Growth-Stage Startups

Evaluating Swiss Deep Tech Startups: A Framework for Private Investors

September 4, 2025

Switzerland leads Europe in deep tech innovation. In the country, 60% of all venture capital funding goes toward deep tech companies, the highest share globally.[1] For investors in Zurich, Geneva, and Lausanne, this concentration creates opportunity. But deep tech startups differ fundamentally from those in other sectors like software. They require longer development cycles, heavier capital investment, and specialized evaluation frameworks.

Most non-professional investors broadly understand how to assess a software startup. You can test the product, see the user interface, and understand the business model in an afternoon. Deep tech is different. A quantum computing company, a biotech startup developing novel therapeutics, or a cleantech firm building industrial-scale carbon capture technology requires years of research before generating revenue. Traditional due diligence methods are not good enough.

This article provides a practical framework for evaluating Swiss deep tech opportunities. Whether you are building a startup investment portfolio or representing a family office exploring alternative capital, understanding these evaluation principles helps you make informed decisions.

Why Swiss Deep Tech Demands Specialized Evaluation

Deep tech represents technology based on scientific breakthroughs or engineering innovations applied for the first time as commercial products. The category includes artificial intelligence, robotics, advanced materials, biotechnology, quantum computing, and clean energy solutions. Swiss startups in these sectors raised $1.9 billion in 2024, with projections reaching $2.3 billion in 2025.[2]

Three characteristics separate deep tech from regular tech startups:

Extended development timelines: A software company can launch a minimum viable product in six months. A medtech device seeking Swissmedic approval requires 24 to 48 months from submission to market clearance for high-risk classifications. Cleantech hardware companies need years to move from laboratory prototypes to pilot-scale demonstrations. These extended timelines mean investors must evaluate scientific milestones, not just customer traction.

Capital intensity ahead of revenue: Software startups can achieve product-market fit with modest funding. Deep tech companies require substantial capital before seeing their first customer. A biotech company developing therapeutics might spend CHF 5 million on pre-clinical studies before human trials begin. This capital intensity changes the investment risk profile.

Technical validation complexity: Investors can evaluate a mobile app by downloading it. Evaluating whether a novel battery chemistry will scale to commercial production requires domain expertise most investors lack. This complexity requires different validation signals.

The Academic Advantage: ETH and EPFL Spin-Offs

Switzerland’s deep tech ecosystem centers on two academic pillars: ETH Zurich and EPFL (École polytechnique fédérale de Lausanne). These universities rank among the top four in Europe for deep tech spin-out value creation, alongside Oxford and Cambridge.[3]

The academic origin creates a measurable advantage. ETH and EPFL spin-offs demonstrate a 90% five-year survival rate, compared to the 50% national average for all new Swiss companies.[4] This survival advantage stems from systematic support infrastructure, advanced selection processes, and access to world-class research resources.

For investors: the ETH or EPFL affiliation serves as a first-level quality filter. The universities employ formal technology transfer processes that verify intellectual property ownership, assess market potential, and connect startups with experienced mentors. ETH Zurich created 37 spin-offs in 2024 and 43 in 2023.[5] EPFL has similar output, with its Innovation Park hosting over 200 tenant companies.[6]

For founders: these institutions provide access to specialized equipment, technical expertise, and a network of potential corporate partners. The survival rate advantage reflects this structural support.

Geographic concentration follows university strengths. Zurich excels in artificial intelligence, robotics, and fintech, reflecting ETH’s computer science and engineering departments. Lausanne dominates in life sciences, cleantech, and advanced materials, following EPFL’s research focus. Basel leverages its pharmaceutical industry concentration for biotech startups.

Validation Programs: Decoding Innosuisse and Venture Kick

Swiss deep tech startups have access to several validation programs that signal technical merit and execution capability. Understanding these programs helps investors interpret a startup’s development stage.

Innosuisse Support Programs

Innosuisse, Switzerland’s innovation agency, operates multiple startup support mechanisms. Each program provides different validation signals:

Core Coaching Program: This program received 552 applications in 2024, with a 76% approval rate for startups meeting basic criteria.[7] Over 640 startups were receiving active support by the end of 2024. While this program has relatively open access, participation signals that founders have engaged with structured business development processes. Impact monitoring shows that coached startups achieve more successful financing rounds, with 19% raising over CHF 5 million.[8]

Start-up Innovation Projects: Introduced in 2023, this program provides direct funding to science-based startups before market entry. It received 225 applications in 2024, demonstrating high demand.[9] Unlike coaching, the mechanism involves competitive grant selection. Receiving this funding indicates that Innosuisse evaluators assessed the startup’s technical innovation as both scientifically sound and commercially promising.

