The Nelson Advisors Guide for Founders to HealthTech and MedTech Success in 2026: 10 European Case Studies

Nelson Advisors
4 hours ago
36 min read

Executive summary

Europe has quietly become one of the most productive regions on earth for building health and medical technology companies of genuine scale. Over the past decade a cohort of founders has turned the continent's structural disadvantages, fragmented markets, multiple regulators, conservative public payers, slow procurement, into the very foundations of durable, defendable businesses. This guide examines ten of them in depth.

The companies profiled here are deliberately varied. They span pure software platforms (Doctolib, Ada Health, Kry/Livi), deep-science instrumentation (Oxford Nanopore), surgical robotics (CMR Surgical), applied genomics and AI biology (SOPHiA GENETICS, Owkin), clinical-AI infrastructure (Corti), tech-enabled care delivery (Cera) and preventive consumer medicine (Neko Health). Between them they have raised well over five billion euros from grants, venture capital, private equity, strategic pharmaceutical partners and the public markets. Some are now profitable; several are not; one cautionary comparator in this report went bankrupt after a multi-billion-dollar valuation.

Three themes run through every case.

First, the most defendable moats in European health technology are built not from clever algorithms but from proprietary data, regulatory clearance and entrenchment in the workflows of clinicians and payers, assets that take years to accumulate and are extremely hard to copy.

Second, the distinction between being “AI-native” and an “AI enabler” matters far less than founders assume: what matters is whether AI is layered on top of a genuine distribution and data advantage.

Third, the 2022–24 funding winter separated the survivors from the casualties almost entirely on the basis of capital discipline and proximity to whoever actually pays for healthcare.

This is not a ranking. It is a working reference for founders and boards: a set of strategies, tactics, funding journeys, critical decisions and hard-won lessons drawn from companies that have already walked the path you are on. Read it as a map of what has worked, what has nearly killed otherwise excellent businesses, and where the genuinely defensible value in this market sits in 2026.

How to read this guide

▸ Each case study opens with a fact box (HQ, founders, funding, valuation, AI posture) followed by analysis of strategy, scalability, moat, risks, critical decisions and lessons.

▸ Funding figures are drawn from company announcements and reputable reporting to mid-2026; private valuations are last-round marks and may be stale. Treat them as directional, not audited.

▸ Cross-cutting chapters after the case studies synthesise the patterns: AI-native versus AI-enabler, moat construction, the funding map, and the risks that have sunk well-capitalised peers.

The 2026 landscape: why these lessons matter now

Founders entering HealthTech and MedTech in 2026 face a market that looks superficially similar to 2021 but behaves completely differently. The exuberant capital of the pandemic era has gone; in its place is a more demanding investor base that wants evidence of clinical impact, a credible path to profitability and a defensible position before it writes large cheques. At the same time, the underlying demand has never been stronger. Ageing populations, chronic disease burden, workforce shortages across European health systems and the maturation of artificial intelligence have combined to make health one of the few categories where technology can simultaneously cut cost and improve outcomes.

The opportunity is therefore real, but the bar is higher. The companies in this report succeeded not because they rode a hype cycle but because they solved a structural problem for a payer, a clinician or a patient and then made that solution progressively harder to displace. Understanding how they did it and where comparable companies failed, is the single most useful preparation a founder can do.

A simple framework: scalable, sustainable, defendable

Throughout this guide we assess each company against three properties that, in our advisory work, separate enduring health-technology businesses from those that raise large rounds and then stall.

• Scalable — can the business add revenue faster than it adds cost? Software platforms scale through near-zero marginal cost; hardware and care businesses scale through repeatable unit models, manufacturing capacity or acquisition roll-ups. The question is always: what is the unit, and does the next unit get cheaper?

• Sustainable — does the business have a route to funding itself? In a market where reimbursement is slow and capital is expensive, the survivors are those with recurring revenue, demonstrated unit economics and the discipline to reach profitability before the money runs out.

• Defendable — once the business works, how hard is it to copy? In health the durable moats are proprietary longitudinal data, regulatory clearance, entrenchment in clinical workflows, patent estates and trusted brand. Features are not moats; the assets that take years to build are.

With that framework in mind, we turn to the ten companies.

Why these ten companies

These ten were not chosen as a definitive ranking of the best European health-technology companies, nor as an investment recommendation. They were selected because, between them, they illustrate the full range of strategies, business models, funding paths and moats a founder is likely to consider, also because each has progressed far enough to offer real, tested lessons rather than early promise.

The selection deliberately spans the spectrum from pure software to deep science. Doctolib, Ada Health and Kry/Livi represent software-led platforms and digital care. Oxford Nanopore and CMR Surgical represent capital-intensive deep-science hardware. SOPHiA GENETICS and Owkin sit at the data-and-AI layer of genomics and drug discovery. Corti represents pure clinical-AI infrastructure. Cera represents technology-enabled care delivery, and Neko Health represents consumer-facing preventive medicine. The set also spans the funding journey, from companies still private and venture-backed, through those that took strategic capital from industry partners, to two that have weathered the public markets — and several European geographies, with companies anchored in France, the United Kingdom, Switzerland, Germany, Sweden and Denmark.

Just as importantly, the cohort spans outcomes. Some are profitable; several are not; valuations have risen for some and fallen sharply for others. We have resisted the temptation to present only unambiguous successes, because the most useful lessons frequently lie in the tensions, a brilliant business with a punished share price, a category creator still searching for profitability, an AI pioneer confronting a new paradigm. Read together, they form a more honest map of the territory than a list of winners would.

Part One: Ten European case studies

The ten companies are grouped loosely from broad software platforms through to deep science and consumer medicine. Each can be read independently.

1. Doctolib — the operating system for European healthcare

HQ / country	Paris, France (major hub in Berlin)
Founded	2013
Founders	Stanislas Niox-Château (CEO) and co-founders
Sub-sector	Practice-management SaaS, e-booking and teleconsultation
Funding to date	~€790m+ private; never IPO’d (Series F €500m, 2022)
Key backers	Accel, General Atlantic, Eurazeo, Bpifrance
Valuation	~€5.8bn / $6.4bn (2022, reportedly held flat in 2025)
AI posture	AI enabler — marketplace first, AI layered on from 2024

What it does and why it matters

Doctolib is the closest thing Europe has to an operating system for healthcare. Patients use it to find a practitioner, book an appointment and run a video consultation; the much more valuable side of the business sells practice-management software, scheduling, records, billing and, increasingly, AI tools, to doctors, clinics and hospitals. Founded in Paris in 2013 by Stanislas Niox-Château, it now serves more than 135,000 healthcare professionals and is the most valuable startup France has produced.

Its significance for founders is that it demonstrates how to build a category-defining platform in a regulated, fragmented, trust-sensitive market and how distribution, not technology, became the moat.

