Connect with us

NEWS

Lupa, Once Prismus, Wins a Swiss Prize With 30 Customers

Zurich AI firm Lupa, formerly Prismus, finished Venture Kick’s CHF 150,000 path. Thirty scientific-hardware companies already pay to find labs by technical fit.

Published

on

Lupa, the Zurich AI firm formerly called Prismus, has completed Venture Kick’s three-stage path for CHF 150,000 (€163,000), with 30 paying customers already on the books.

The people writing cheques are scientific hardware companies that need labs and researchers whose work matches a laser, a chip, or an electrode array, not a job title in a sales database.

A Three-Stage Prize of CHF 150,000

Venture Kick’s own post on October 29, 2026, used the phrase it prints for every team that clears the last jury: Lupa (ex-Prismus) receives the CHF 150,000 Stage III path. That figure is the program total, not a fresh seed cheque dropped on a blank cap table.

The public record already showed a Stage II award of CHF 40,000 on May 7, 2026. Across the full model, selected Swiss teams present to juries for a CHF 10,000 grant, then a CHF 40,000 convertible loan, then a CHF 100,000 convertible loan. The first slice is a grant with a moral pledge to return it if the company works; the later slices convert at the next significant equity round.

For 2026 the philanthropic program set aside CHF 9,580,000 for 283 kicks, with Kickers Camps after each stage and a stated selection chance above 50 percent at each jury. Stage III is the finish of that nine-month track, and it is also the ticket to later cheques that dwarf the prize in the headline.

Co-founder and CEO Elisabeth Blyumin, who trained in computational physics at ETH and worked in tribology and quantum sensing, put the camps, not the cash, at the centre of the note that went out with the award.

VentureKick has excited us, challenged us and pushed us into the next phase of our company. In particular, the Kickers Camps were intense and rewarding to complete.

Elisabeth Blyumin, co-founder and CEO of Lupa

THE PUBLIC MONEY ON THE PATH

Award Date Amount named
Stage II May 7, 2026 CHF 40,000
Stage III (full path) October 29, 2026 CHF 150,000
Program split Three juries CHF 10,000 grant, then CHF 40,000 and CHF 100,000 loans

CTO Timen van Gelderen, Chief AI Officer Gonçalo Hora de Carvalho, and Chief Revenue Officer Peter Kempin complete the named executive bench. The company’s LinkedIn page lists four employees against a headcount band of two to ten, founded in 2025 and based in Zurich.

Menhir Photonics and MaxWell Are Already Using Lupa

By the Stage II interview in May, the turning point had already been three paying clients, all Swiss deep-tech firms. The team told Venture Kick its scientific hardware companies as first users were hunting prospects from underlying science and applications rather than surface market data, and that it had narrowed a broad vision to technical sales teams in deep-tech industries.

That wedge still shows on the product site. Florian Emaury, CEO and co-founder of Menhir Photonics in Dübendorf, is quoted praising organised data and the founders’ support. Menhir, started in 2018, sells femtosecond lasers for labs and field use, with systems the company says reach a 2.5 GHz repetition rate. Emaury’s own path runs from the Institut d’Optique through a physics PhD at ETH in Ursula Keller’s group, which is the same talent pool Lupa is trying to turn into a lead list.

The sharper commercial line sits with Giulio Zorzi, head of EMEA at MaxWell Biosystems, an ETH Zurich spin-off from 2016 that builds high-density microelectrode arrays. MaxWell’s hardware records from 26,400 electrodes, and Zorzi was hired to run a regional sales and success bench across Europe, the Middle East, and Africa. On Lupa’s site he does not talk about patents in the abstract. He talks about a pipeline.

Lupa is helping us in identifying and highlighting the scientific researchers whose work actually matches our technology. Their analyses added valuable clarity to our pipeline and are helping us direct our commercial strategy toward the opportunities with the highest potential impact.

Giulio Zorzi, Head of EMEA, MaxWell Biosystems, on Lupa’s site

A laser maker and a neuroscience-tool vendor are not “industry” in the press-release sense. They are small commercial teams that have to find the ten labs whose protocols actually need their box, which is why the MaxWell Biosystems EMEA commercial team is the kind of desk Lupa is built for. Photonics, quantum, physics, and life science are the fields the company says those 30 paying customers already span.