Innogrant and BRIDGE Programs: These highly competitive programs provide substantial funding. While precise acceptance rates are not publicly disclosed, program descriptions point to detailed peer review by academic experts. The BRIDGE program offers awards up to CHF 930’000 for projects moving from fundamental research toward commercialization.[10] Any startup successfully navigating these application processes has survived intense technical scrutiny.

For investors: Ask whether the startup has received Innosuisse funding beyond basic coaching. Direct funding through Start-up Innovation Projects, Innogrant, or BRIDGE show meaningful third-party validation of technical merit.

ETH and EPFL Fellowship Programs

ETH Pioneer Fellowship provides CHF 150’000 to support early-stage ventures. The program receives over 100 applications every year for 10 to 15 fellowship slots, creating approximately 10% to 12% acceptance.[11] This competitive selection filters for both technical excellence and entrepreneurial potential. ETH reports that enterprises founded by Pioneer Fellows have significantly higher chances of successful exits compared to average startups.[12]

EPFL Innogrant awards CHF 100’000 to support spin-off development. Since launching in 2005, the program has supported 145 spin-offs.[13] This program pioneered the institutional grant model in Switzerland, establishing standards for academic entrepreneurship support.

Venture Kick operates a three-stage funding program providing up to CHF 150’000 total (CHF 10’000 in Stage 1, CHF 40’000 in Stage 2, CHF 100’000 in Stage 3). Approximately 40% of startups progress from Stage 1 to Stage 2, and about 50% advance from Stage 2 to Stage 3.[14] Only around 20% of initial entrants reach the final stage. This staged progression creates a clear validation timeline. A startup that secured all three Venture Kick stages has repeatedly demonstrated milestone achievement over an extended period.

Sector-Specific Evaluation Frameworks

Deep tech evaluation varies substantially by sector. A biotech company faces different regulatory pathways than a cleantech hardware startup. Understanding sector-specific validation methods helps investors know what to expect.

Biotechnology and Medical Technology

Swiss biotech and medtech startups operate within established regulatory frameworks. Swissmedic, the Swiss regulatory authority, classifies medical devices into risk categories that determine approval requirements:

Device classifications range from Class I (low risk, such as bandages) through Class III (high risk, such as implantable devices). Higher classifications require more extensive clinical evidence. A fast-track pilot program launching July 1, 2025, aims to reduce approval timelines for trials addressing high medical need.[15] This regulatory development signals Switzerland’s commitment to innovation while maintaining safety standards.

Clinical validation milestones provide investor checkpoints. For seed-stage investments, successful pre-clinical animal studies and a clear human trial protocol indicate progress. For Series A investments, completed Phase I human trial data demonstrating safety represents a major de-risking event. Investors should verify that startups have formal partnerships with Swiss university hospitals such as Universitätsspital Zürich, Centre hospitalier universitaire vaudois (CHUV), or Inselspital Bern. These partnerships provide access to clinical trial infrastructure and medical expertise.

For medtech founders, regulatory approval represents the dominant risk. Investors should verify that the team includes regulatory affairs expertise or has engaged specialized consultants. Ask to see the regulatory strategy document and timeline assumptions.

Clean Technology and Climate Innovation

EPFL reports that 17% of startups founded since 2021 focus on clean technology, reflecting growing investment in climate solutions.[16] Evaluating cleantech requires understanding technology readiness levels (TRLs) and certification pathways.

Pilot scale validation serves as a critical checkpoint. Laboratory demonstrations (TRL 4 to 5) prove basic scientific principles but do not validate commercial viability. Investors should look for pilot projects at meaningful scale, typically defined as one-tenth of a planned commercial unit. For carbon capture technology, this might mean demonstrating capture of specific tonnage at industrial facilities. For plastic recycling processes, pilot scale should process measurable daily volumes with verified chemical purity outcomes.

Third-party technical validation from respected Swiss institutions adds credibility. Empa (Swiss Federal Laboratories for Materials Science and Technology) evaluates projects for innovation potential, market relevance, and commercial feasibility.[17] A validation report from Empa or similar institutions (Paul Scherrer Institute for energy systems, CSEM for microtechnology) provides independent technical verification.