Strategy and tactics

Doctolib executed a textbook land-and-expand vertical-SaaS strategy. It began with the single sharp problem of appointment booking, then layered teleconsultation, full practice management, payments and AI on top, increasing revenue per practitioner over time. Crucially, growth was driven by a large human field-sales force that built density city by city, signing up doctors in person to overcome the trust barrier that gates technology adoption in medicine.

This created a two-sided flywheel: more doctors made the platform more useful to patients, and more patient demand pulled in more doctors. The company’s decision to run COVID-19 vaccination booking for the French state in 2020–21 cemented both its brand and its status as quasi-national infrastructure.

Scalable, sustainable, defendable

The model is highly scalable: subscriptions of roughly €109–€229 per practitioner per month carry high gross margins and low marginal cost once local density is achieved. On sustainability, 2024 annual recurring revenue reached €348m, up 22.5%, while losses were cut 38% to €53.8m, putting breakeven within reach. Defensibility comes from per-geography network effects, the high switching costs of being a practice’s system of record, GDPR-compliant health-data hosting, and a deep brand of trust.

AI-native versus AI-enabler

Doctolib is an AI enabler, not an AI-native company. It was built as a marketplace and bolted AI on from 2024 — an ambient-transcription Consultation Assistant (built on Azure OpenAI and the acquisition of Typeless) and an AI phone assistant in 2025. The instructive point is that its AI is valuable precisely because it sits on top of an existing network, distribution and a decade of appointment and clinical data. AI is the layer; the moat is underneath it.

Risks and critical decisions

The principal risks are recurring data-privacy and regulatory scrutiny, including controversy over using anonymised patient data to train AI, a continued path to profitability and roughly 80% revenue concentration in France. Three critical decisions defined the company: choosing a heavy human field-sales model over pure product-led growth; stepping up to run national vaccination booking; and the decisive 2024 pivot to position itself as an AI clinical platform ahead of a possible 2026–27 IPO.

Founder lessons from Doctolib

▸ In regulated verticals, owning the system of record and the local network beats winning on features.

▸ Boots-on-the-ground sales can itself be a moat when trust gates adoption.

▸ A crisis (COVID) can be a catalyst that converts a product into infrastructure.

▸ AI is a layer on top of distribution and data — not a substitute for them.

2. Oxford Nanopore Technologies — owning a new sequencing category

HQ / country	Oxford, United Kingdom (LSE: ONT)
Founded	2005 (University of Oxford spin-out)
Founders	Hagan Bayley, Gordon Sanghera, Spike Willcocks
Sub-sector	Genomics instrumentation and consumables (sequencing)
Funding to date	>$1.1bn pre-IPO; ~£350m raised at 2021 IPO
Key backers	IP Group, Invesco, Temasek, Nikon, Wellington
Valuation	~£1.2bn market cap (2026), down ~75% from IPO peak
AI posture	AI-enabled deep tech — ML basecalling on a physics moat

What it does and why it matters

Oxford Nanopore sells a fundamentally different way to read genetic material. Its sequencers pass single DNA or RNA molecules through protein nanopores and read the resulting electrical signal in real time, enabling portable, long-read sequencing, from the USB-stick-sized MinION to the high-throughput PromethION. Spun out of the University of Oxford in 2005 around Professor Hagan Bayley’s chemistry, it created and still leads a sequencing category distinct from the dominant short-read, sequencing-by-synthesis approach.

It matters to founders as the archetype of patient, deep-science company building: more than $1.1Bn raised over roughly sixteen years before a landmark IPO, and a moat rooted in physics, biochemistry and intellectual property rather than software alone.

Strategy and tactics

The strategy was to build one scalable sensing platform, “analysis of anything, by anyone, anywhere” and ladder products from the entry-level Flongle up to PromethION. Tactics included democratising access and seeding academic users with grant programmes (the platform now underpins more than 14,000 publications), continuous accuracy improvements, and a deliberate move up into higher-margin Clinical and BioPharma segments. From 2025 the company showed new discipline, sunsetting weaker products to focus on its profitability path.

Scalable, sustainable, defendable

Scalability comes from a common sensor across the entire product line and a razor-and-razor-blade model: consumable flow cells and reagents generate recurring revenue tied to the installed base. FY2025 revenue reached £223.9m, up around 24% at constant currency, with Clinical growing roughly 60%. With about £302m of cash at year-end 2025 and guidance for adjusted EBITDA breakeven in 2027, the business is funding its own path to sustainability. Defensibility rests on a deep, litigated patent estate covering pores, enzymes and electronics.

AI-native versus AI-enabler

Oxford Nanopore is best described as AI-enabled deep tech. The core invention is physical and biochemical, but its Dorado basecaller uses neural networks to convert raw electrical signal into sequence in real time, and machine learning is now used to design novel pore proteins. AI is mission-critical software, but it sits on a moat made of physics and biology, the inverse of a pure-software AI company.

Risks and critical decisions

The clearest risk is illustrated by the share price: down roughly 75% from its IPO peak, a reminder that a blockbuster listing is a milestone, not a finish line. The company also has meaningful exposure to US federal research funding and is in litigation with BGI/MGI. Three critical decisions shaped it: spinning out around nanopore science in 2005 as a contrarian long-horizon bet; re-engineering its pore chemistry after Illumina’s 2016 patent suit (switching to a CsgG pore) to defuse the litigation; and choosing a London listing in 2021.

Founder lessons from Oxford Nanopore

▸ Own a category rather than fight an incumbent head-on.

▸ Razor-and-blade economics turn hardware sales into recurring revenue.

▸ Patents are a living asset — be prepared to re-engineer around your own IP under litigation.

▸ Deep science needs deep, patient capital; a spectacular IPO is the start of public-market discipline, not the end of the journey.

3. CMR Surgical — challenging a monopoly in surgical robotics

HQ / country	Cambridge, United Kingdom
Founded	2014 (as Cambridge Medical Robotics)
Founders	Mark Slack, Luke Hares, Martin Frost, Paul Roberts, Keith Marshall
Sub-sector	Surgical robotics (minimally invasive soft-tissue surgery)
Funding to date	~$1.4–$1.5bn private (Series D $600m, 2021)
Key backers	SoftBank Vision Fund 2, Ally Bridge, Tencent, GE Healthcare
Valuation	$3bn (2021); explored a sale up to ~$4bn in 2025
AI posture	AI enabler — mechatronics core, AI/data features added

What it does and why it matters

CMR Surgical makes Versius, a modular, portable robotic-assisted surgery system that competes directly with Intuitive Surgical’s dominant da Vinci. Where da Vinci is a single large console-and-cart, Versius uses small, independent bedside arm carts that fit existing operating theatres and can be moved between them. Founded in Cambridge in 2014, Versius has now been used in more than 30,000 procedures across 30-plus countries.