What Lupa Searches for a Sales Team

Lupa Agents take a company’s scientific context once, then rank markets, labs and partnerships so a seller can act on a scored list instead of a pile of PDFs. A second door on the same data layer is aimed at researchers who want traced answers from their own collections, citation graphs, and government databases. The company calls that connected layer the Substrate, kept current with live search rather than a snapshot that aged in a crawl.

The product page walks through a specimen that is worth taking at face value as an interface, not as a new finding about biology: a traced note on cryo-EM structure depositions, with 7,200-plus structures a year, 50,000-plus datasets, and 12,000-plus patents, then a “Takagi Lab” card marked a qualified lead with score 92, checked across 98 sources. The point of the mock card is the score and the source trail, which is the thing a hardware seller can hand to a colleague without hiding the homework.

LUPA’S LIVE SOURCES

  • Papers and grants: Publications, grant records, and funding news the agents read the way an analyst would.
  • Patents and firms: Patent files plus company and lab records, tied back to the document they came from.
  • People networks: Citation graphs and co-author webs that name researchers a CRM never stored.
  • Open web refresh: Live search on top of the Substrate, so the list is not frozen from last quarter.

The hard part is not collecting those files. Anyone can dump patents into a model and watch it invent a market. The open worry around this kind of tool is how it ranks noisy patent and market records without sending a sales team on junk calls, and Lupa’s answer is boring on purpose: every result is supposed to be followable to a publication, a patent, or a record, and contradictions are supposed to surface when sources disagree. If that holds, a four-person shop can sell trust. If it slips, the ranked list is just a faster way to be wrong.

Stop Hunting Labs by Job Title

A photonics seller cannot type “head of optics” into a standard lead tool and expect the right lab. The buyer is often a postdoc, a principal investigator, or a process engineer whose purchasing power is buried in a paper, a grant, or a methods section. Job titles lag the science by years, and the useful signal is a protocol, a wavelength, or an animal model, not a seniority tag.

That is why Lupa’s first users were hardware companies rather than university licensing offices. A tech-transfer manager wants licensees. A Menhir or MaxWell seller wants the next ten groups whose work already implies the product, which is a different query run on the same messy record of science. Blyumin’s ETH years in tribology and quantum sensing, and Kempin’s two decades in corporate finance and business development at UBS before he left to build Lupa full-time, map onto that desk: one founder who lived the lab pain, one who knows how a commercial team buys a tool.

Carvalho spent a decade building AI systems for hospitals, universities, governments, and companies. Van Gelderen is the full-stack engineer who has to make those ranked lists load as a product, not a slide. The Stage II cash was meant to move them from founder-led selling toward a product-led motion and a larger scientific knowledge graph. Stage III, per the announcement, is for scaling in deep-tech markets, opening individual users, and pushing customer acquisition, which is the same wedge with a wider door.

A Different Seat From PatSnap and Derwent

The crowded shelves in this category are not empty. PatSnap, founded in Singapore in 2007, sells Eureka agents and patent analytics to R&D and IP teams, with Pro plans from $100 a month and a corpus it puts at 215 million-plus patents and 260 million-plus papers. Clarivate’s Derwent and Questel’s Orbit still own the professional search desk. Newer agent layers such as Cypris pitch landscape, freedom-to-operate, and scouting workflows with citations for corporate R&D.

Those seats are priced and staffed for patent counsel, information scientists, and central innovation groups. They are not a Swiss instrumentation firm’s weekly prospecting list. Lupa is trying to sell the same underlying records, papers, patents, grants, labs, to a revenue team that will only keep paying if the ranked names convert. Traceable sources are the procurement language that lets a cautious scientist-seller defend a new tool to a CFO who has already been burned by a black-box demo.

The overlap is real, and it will get louder if Lupa’s individual-user plan works. A cheap, sourced agent that names labs is a threat to the lower rungs of tech-scouting retainers, not to a global IP department that still needs family data, claim charts, and a human searcher. For now the company is not picking a fight with Derwent. It is picking off the hardware sellers those platforms were never asked to serve.

