades. Why have they remained so rare across factory floors and warehouse logistics?
Erik Startupticker notes traditional haptic hardware is notoriously fragile and expensive. Most solutions measure force at single contact points on fingertips, which break quickly under relentless industrial wear.
Emilia You graze a gripper against a rigid steel beam during a fast cycle, and an expensive miniature sensor is instantly destroyed.
Erik Enact is pursuing a completely different architecture. They measure force directly inside the actuators, similar to how human bodies sense load and muscle tension throughout a limb.
Emilia Whole-body proprioception instead of delicate fingertip patches. That means redesigning the entire stack, from motor control to real-time neural inference.
Erik They call it full stack, unifying data pipelines, motor physics, and model architectures. Their backers include Redalpine, Notion Capital, firstminute capital, Arthur Mensch, and Yann LeCun.
Emilia Alongside Andrew Lonsberry from Path Robotics. When foundational AI researchers back physical hardware, it signals that bodily mechanics and sensory feedback are the defining frontier.
Erik They plan to quintuple their team size by late 2027, accelerate hardware pipelines, and start deploying pilot machines directly with industrial partners.
Emilia The real crucible will be mechanical reliability. Real factories operate around dust, thermal fluctuations, and strict speed requirements. Can whole-body force sensing endure twenty-four-seven operating cycles?
Erik That is why their focus on unit cost and robustness from day one is critical. Plant directors will not adopt delicate machinery, no matter how clever the neural network is.
Emilia It represents an engineering philosophy that refuses to bypass basic physics. You see that same stubborn mindset applied to everyday environmental nuisances.
Erik You really do. Solving invisible, grinding physical headaches is a recurring theme here, especially when you examine industrial acoustics.
Emilia You are talking about Sonexos. They just raised five million Swiss francs in equity financing to cancel acoustic noise using plasma physics.
Erik Absorbing acoustic energy with an ionized layer of gas sounds like science fiction at first glance. But the problem they target is remarkably widespread.
Emilia Low-frequency hums. The relentless drone generated by massive cooling fans in data centers, heavy industrial heat pumps, compressors, and building ventilation ducts.
Erik Those long acoustic wavelengths travel straight through building concrete, partitions, and triple glazing. They are the hardest sounds to muffle.
Emilia And traditional passive acoustic baffles must be huge and dense to absorb long waves, which clogs air ducts and chokes ventilation airflow.
Erik Standard active noise cancellation works neatly inside the sealed cavity of personal headphones. In sprawling open ducts, complex three-dimensional sound fields easily overwhelm mechanical speakers.
Emilia Sonexos bypasses moving cones entirely with their patented Plasmacoustic Metalayer. It is a solid-state, software-controlled surface that absorbs sound without mechanical wear.
Erik Their engineered prototype benchmarks show low-frequency attenuation reaching up to twenty decibels in ventilation setups, and sixteen decibels in home appliances.
Emilia A twenty-decibel drop on a logarithmic scale means the human ear perceives that sound at roughly a quarter of its original loudness. That is a noticeable difference.
Erik Valquest Partners led this Pre-Series A round, with Bosch Ventures participating as well. The presence of Bosch Ventures is very instructive here.
Emilia Bosch manufactures heating systems, industrial motors, and major household appliances. If you can make residential heat pumps near-silent without adding bulk, you solve a massive consumer friction.
Erik Sonexos is already running paid integration tests with HVAC and appliance manufacturers, working toward series production by early 2028.
Emilia Targeting 2028 highlights the reality of deep-tech engineering. You cannot rush physical hardware through continuous software deployments. You need methodical validation cycles.
Erik You have to satisfy electrical certifications, long-term component lifespans, and existing assembly tolerances. You cannot push an emergency patch to an installed air handler.
Emilia Even when you turn to software coming out of Zurich's academic community, you notice that same rigorous, controlled approach to real-world integration.
Erik You are thinking of Klaus Wellershoff partnering with two ETH Zurich computer science students to launch their fintech venture, Turiq.
Emilia Yes. Ryan Fuchs and Boris Kantor teamed up with Wellershoff, following his sale of Zwei Wealth to Swiss Life, to build an AI platform tailored for financial institutions.
Erik The platform helps chief investment officers and asset managers synthesize incoming market information with internal portfolio allocations to prepare investment memos and client communication.
Emilia What catches my eye is their explicit guardrail against automated execution. They deliberately do not let models execute trades or broadcast client notes autonomously.
Erik Human advisors select data inputs, establish classification guidelines, and approve every recommendation before release. Legal and financial accountability stays entirely with licensed professionals.
Emilia Plus, their architecture follows a bring-your-own-cloud design. Banks run the system within their own enterprise infrastructure rather than shipping data to external providers.
Erik Swiss wealth managers are uncompromising when it comes to banking secrecy and privacy laws. Sending client portfolio figures to external public servers is an instant dealbreaker.
Emilia They also spent roughly a year testing and refining the software directly inside Wellershoff and Partners before approaching external private banks.
Erik That practical incubation environment gives young technical founders immediate, candid feedback on what portfolio analysts actually need during turbulent trading mornings.
Emilia It keeps builders from creating theoretically elegant software that completely misses the regulatory realities of Swiss wealth administration.
Erik This reflects the wider commercialization push from both ETH Zurich and the University of Zurich through their entrepreneur fellowships and startup programs.
Emilia The University of Zurich highlights more than one hundred and sixty historical spin-offs, running competitive funding calls through DIZH and specialized entrepreneur initiatives.
Erik And that applied engineering spirit reaches well beyond private banking or heavy industry into surprisingly specialized niches across European sports and consumer logistics.
Emilia Look at 49ing applying specialized computer vision to ice hockey and floorball analytics. At first that sounds narrow, until you examine their customer roster.
Erik They already serve over five hundred institutional clients. They ingest match broadcasts to automatically produce player tracking data, tactical metrics, and searchable video archives.
Emilia They work with Switzerland's National League, top clubs in Finland's Liiga, and prominent North American collegiate hockey conferences like Hockey East and Atlantic Hockey.
Erik Now they are translating that multimodal tracking architecture into floorball and preparing a Series A funding round to back their international commercial expansion.
Emilia Ice hockey is blisteringly fast, with obstructed sightlines, rapid line changes, and tiny pucks. Teaching vision models to track players reliably in that chaos is genuinely difficult.
Erik If you conquer a high-speed, dynamic visual environment like that, you possess a technical defensibility that generic large vision models simply cannot touch.
Emilia You see that same sharp, single-minded focus in consumer software too. Take Airglitch, which just publicly unveiled their machine learning flight platform.
Erik Their system evaluates split tickets, multi-leg combinations, and bundled journeys, claiming to beat Google Flights fares in over twenty percent of searches.
Emilia Instead of an experienced traveler spending an entire evening juggling twelve open browser tabs to piece together split routings, an algorithm automates the search puzzle instantly.
Erik They are already discussing pilot tests with corporate partners. It reflects how applied software here targets tangible friction rather than chasing open-ended conversational novelties.
Emilia From robotic proprioception inside industrial actuators and plasmacoustic sound barriers to private wealth analysis, Zurich teams seem focused on difficult, defensible physical and data challenges.
Erik It feels like a deliberate ecosystem choice: leave superficial apps to others, and build serious industrial engineering answers that customers and partners will actually pay for.
Emilia We will be watching how these commercial pilots and series production milestones unfold over the coming years.
Erik Thanks for joining us on Europe Builds.
Emilia See you next time.