It matters because it is a credible European challenger to one of medtech’s most entrenched monopolies — and a live lesson in the capital intensity and patience that hardware in medicine demands.

Strategy and tactics

The strategy was to win as the challenger by competing on cost, footprint and accessibility rather than trying to out-feature the incumbent. Tactically, CMR commercialised internationally first, across Europe, Latin America, Asia and Australia, becoming the second most-adopted soft-tissue robot outside the US before tackling the American market. It built recurring revenue through single-use instruments, service contracts and its Versius Connect data ecosystem, and invested in its own manufacturing capacity in Ely.

Scalable, sustainable, defendable

Scalability rests on a razor-and-blade model (system plus single-use instruments plus service) and purpose-built manufacturing capacity of around 500 systems a year. The defensibility is real but shallower than the incumbent’s: surgeon and theatre switching costs, multi-jurisdiction regulatory clearances and a growing procedure dataset, set against Intuitive’s roughly 60% global share, 9,000-plus installed systems and around 85% recurring revenue. CMR’s edge is economics and form factor, not scale.

AI-native versus AI-enabler

CMR is firmly an AI enabler, not AI-native. The core value is mechatronics, robotics and surgeon ergonomics; the robot is teleoperated, not autonomous. AI and data features, procedure analytics, the digital ecosystem, near-infrared imaging, are add-ons layered onto excellent hardware.

Risks and critical decisions

The risks are formidable: a dominant incumbent, deep-pocketed new US entrants (Medtronic’s Hugo, J&J’s Ottava), high cash intensity, slow surgical sales cycles, a 2025 reliance on debt and a valuation that has been flat since 2021. Three critical decisions stand out: the founding bet on a modular, portable architecture; the decision to internationalise before entering the US; and new leadership’s high-conviction call to scrap the planned first-generation US launch and wait for the cleared Versius Plus, vindicated by FDA clearance in December 2025.

Founder lessons from CMR Surgical

▸ Against an entrenched incumbent, differentiate on architecture and economics — don’t try to out-feature them.

▸ Sequence your markets: prove the product in less-defended geographies before the hardest one.

▸ Build recurring revenue into hardware from day one.

▸ Be willing to delay a flagship launch for the right product — and respect that hardware in medicine is brutally capital-intensive.

4. SOPHiA GENETICS — owning the genomic data layer

HQ / country	Lausanne, Switzerland / Boston, USA (NASDAQ: SOPH)
Founded	2011 (EPFL spin-out)
Founders	Jurgi Camblong, Pierre Hutter, Lars Steinmetz
Sub-sector	Genomic data analytics / clinical bioinformatics SaaS
Funding to date	~$250m pre-IPO; $234M gross at 2021 NASDAQ IPO
Key backers	Balderton, Generation IM, aMoon, Hitachi Ventures
Valuation	~$330–360M market cap (2026), down ~70% from IPO
AI posture	AI-native within its niche (proprietary ML is the product)

What it does and why it matters

SOPHiA GENETICS sells SOPHiA DDM, a cloud-based, AI-powered platform that helps hospitals and laboratories analyse and interpret complex genomic and multimodal data for precision medicine, principally in oncology. It sits one layer above the sequencing hardware companies, turning raw sequencer output into clinically actionable insight. Founded in 2011 as an EPFL spin-out, its platform has analysed more than 2.5 million cases across some 70 countries.

It matters as a rare European deep-tech company that listed on NASDAQ, and as a case study in owning the data and analytics layer of a value chain rather than the hardware.

Strategy and tactics

The strategy is a decentralised “data-as-a-network” model: hospitals sequence locally and analyse in SOPHiA’s cloud, so each analysis enriches a shared dataset that improves the algorithms for everyone. Tactics include classic land-and-expand SaaS economics, licensing marquee assays (notably Memorial Sloan Kettering’s MSK-ACCESS and MSK-IMPACT), a 2026 joint venture with MSK in multimodal precision oncology, and monetising the platform with biopharma partners.

Scalable, sustainable, defendable

The cloud SaaS model is scalable with improving margins (adjusted gross margin around 74%), and revenue reaccelerated to roughly $77m in FY2025, up about 19%. It is not yet self-sustaining, net losses remain large, with adjusted EBITDA breakeven targeted by end-2026. Defensibility comes from a genuine data network effect, a dataset of more than two million genomic profiles, around 31 patents and deep clinical integration that creates high switching costs.

AI-native versus AI-enabler

SOPHiA is AI-native within its niche: proprietary machine learning is the product, not an add-on. It is, however, an applied AI company for genomics rather than a frontier-model builder, a useful distinction for founders, because applied, domain-specific AI on proprietary data is often more defensible than generic model capability.

Risks and critical decisions

Risks include persistent losses, a stock down around 70% from IPO, intellectual-property litigation (a European dispute with Guardant Health) and a leadership transition. Three critical decisions defined it: building a decentralised analytics layer rather than a centralised lab; listing on NASDAQ with a dual Swiss-Boston identity; and going all-in on multimodal oncology through the MSK alliance, alongside a deliberate founder-to-CEO succession in 2026.

Founder lessons from SOPHiA GENETICS

▸ Own the data layer, not just the workflow — a data network effect compounds with every customer.

▸ Partner with the prestige incumbent (MSK) for instant clinical credibility.

▸ A great moat does not guarantee a great stock; unit economics must eventually deliver.

▸ Plan IPO timing, venue and founder succession deliberately, not reactively.

5. Owkin — strategic capital and an AI-biology moat

HQ / country	Paris, France / New York, USA
Founded	2016
Founders	Thomas Clozel (CEO), Gilles Wainrib (President/CSO)
Sub-sector	AI for drug discovery, biomarkers and trial optimisation
Funding to date	~$300m+ equity (Sanofi $180m equity + $90m alliance, 2021)
Key backers	Sanofi, Bristol Myers Squibb, GV (Google Ventures), Bpifrance
Valuation	~$1bn unicorn (last mark; likely stale)
AI posture	Clearly AI-native — founded around ML on patient data

What it does and why it matters

Owkin is a French-American “TechBio” company that applies AI and federated learning to hospital data to discover drug targets and biomarkers, optimise clinical trials and build diagnostics. Federated learning lets it train models across hospital datasets without the data ever leaving the institution, a privacy-preserving approach that unlocked access to data others could not touch. It became a unicorn in 2021.

It matters as the clearest example in this report of strategic capital: rather than raising plain venture money, Owkin sold equity to the pharmaceutical giants whose problems it solves.

Strategy and tactics

The strategy is to secure privacy-preserving access to multimodal hospital data, train biology-specialised AI on it, and monetise through pharma partnerships and its own diagnostics and pipeline. Tactically, the defining move was Sanofi’s November 2021 investment of $180m in equity plus $90m for a three-year R&D alliance, the round that made Owkin a unicorn, followed by a deal with Bristol Myers Squibb. From 2025 the company pivoted toward a more scalable platform and licensing model built around an agentic AI co-pilot and a foundational biology model, and spun out its diagnostics business.