Carbon credit certification through recognized bodies like Gold Standard or Verra demonstrates that climate impact claims can withstand independent audit.[18] These certifications require verification of carbon reduction methods, additionality proofs, and monitoring protocols. For cleantech startups claiming environmental benefits, ask whether they have begun or completed this certification process.

Artificial Intelligence and Machine Learning

AI and machine learning startups represented 23% of new Swiss deep tech companies founded since 2021.[19] This growth reflects Switzerland’s research strengths but also creates evaluation challenges in a crowded sector.

Swiss National AI Institute (SNAI) launched in 2024 as a joint initiative of ETH Zurich and EPFL. The institute provides a new route to validation for AI startups. Affiliation with SNAI means access to over 70 AI-focused professors and collaboration with the Swiss National Supercomputing Centre and its Alps supercomputer.[20] This institutional connection indicates that the startup’s technical approach has academic credibility and can access computational resources often required for advanced AI development.

Data moats determine competitive sustainability in AI. Investors should ask: Where does your training data originate? What makes your data access exclusive? Generic answers referencing publicly available datasets like ImageNet suggest a weak competitive advantage. Strong answers will cover proprietary data generated through exclusive partnerships with hospitals, industrial partners, or unique sensor deployments. For Swiss AI startups, partnerships with CHUV for medical imaging data or collaborations with Swiss industrial companies for manufacturing data represent meaningful competitive advantages.

For AI founders, the question investors should ask is whether the product improves with each new customer. This property indicates a data network effect, where the product naturally becomes more valuable over time. Generic AI applications rarely possess this characteristic.

Due Diligence for Non-Technical Investors

Most private investors lack deep scientific expertise. You do not need to understand quantum mechanics to invest in quantum computing. You need frameworks for verifying that credible experts have evaluated the technology and endorse its commercial potential.

Intellectual Property Verification

Intellectual property clarity represents the most critical legal checkpoint. Swiss universities maintain formal technology transfer processes. ETH Transfer manages ETH Zurich’s spin-off support, while EPFL operates similar technology transfer functions. These offices verify IP ownership before approving spin-off creation.

Red flag: Ambiguous IP ownership. Ask to see the signed technology transfer agreement from the university to the startup. If core intellectual property remains university-owned with only a license to the startup, this arrangement creates complications for future funding rounds. Exclusive, worldwide licenses are acceptable. Direct IP assignment to the company is strongest.

Red flag: The Professor Problem. Occasionally, a founding professor retains personal IP rights separate from the university assignment. This situation creates significant obstacles for venture financing. Any investor should verify that all relevant background IP from the founding laboratory has been exclusively assigned or licensed to the startup.

Investors should ask: “Can you show me your IP docket?” This document lists all patents and patent applications. Look for granted patents in major markets (Switzerland, European Union, United States, China, Japan), not just provisional applications. Provisional patents expire after 12 months if not converted to utility patents. A portfolio consisting only of expired provisional patents provides no protection.

Team Composition and Red Flags

Deep tech teams require a specific mix of skills. The “Ivory Tower Syndrome” describes a common early-stage weakness: a founding team of three PhD graduates from the same laboratory with no dedicated business leader. This structure often succeeds through seed stage when technical development dominates. The problem emerges at Series A when commercial execution becomes essential.

For investors, a founding team entirely composed of technical experts with no clear plan to add commercial leadership represents a yellow flag at seed stage. This warning becomes red at Series A if unaddressed. Ask: Who owns the go-to-market strategy? Who has direct sales experience? If all founders answer “I’m focused on the technology,” the team needs strengthening.

Green flag example: A medtech startup with a founding scientist from EPFL, a co-founder with 10 years of medical device sales experience, and an advisory board including a cardiology professor from CHUV and a former executive from a medical device company demonstrates balanced capabilities.

Validation Through Advisory Boards

For non-technical investors, advisory board composition serves as a proxy for technical quality. A strong deep tech advisory board typically includes three distinct types of expertise:

Academic knowledge to validate the underlying science (for example, from an ETH professor in the relevant field)
Industry knowledge to validate the market need (for example, from a senior R&D leader from Roche for biotech)
Management knowledge to validate the execution plan (for example, from a founder who successfully exited a previous Swiss startup)

Generic business advisors without domain expertise signal weak technical validation. The advisory board should be able to show that respected experts in the specific field have reviewed the technology and endorse its commercial potential.