Kickfund Can Add CHF 850,000 After the Jury

Stage III’s quiet clause is access, not the CHF 150,000 label. Teams that clear the last jury gain a shot at Kickfund, an additional equity opportunity of up to CHF 850,000. Stage 2 winners may also apply for a CHF 150,000 InnoBooster grant. Stacked with the three kicks, Venture Kick markets a path of up to CHF 1,150,000, which is the number that matters if Lupa’s next round is a priced equity deal rather than another prize night.

The loans convert at the next significant equity round, so the Stage II and Stage III paper is already waiting for a valuation. That is a feature of the Swiss model, and it is also a clock. Thirty paying customers give the team something to take into that conversation that a slide about “bridging science and industry” never would.

Venture Kick’s January 29, 2026 release said private donors have put CHF 102 million into 1,251 startup projects since 2007, with 16,100 jobs attached to the portfolio. In 2025 it backed 130 new projects with CHF 8.3 million directly and said those companies attracted CHF 1.25 billion in follow-on financing. Survival of incorporated Kick companies is put at 84 percent, and 32 percent of founding teams include women. Lupa is one more file in that stack, with a woman CEO and a product already charging.

FROM PRISMUS TO STAGE III

  1. 2025: Founds the company in Zurich as Prismus, later listed at two to ten people.
  2. May 7, 2026: Collects the CHF 40,000 Stage II award after the first three paying Swiss deep-tech clients.
  3. August 2026 onward: Ships the Lupa site and names hardware users in photonics and life science tools.
  4. October 29, 2026: Clears Stage III under the Lupa name, with 30 paying customers and a plan to add individual seats.

The rebrand is already in the award title. Searchers still hitting Prismus will land on Lupa, which is a small tax on a young company and a sign the product, not the legal name, is what the jury was buying.

Inside the Andreasturm Beta

The team has moved into the Andreasturm with the ETH Entrepreneur Club and other startups, with the ETH AI Center a couple of floors up in the same tower. Kempin has been on the road in Lausanne around EPFL and Biopôle, and the company showed up at ECIO in Zurich among integrated-optics and silicon-photonics groups, the rooms where a hardware seller actually stands next to a potential lab customer.

After more than a year of building, Lupa opened a beta limited to the first 1,000 sign-ups, asking new users to show what is missing. Stage III money is earmarked to scale deep-tech adoption and to onboard those individual users, which is a different motion from closing another instrumentation firm. A named seat for a single scientist is how the Substrate either becomes a habit or stays a boutique login for 30 commercial teams.

LUPA IN FIGURES

  • Paying customers: 30, according to the Stage III announcement.
  • Beta cap: First 1,000 sign-ups on the public waitlist.
  • Founded: 2025 in Zurich, now operating as Lupa.
  • Next money on paper: Kickfund access up to CHF 850,000, plus an InnoBooster application path of CHF 150,000.

The Stage III cash will not staff an enterprise patent army. It will keep a small Zurich team shipping ranked, sourced names to hardware sellers who already pay, and it will test whether a single researcher will log in for the same trail. Kickfund, at up to CHF 850,000, is the larger cheque that last jury made reachable.

Harry is the editor of IAQABA, an independent publication he owns and runs. A decade in journalism, beginning as a reporter and now as the editor of his own titles, has left him with a clear test for what deserves a story: it has to change what a reader knows or decides, and it has to rest on something he can point to. That rules out recycled press releases, forecasts with no data behind them and rumours that no document supports. It leaves room for a great deal, and the site covers news, business, science and technology alongside sports, entertainment and lifestyle, with travel, auto and gaming given the same standard rather than lighter treatment. Sources are primary wherever possible: the regulator's filing, the company's own statement, the transcript, the dataset, or the product on Harry's desk. Figures are checked before they are published and rechecked if a reader questions them. Mistakes are corrected under a published policy. Readers across the world can reach him directly at support@iaqaba.com.

Click to comment

Leave a Reply

Your email address will not be published. Required fields are marked *

Trending