Scalable, sustainable, defendable

Scalability improves markedly with the shift from bespoke services to a licensable software platform. Sustainability is trending the right way, revenue around $86m and roughly doubling, with about $200m of cash and shrinking losses, though the business remains pre-profit and capital-intensive. Defensibility rests on a decade of hospital trust and data access (including a proprietary multi-omics cancer atlas), federated-learning IP and credibility, and embedded relationships with pharma partners who are also shareholders.

AI-native versus AI-enabler

Owkin is unambiguously AI-native, it was founded around machine learning on patient data and is now building foundational biological models. For founders this case shows both the upside and the dependency of being AI-native: the technology is the company, which makes the data and partnerships that feed it existential.

Risks and critical decisions

Risks include admitted over-hiring after its mega-rounds, concentration in a few large partners, a potentially stale billion-dollar valuation (down-round risk), an unproven drug pipeline and a frontier-model arms race against far larger players. Three critical decisions defined it: taking Sanofi’s strategic equity and alliance over pure VC; pivoting from services toward a licensable platform; and spinning out diagnostics to refocus the core.

Founder lessons from Owkin

▸ Strategic capital — equity plus data plus credibility from an industry partner — can beat plain VC, but it adds dependency.

▸ A regulated-industry data moat is built slowly and is correspondingly hard to copy.

▸ Be honest about over-hiring after mega-rounds; capital can mask a lack of focus.

▸ Evolve from bespoke services toward scalable software, and divest non-core assets to stay focused.

6. Ada Health — turning regulation into a moat

HQ / country	Berlin, Germany
Founded	2011 (consumer app launched 2016)
Founders	Claire Novorol, Martin Hirsch, Daniel Nathrath (CEO)
Sub-sector	AI symptom assessment, triage and care navigation
Funding to date	~$190m+ (Series B $90m, 2021; €30m venture debt 2023)
Key backers	Access Industries, Leaps by Bayer, Samsung Catalyst Fund
Valuation	~$600m last reported (2022); likely repriced lower
AI posture	AI-native, but pre-LLM expert-system AI

What it does and why it matters

Ada Health builds an AI-powered symptom assessment, triage and care-navigation platform. A consumer app launched in 2016 lets users describe symptoms and receive a probabilistic assessment and guidance on what to do next; by late 2024 it had handled more than 35 Million assessments. Founded in Berlin (and not to be confused with the unrelated Canadian customer-service company Ada Support), it is one of Europe’s most-used digital triage tools.

It matters as a case study in converting regulatory clearance into a durable moat, and in the difficult art of monetising a free consumer health product.

Strategy and tactics

The central strategic move was a pivot from a free consumer app, which built brand, reach and a vast dataset, to an enterprise and B2B2C model, selling “Ada Assess” to health systems, payers, pharma and governments including Bayer, Novartis, Pfizer and US health systems. After the 2022 downturn the company repositioned hard around profitability and clinical credibility, and used venture debt to extend runway with less dilution.

Scalable, sustainable, defendable

The product is highly scalable: software-only AI triage with near-zero marginal cost, available in multiple languages. On sustainability, the company stated it reached profitability at the end of 2023 with 260% year-on-year revenue growth, driven by enterprise contracts. Defensibility comes from a physician-curated medical knowledge base, EU-MDR Class IIa certification and ISO 13485, a trusted consumer brand and 35 million-plus structured real-world assessments.

AI-native versus AI-enabler

Ada is AI-native but in a specific, instructive sense: it was built around a proprietary probabilistic medical-reasoning engine, not a bolt-on, yet that engine predates the generative-LLM wave. Large language models are now both a competitive threat and a strategic question for the company, illustrating that being AI-native at one moment does not guarantee defensibility against the next paradigm shift.

Risks and critical decisions

Risks include LLM-driven disruption (ChatGPT, Google and others), medical-advice liability, thin consumer monetisation and enterprise concentration. Three critical decisions defined it: the early pivot from a doctor-facing network to a consumer symptom checker; the later pivot to a monetised enterprise model prioritising profitability after 2022; and the early investment in EU-MDR certification, which turned a compliance cost into a competitive moat.

Founder lessons from Ada Health

▸ In regulated markets, certification and clinical validation are a durable moat, not just a cost.

▸ A free consumer product can be a data-and-brand engine — but it rarely funds deep tech. Find the enterprise payer.

▸ In a downturn, choose profitability over vanity growth.

▸ Being AI-native at one moment is not permanent defensibility — you must actively renew it against paradigm shifts such as LLMs.

7. Kry / Livi — discipline and the hybrid care model

HQ / country	Stockholm, Sweden (Livi in UK and France)
Founded	2014
Founders	Johannes Schildt (CEO), Josefin Landgård and co-founders
Sub-sector	Digital-first primary care / telehealth (hybrid)
Funding to date	~$700m total (Series D ~$300m, 2021); no IPO
Key backers	Accel, Index Ventures, Ontario Teachers’, CPP Investments, Fidelity
Valuation	~$2bn last mark (2021); likely stale post-repricing
AI posture	AI adopter — clinical-ops company, AI as a margin lever

What it does and why it matters

Kry, branded Livi in the UK and France, is a digital-first primary care company offering app-based video consultations alongside owned physical clinics, a hybrid “phygital” model across Northern Europe. Founded in Stockholm in 2014, it now serves more than seven million patients with over 1,300 clinicians.

It matters above all as a survival story: Kry came through the telehealth crash that bankrupted Babylon Health, and did so through capital discipline rather than momentum.

Strategy and tactics

The strategy combined digital and physical care, the app as a funnel, owned clinics for higher-value and chronic care, with multi-country expansion built on genuine localisation rather than copy-paste (per-visit payment in Sweden versus free-at-point-of-use in the NHS). Tactics included expanding wallet share into mental health, weight management and occupational health, acquiring clinics, and anchoring revenue to public payers with a cost-saving argument of roughly 40% versus in-person care.

Scalable, sustainable, defendable

The model proved scalable, the patient base grew roughly a hundredfold in five years and AI now cuts clinician administrative time by around 30%. On sustainability, all of its markets were individually profitable by end-2024, with group profit deliberately delayed to fund generative-AI investment; FY2024 revenue was around €220m, up 13%. Defensibility comes from an employed clinician network (quality control versus US-style marketplaces), public-payer and government contracts, multi-country regulatory know-how and owned clinic assets.

AI-native versus AI-enabler

Kry is an AI adopter, not AI-native. It was built as a telehealth and clinical-operations company, with AI layered in for efficiency. Tellingly, its CEO has publicly framed generative AI as an overhead that is delaying group profitability, a candid view of AI as a margin lever rather than the core product.

Risks and critical decisions

Risks include the commoditisation of video consultation (now a feature inside larger platforms such as Doctolib), roughly 78% revenue concentration in Sweden, and dependence on reimbursement policy. Three critical decisions defined it: the 2022 pivot from hyper-growth to profitability, cutting around 400 jobs and exiting the German consumer market — which arguably saved the company; the hybrid owned-clinic strategy; and the deliberate decision to delay group profitability to fund AI.

Founder lessons from Kry / Livi

▸ Multi-country health expansion needs real localisation, not copy-paste.

▸ Discipline beats momentum in a downturn — Kry survived where Babylon failed.

▸ Own the hard, defensible parts (employed clinicians, clinics), and anchor to whoever actually pays.

▸ Treat AI as a margin lever and be transparent with investors about the trade-off.

8. Cera — a data moat from owning care delivery

HQ / country	London, United Kingdom
Founded	2015–16
Founders	Dr Ben Maruthappu (CEO), Marek Sacha
Sub-sector	Digital-first home healthcare / care delivery
Funding to date	~$571m (£450m) across ~13 rounds; latest >$150m, 2025
Key backers	Kairos, BDT & MSD Partners, Schroders Capital, Vanderbilt Endowment
Valuation	Unicorn (>$1bn) as of 2025; exact figure undisclosed
AI posture	AI enabler evolving to AI-led — care operator first

What it does and why it matters

Cera is Europe’s largest digital-first home healthcare provider, delivering AI-powered care, nursing and telehealth in patients’ homes as an alternative to hospital and residential care. Founded in 2015–16 by NHS doctor Ben Maruthappu after his mother received fragmented post-fracture care, it employs more than 10,000 carers and nurses, works with over 150 local authorities and around two-thirds of NHS Integrated Care Systems.

It matters because, unusually for the sector, Cera is profitable, and because its decision to own care delivery created a proprietary data moat that pure-software rivals cannot replicate.

Strategy and tactics

The strategy is to replace expensive hospital and residential beds with technology-enabled care at home. Cera built an actual care-delivery business, it employs carers and layered a proprietary digital platform on top, using AI to predict deterioration and falls from daily carer-collected data and intervene early (it claims to predict up to around 80% of deterioration and cut hospitalisations by roughly a third). It grows by acquiring and rolling up fragmented home-care agencies and anchors revenue on recurring public-sector contracts.

Scalable, sustainable, defendable

Scalability comes from combining software with an acquisition roll-up across a highly fragmented market. On sustainability, Cera reported being EBITDA-positive in 2023 and free-cash-flow positive in 2024, exceptionally rare for a healthtech scale-up, which de-risked its 2025 unicorn round of more than $150M.

Defensibility rests on the proprietary longitudinal dataset generated by millions of daily home visits, deep entrenchment with NHS and local-authority commissioners, and scale as the UK’s largest operator.

AI-native versus AI-enabler

Cera is an AI enabler evolving toward AI-led. Its core business is physical care delivery; AI is layered on to optimise it through predictive risk, scheduling and documentation. It markets itself as “AI-led home healthcare,” but the underlying model is operational first and algorithmic second and that operational layer is precisely what generates the data the AI needs.

Risks and critical decisions

Risks include heavy dependence on public-sector budgets and contract renewals, margin pressure in low-margin care, labour intensity and wage inflation, scrutiny of its AI claims, and integration risk from acquisitions. Three critical decisions defined it: building a care operator rather than a pure software platform; anchoring on public-sector contracts for defensive recurring revenue; and prioritising profitability over pure growth, which gave it negotiating leverage and resilience.

Founder lessons from Cera

▸ Owning the full delivery stack creates a proprietary data moat that software-only rivals lack.

▸ In healthcare, demonstrable cost savings to the payer is the strongest sales argument.

▸ Reaching profitability buys leverage and resilience in a sector littered with cash-burning failures.

▸ Roll-up of a fragmented market can be a legitimate scaling engine — if integration is disciplined.

9. Corti — AI infrastructure for healthcare

HQ / country	Copenhagen, Denmark
Founded	2014–16
Founders	Andreas Cleve (CEO), Lars Maaløe (CTO) and co-founders
Sub-sector	Clinical-AI infrastructure / ambient clinical intelligence
Funding to date	~$103m (Series B $60m, 2023)
Key backers	Prosus Ventures, Atomico, Eurazeo, EIFO
Valuation	Undisclosed
AI posture	Genuinely AI-native — healthcare-specific foundation models

What it does and why it matters

Corti builds healthcare-specialised AI infrastructure, foundation models and APIs for speech-to-text, text generation, agentic workflows and clinical documentation, that act as a real-time co-pilot during clinical conversations. It first became known for AI that detects cardiac arrest during emergency calls. Founded in Copenhagen, its technology now powers applications serving more than 100 Million patients a year and over a million interactions a week, including in the NHS.

It matters as one of the clearest AI-native companies in this report, and as a model for scaling through infrastructure rather than selling a single product.

Strategy and tactics

Corti’s defining strategic move was to pivot from selling a finished co-pilot to becoming AI infrastructure for healthcare developers, opening its API to the world in 2025 and launching healthcare-specific foundation models trained exclusively on healthcare data, explicitly positioned against general-purpose LLMs as safer and more suitable for clinical use. Tactics include an API and SDK platform play, partnerships with EHR vendors and virtual-care platforms, and aggressive US expansion.

Scalable, sustainable, defendable

The platform model is highly scalable: others build on Corti, multiplying its reach (it reported a roughly fortyfold increase in projects built on its platform over two quarters). Sustainability is less proven, reported revenue of around $13M in 2025 is modest against the capital raised, so the infrastructure pivot must monetise. Defensibility comes from nearly a decade of peer-reviewed research and proprietary clinical training data drawn from millions of real patient interactions, a safety and trust advantage in a domain where general models are risky.

AI-native versus AI-enabler

Corti is genuinely AI-native: the entire company is built around proprietary healthcare AI models, and AI is the product rather than an add-on. It is the strongest example in this report of a company whose moat is domain-specific AI, although, importantly, the moat is the proprietary clinical data and published evidence, not the model architecture itself.

Risks and critical decisions

Risks include intense competition from well-funded rivals (Microsoft/Nuance DAX, Abridge, Suki, Nabla) and from general-purpose LLM providers moving into healthcare, clinical-safety and liability burdens, long enterprise sales cycles and modest revenue relative to capital. Three critical decisions defined it: the life-or-death origin in cardiac-arrest detection that built clinical credibility and data; the 2024–25 pivot to an open API and foundation-model business; and the deliberate choice to build healthcare-specific models rather than fine-tune general LLMs.

Founder lessons from Corti

▸ In clinical AI, proprietary domain data and published evidence are the moat — not the model architecture.

▸ Specialisation can beat general-purpose incumbents on safety and trust.

▸ Becoming infrastructure that others build on can scale reach far faster than selling one product.

▸ Watch the gap between reach and revenue: an infrastructure pivot must ultimately monetise.

10. Neko Health — brand, hardware and preventive medicine

HQ / country	Stockholm, Sweden
Founded	2018 (publicly launched 2023)
Founders	Hjalmar Nilsonne (CEO), Daniel Ek (chairman)
Sub-sector	Preventive health / consumer medtech (hardware + AI + clinics)
Funding to date	~$325m+ (Series B $260m at ~$1.8bn, 2025)
Key backers	Lightspeed, General Catalyst, Lakestar, Atomico
Valuation	~$1.8bn (January 2025)
AI posture	Hardware-and-AI-native — proprietary scanner plus AI analysis

What it does and why it matters

Neko Health operates AI-powered, non-invasive full-body scanning clinics that capture millions of data points, skin and moles, cardiovascular and metabolic markers, in minutes, for early detection and prevention. Co-founded in 2018 by Hjalmar Nilsonne and Spotify’s Daniel Ek, it scanned roughly 10,000 patients across Stockholm and London by early 2025, with a waitlist that grew past 100,000.

It matters as a high-conviction bet on shifting healthcare from treatment to prevention, and as a study in how a trusted founder brand and a premium consumer experience can generate demand at remarkable speed.

Strategy and tactics

The strategy is to make preventive, broad-spectrum screening fast, affordable, convenient and consumer-friendly. Tactics include building proprietary scanning hardware and AI analysis in-house (vertical integration), a premium consumer-pay model with a curated, retail-quality clinic experience, demand generation through waitlists and Ek’s brand halo, deliberate one-market-at-a-time geographic rollout, and reinvestment in diagnostics to widen what the scans can detect.

Scalable, sustainable, defendable

Scalability comes from a repeatable clinic-plus-standardised-hardware unit model, where each new clinic replicates a proven format. Sustainability is supported by direct consumer revenue (around £299 per scan), reducing reliance on slow public payers, though the model is capital-intensive. Defensibility rests on proprietary scanner hardware and IP, a strong consumer brand, and a fast-growing longitudinal dataset that improves the AI with every scan, a data network effect.

AI-native versus AI-enabler

Neko is best described as hardware-and-AI-native. It was built around proprietary scanning hardware and AI from day one, with AI interpreting the data. It is not a pure-software AI company, but AI is core to the product, sitting between Cera’s adopter stance and Corti’s pure AI-native model.

Risks and critical decisions

The defining risk is clinical: preventive whole-body screening attracts overdiagnosis and false-positive criticism from clinicians, and faces variable regulation across markets including the US FDA. The model is also capital-intensive and reliant on out-of-pocket demand. Three critical decisions stand out: building proprietary scanning hardware in-house rather than assembling off-the-shelf devices; choosing a consumer-pay premium model over chasing public payers first; and a deliberate, evidence-led geographic expansion rather than blitz-scaling.

Founder lessons from Neko Health

▸ A trusted founder brand and a premium experience can build demand extraordinarily fast in healthcare.

▸ Owning the hardware creates a durable moat and proprietary data.

▸ Preventive screening lives or dies on clinical evidence and avoiding overdiagnosis — earn regulator and clinician trust early.

▸ Consumer-pay can sidestep slow public payers, but it is capital-intensive and demand-sensitive.

Part Two: The patterns that decide success

AI-native versus AI enablers: a distinction founders over-weight

It has become fashionable to divide health-technology companies into the “AI-native”, built from the ground up around machine learning, and “AI enablers” that adopt AI to improve an existing business. Our ten companies map cleanly onto a spectrum. Owkin and Corti are genuinely AI-native, with foundational models at their core. SOPHiA and Ada are AI-native within their niches, built around proprietary engines, though Ada’s predates the LLM era. Neko is hardware-and-AI-native. Oxford Nanopore is AI-enabled deep tech, with neural-network basecalling on a physics moat. And Doctolib, CMR Surgical and Cera are AI enablers, platforms, robots and care operations with AI layered on top.

The striking conclusion is that the AI-native label correlates poorly with durability. Several of the most defendable businesses here, Doctolib, Cera, CMR, are emphatically not AI-native, while at least one cautionary tale in the next chapter was an AI-native company whose AI was over-hyped. What consistently matters is not whether AI is the origin of the company but whether AI sits on top of a genuine, hard-to-copy advantage in data, distribution, regulation or hardware.

This has a direct implication for founders. The right question is not “am I an AI company?” but “what is my durable advantage, and does AI compound it?” In every successful case here, the AI is valuable precisely because it is fed by proprietary data or embedded in a workflow the company already owns. Where AI is the only differentiator, it is rarely a moat: in 2026, general-purpose models are cheap, capable and improving fast, and a thin AI wrapper is the most easily disrupted position in the market.

How European health technology moats are actually built

Across all ten companies, four moat types recur, almost always in combination. Understanding which ones you are building and how long each takes, is central to a defensible strategy.

• Proprietary, longitudinal data. The single most common moat. SOPHiA’s two-million-profile dataset, Cera’s daily home-visit data, Corti’s clinical interactions, Owkin’s federated hospital data and Neko’s scan dataset all improve their products with scale and cannot be bought off the shelf. The key insight: data moats are usually a by-product of owning a workflow or delivery channel, not a thing acquired directly.

• Regulatory clearance and clinical validation. Ada’s EU-MDR certification, CMR’s FDA De Novo and 510(k) clearances and Oxford Nanopore’s clinical validation are assets, not just compliance. They take years and capital to obtain, which is exactly what makes them defensible against fast followers.

• Workflow entrenchment and switching costs. Doctolib as a practice’s system of record, Kry’s employed-clinician network, CMR’s trained surgeons and Cera’s commissioner relationships all make the company painful to rip out once embedded. In health, where safety and continuity matter, switching costs are unusually high.

• Intellectual property and deep science. Oxford Nanopore’s litigated patent estate and CMR’s modular-architecture IP protect genuinely hard engineering. This moat is strongest where the underlying science is difficult and patient capital has funded a long lead.

The pattern is that the strongest companies stack several of these. Doctolib combines network effects, switching costs, data and brand; Oxford Nanopore combines IP, razor-and-blade economics and an ecosystem of published users. A single moat can be eroded; a stack of complementary moats is what produces durability. Founders should map their intended moats explicitly and ask, for each, how many years and how much capital a well-funded competitor would need to replicate it.

The funding map: grants, venture, private equity and public markets

The capital journeys of these ten companies reveal as much about strategy as their products do. Several lessons stand out for founders planning their own financing.

Grants and non-dilutive capital play a real but limited role. Oxford Nanopore benefited from early NIH and Innovate UK support and Owkin from French sovereign backing through Bpifrance, but in every case grants were catalytic seed money, not the engine.

For deep-science companies they de-risk the earliest, least investable phase; founders should pursue them aggressively but not mistake them for a business model.

Venture capital remains the dominant fuel, and the rounds here are large: Doctolib’s €500m Series F, CMR’s $600m Series D led by SoftBank, Kry’s ~$300m Series D, Neko’s $260m Series B. The clear pattern is that hardware and deep science (Oxford Nanopore raised over $1.1bn before listing; CMR around $1.4–$1.5bn) require far more capital and patience than software. Founders in capital-intensive categories must plan for a decade-long financing arc and choose investors who can sustain it.

Strategic and quasi-private-equity capital is a distinctive European feature. Owkin’s decision to sell equity to Sanofi and Bristol Myers Squibb, the customers whose problems it solves, brought data, credibility and distribution alongside money, at the cost of dependency. Eurazeo’s lead of Doctolib’s 2022 round and pension funds such as Ontario Teachers’ and CPP Investments backing Kry show how growth and PE-style capital now anchor the late-stage European market.

The public markets are the hardest lesson. Of the two companies here that listed, Oxford Nanopore and SOPHiA, both saw their shares fall roughly 70–75% from peak despite sound underlying businesses. The message is not that listing is wrong, but that an IPO is the beginning of public-market discipline and a milestone valuation must then be grown into, not celebrated as an exit.

Risks and rewards: lessons from the casualties

Every company in this report carries real risk, regulatory scrutiny, payer dependence, capital intensity, competition from better-funded incumbents and, for the AI-natives, the relentless advance of general-purpose models. But the sharpest lessons come from the companies that failed, and three cautionary cases should be studied by every founder.

• Babylon Health. The headline collapse. An AI symptom-checker and telehealth company valued at $4.2bn at its 2021 SPAC listing, bankrupt and sold for parts by 2023. It over-hyped AI that was reportedly closer to a decision tree, took on a ruinous US value-based-care model that paid out almost as much in medical claims as it earned, and expanded recklessly across the UK, US and Africa before reaching profitability.

• Pear Therapeutics. The first FDA-cleared prescription digital therapeutics company, bankrupt in 2023 with assets auctioned for a few million dollars. The lesson is brutal: regulatory clearance does not equal reimbursement. Payers would not pay, so there was no viable revenue model despite a validated product.

• Olive AI. A healthcare-automation unicorn that raised around $400m at a $4bn valuation and wound down in 2023. It over-promised on AI-driven savings, expanded too fast and failed to deliver — destroying trust and capital.

The common thread is unmistakable: inflated AI claims without clinical evidence, ignoring the reimbursement and payer reality, and scaling before proving unit economics. These are precisely the traps the survivors in this report avoided. Kry pivoted to discipline where Babylon chased growth. Cera and Ada reached profitability where Olive burned cash.

Neko is deliberately building clinical evidence to avoid the over diagnosis critique. The reward for getting this right is enormous, the companies here have created billions in value and, in several cases, genuine improvements in care, but the asymmetry is severe, and the graveyard is full of well-funded companies that confused a large round for a durable business.

The critical decisions that recur

Reading across thirty-odd critical decisions made by these founders, a handful recur often enough to count as patterns worth internalising.

• Choosing distribution as a moat. Doctolib’s human field sales and Kry’s employed clinicians were expensive, unfashionable choices that became the moat.

• Sequencing markets deliberately. CMR proved Versius internationally before the US; Neko expanded one city at a time; Kry localised rather than copy-pasting.

• Owning the hard part. Cera owns care delivery, Neko owns the hardware, Owkin owns the data access, each chose the difficult, defensible layer over the easy, copyable one.

• Pivoting decisively under pressure. Ada and Kry pivoted to profitability in the 2022 downturn; Owkin and Corti pivoted from services to scalable platforms; CMR delayed a launch for the right product.

• Treating regulation and IP as strategy. Ada’s certification, CMR’s clearances and Oxford Nanopore’s re-engineering around a patent suit were all offensive moves, not defensive chores.

Commercial models: who pays and how you reach them

If the moat determines whether a company survives, the commercial model determines how fast it grows and how much capital it consumes along the way. The ten companies here use markedly different routes to revenue, and the choice of model is one of the most consequential a founder makes.

Three broad commercial archetypes appear. The first is recurring software subscription, used by Doctolib and Ada Health, where practitioners or enterprises pay a monthly or annual fee. This model carries the highest gross margins and the cleanest scalability, but it requires solving the cold-start problem of adoption in a conservative profession, which is why Doctolib invested so heavily in human sales. The second is the razor-and-blade hardware model, used by Oxford Nanopore and CMR Surgical, where an installed base of instruments drives recurring consumable and service revenue. This model produces durable, sticky revenue but demands enormous upfront capital and long sales cycles. The third is service or care-delivery revenue, used by Cera and Kry, where the company is paid per episode of care, usually by a public payer; margins are thinner and the business is operationally heavy, but the revenue is recurring, defensive and anchored to genuine demand.

Cutting across these is the question of who pays. In Europe the dominant payer is the state, national health systems, regional authorities and statutory insurers, which shapes everything. Public payers are slow to procure and price-sensitive, but once won they provide stable, large-volume, recurring demand and a powerful reference. Cera’s cost-saving argument to the NHS, Kry’s roughly 40% cost advantage over in-person care, and Doctolib’s role in national vaccination booking all turned public-payer alignment into a growth engine. The contrasting route is consumer-pay, chosen by Neko Health, which sidesteps slow procurement entirely but trades it for the need to generate consumer demand and the capital intensity of a clinic network. Owkin and Corti, meanwhile, sell to enterprises, pharmaceutical companies and health systems, where contracts are large but sales cycles are long and concentration risk is real.

The lesson for founders is to choose the commercial model that matches both the product and the appetite of available capital, and then to design the company around its payer from day one. The companies that struggled were frequently those that built a product first and discovered the payer later, Pear Therapeutics being the starkest example, with a cleared product and no one willing to pay for it.

The European advantage: constraints that became strengths

It is tempting to view Europe’s fragmentation, regulatory density and conservative payers purely as handicaps relative to the larger, more homogeneous US market. The companies in this report suggest a more nuanced picture: several of the continent’s apparent disadvantages, handled well, became sources of durable competitive strength.

Fragmentation forced discipline. Because there is no single European market, companies such as Kry, Doctolib and CMR Surgical had to master genuine localisation, different languages, reimbursement systems, regulators and clinical cultures. That is harder and slower than scaling across fifty US states, but the resulting multi-country regulatory and operational know-how is itself a moat that new entrants cannot quickly replicate. The very friction that slowed these companies down also protected them once they had crossed it.

Strong public health systems created uniquely valuable data and demand. Europe’s universal, largely single-payer systems generate longitudinal, population-scale clinical data and a coherent buyer with a direct interest in cost reduction. Owkin’s access to leading oncology hospitals, Cera’s entrenchment with NHS commissioners, SOPHiA’s network of clinical institutions and Corti’s reach into the NHS were all enabled by the structure of European healthcare. The same systems that are slow to buy are, once aligned, extraordinary sources of data and stable demand.

World-class science and a deepening talent base supplied the raw material. Oxford Nanopore emerged from the University of Oxford, SOPHiA from EPFL, Owkin from the French AI research ecosystem and Corti from Danish machine-learning research. Europe’s universities and research institutions remain a genuine comparative advantage in deep-science health technology, and a maturing pool of operators, many of whom cut their teeth at the first wave of European technology champions, increasingly supplies the commercial talent that was historically the weak link.

None of this makes Europe an easy place to build. Capital is still thinner at the latest stages, exits are harder, and the public-market reception for Oxford Nanopore and SOPHiA was punishing. But the founders who treated Europe’s constraints as design parameters rather than obstacles built companies that are, in several respects, more defensible than a faster-moving US equivalent would have been.

Timing, sequencing and the discipline of focus

A final pattern deserves its own treatment because it appears, in some form, in almost every successful case and almost every failure: the discipline of sequencing, doing the right things in the right order and the courage to stay focused under the pressure to expand.

The casualties shared a failure of sequencing. Babylon Health scaled across three continents and into a complex US value-based-care model before it had proven profitability anywhere; Olive AI broadened its product line before it had reliably delivered the savings it promised on a narrower one. The survivors did the opposite. CMR Surgical refused to enter the US until it had a cleared, competitive product, even scrapping a planned launch to wait. Neko Health expanded one city at a time, building clinical evidence as it went. Kry cut roughly four hundred jobs and exited Germany in 2022 to concentrate on markets it could make profitable. Each of these was a decision to do less, sooner, in order to do more, later.

Timing also mattered enormously at the level of capital. The companies that raised large rounds at the 2021 peak and then spent as though the environment would persist, Owkin has been admirably candid about over-hiring, had to retrench painfully. Those that treated the downturn as a forcing function for profitability emerged stronger and, in Cera and Ada’s cases, were able to raise again from a position of strength. For founders, the practical implication is to raise when you can but spend as though you cannot, and to treat every expansion, new geography, new product line, new payer type, as a decision that must be earned by proof in the current one, not assumed by ambition.

Focus, in this market, is not a constraint on growth; it is the mechanism of durable growth. The companies that endured were those willing to be smaller than their funding allowed until each step was proven, and the discipline to sequence carefully is perhaps the single most transferable lesson in this guide.

Conclusion: a founder’s checklist for 2026

The ten companies in this guide are different in almost every respect, software and hardware, consumer and enterprise, profitable and pre-profit, AI-native and AI-enabling, yet they converge on a small set of principles that should shape any founder’s strategy in 2026.

Build your business around a durable advantage, not a feature. In European health technology that advantage is almost always proprietary data, regulatory clearance, workflow entrenchment, deep-science IP or some stack of these. Ask continually how many years and how much capital a well-funded rival would need to copy what you have built.

Be honest about what kind of company you are. If you are capital-intensive deep science or hardware, plan for a decade-long financing arc and choose patient, deep-pocketed investors. If you are software, the prize is scale and margin, but the AI layer alone will not protect you. Either way, get close to whoever actually pays for healthcare, in Europe that is usually a public payer or a clinician and prove the unit economics before you scale.

Treat AI as a compounding layer, not an identity. The most defendable companies here use AI to amplify a moat they already own. A thin AI wrapper on someone else’s data or workflow is the most easily disrupted position in the market.

Respect the asymmetry. The rewards in this market are extraordinary, but the failures, Babylon, Pear, Olive, were not small companies; they were well-funded, celebrated businesses that mistook a large valuation for a durable one. Discipline, evidence and proximity to the payer are what separated the survivors from the casualties. They will do so again in 2026.

Looking ahead, the next wave of European health-technology value is likely to accrue to companies that combine a proprietary clinical-data asset with the new generation of AI, not as a thin application layer, but embedded in workflows and regulatory pathways that took years to earn. The maturation of generative AI lowers the cost of building a feature but raises the premium on the things AI cannot easily manufacture: trusted data, clinical evidence, regulatory clearance and entrenchment with the people who deliver and pay for care. The founders who understand that distinction, who use 2026’s cheap intelligence to compound an advantage rather than to substitute for one, will build the defendable companies of the coming decade.

The ten companies in this guide have already shown how it is done. None had an easy path; several are still on theirs. But each, in its own way, made the hard, defensible, sometimes unfashionable choices that the European health-technology market rewards over time. For a founder setting out in 2026, that is the most valuable inheritance of all: not a formula, but a set of proven instincts about where durable value in this industry truly comes from.

The Nelson Advisors founder checklist

▸ Name your moat(s) explicitly — data, regulation, workflow, IP — and quantify how long they take to replicate.

▸ Match your financing plan to your capital intensity; pursue non-dilutive grants early but never mistake them for a model.

▸ Get to the payer and prove unit economics before scaling; reimbursement is as important as regulatory clearance.

▸ Use AI to compound an existing advantage; avoid being a thin wrapper on someone else’s data.

▸ In a downturn, choose profitability and discipline over momentum — it is what kept the survivors alive.

Nelson Advisors > European MedTech and HealthTech Investment Banking

Nelson Advisors specialise in Mergers and Acquisitions, Partnerships and Investments for Digital Health, HealthTech, Health IT, Consumer HealthTech, Healthcare Cybersecurity, Healthcare AI companies. www.nelsonadvisors.co.uk

Nelson Advisors regularly publish Thought Leadership articles covering market insights, trends, analysis & predictions @ https://www.healthcare.digital

Nelson Advisors publish Europe’s leading HealthTech and MedTech M&A Newsletter every week, subscribe today! https://lnkd.in/e5hTp_xb

Nelson Advisors pride ourselves on our DNA as ‘Founders advising Founders.’ We partner with entrepreneurs, boards and investors to maximise shareholder value and investment returns. www.nelsonadvisors.co.uk

#NelsonAdvisors #HealthTech #DigitalHealth #HealthIT #Cybersecurity #HealthcareAI #ConsumerHealthTech #Mergers #Acquisitions #Partnerships #Growth #Strategy #NHS #UK #Europe #USA #VentureCapital #PrivateEquity #Founders #SeriesA #SeriesB #Founders #SellSide #TechAssets #Fundraising #BuildBuyPartner #GoToMarket #PharmaTech #BioTech #Genomics #MedTech