Questions Non-Technical Investors Should Ask

These questions reveal technical risk without requiring scientific expertise:

“Has your core technology been independently validated by a third party such as Empa, PSI, or CSEM? May I review the report?”
“What is the foundational peer-reviewed publication for your technology? Which journal published it?”
“Walk me through your IP docket. Which core patents are granted, not just filed?”
“Have you conducted a formal Freedom to Operate analysis? What were the key findings?”
“What is the single biggest technical hurdle you still need to overcome to reach commercial scale?”
“How many prototype iterations have you built? What was the most significant failure?”
“Where does your training data come from?” (For AI startups) “What makes your data access defensible?”
“Beyond your team, what is your unfair technological advantage that a competitor with CHF 10 million could not replicate in 12 months?”
“Who on your scientific advisory board has challenged your technical approach most intensely? What was the outcome?”
“What are your technology’s key performance metrics in real-world conditions versus controlled laboratory environments?”

Listen for specific, detailed answers. Unclear responses or defensiveness about these questions suggests founders do not truly know about their technology’s true capabilities or limitations.

Common Red Flags in Swiss Deep Tech Deals

Pattern recognition across failed deep tech investments reveals recurring warning signs:

Unrealistic Timeline Claims

Deep tech development requires patience. Biotech founders claiming human trials will begin in 12 months from incorporation ignore the 24-plus months typically required for pre-clinical work, regulatory submission preparation, and ethics committee approvals. Cleantech hardware founders promising commercial plant operations within two years from laboratory prototype overlook the multi-year process of pilot plant design, construction, commissioning, and iteration.

Compare founder timelines against sector norms. If a medtech founder’s schedule assumes 18 months from concept to Swissmedic approval for a Class III device, this unrealistic projection shows poor experience or wishful thinking.

Customer Over-Dependence

A single pilot project with one corporate partner should not represent more than 70% of a startup’s validation and near-term revenue projection. If that corporate partner exits the relationship, the startup faces potential failure. This concentration risk appears frequently in B2B deep tech companies that secure one marquee pilot but have not validated broader market demand.

Ask: “What happens if your primary pilot customer decides not to convert to a commercial relationship?” You do not want an answer like “We would need to restart customer development.”

Weak IP Protection

Founders claiming “strong IP protection” based entirely on five provisional patents filed 18 months ago without utility patent conversions show poor understanding. Provisional patents provide 12 months of priority protection, then expire. A portfolio of expired provisionals provides zero competitive protection.

Similarly, deep tech startups relying entirely on trade secrets rather than patents face risks. A competitor with sufficient resources can often reverse-engineer physical products or processes. Patents provide legal protection that trade secrets cannot match in most physical science applications.

Why Switzerland Leads European Deep Tech

Understanding Switzerland’s structural advantages helps explain the ecosystem’s strength and the opportunities it creates.

The country ranks first in Europe for per-capita deep tech venture capital funding.[21] This concentration creates a strong cycle. Top researchers come to ETH and EPFL because of research funding. These researchers create spin-offs that attract venture capital. Successful exits produce angel investors and serial entrepreneurs who mentor the next generation.

The 60% allocation of Swiss venture capital to deep tech companies far exceeds other nations’ commitments.[22] For comparison, most European countries direct 25% to 35% of venture funding toward deep tech. This concentration means Swiss investors have developed specialized expertise in evaluating complex technologies.

International capital plays a crucial role. Nearly 96% of late-stage deep tech funding rounds in Switzerland involve global investors.[23] This international participation reflects the quality of Swiss startups and their global market potential. Startups successfully raising seed and Series A rounds in Switzerland typically turn to international venture capitalists for Series B and later rounds.

For founders, it is important to build a company capable of attracting global investors from the start. Swiss-only customer bases or Switzerland-specific solutions will not be as scalable.

For investors, the presence of international venture capitalists in later rounds provides potential co-investment opportunities and exit liquidity. If Sequoia or Lakestar leads a Swiss startup’s Series B, this outside validation confirms the opportunity’s global potential.

Building a Deep Tech Investment Portfolio

Deep tech investing follows the power law distribution observed across venture capital. One or two exceptional outcomes generate the majority of portfolio returns. Most investments return zero or modest multiples. This distribution makes portfolio construction essential.

The SICTIC Angel Investor Handbook, the authoritative Swiss angel investing guide, recommends diversification across 10 to 20 startups.[24] This recommendation applies with particular force to deep tech, where technical and regulatory risks create higher failure rates than software investing.

Practical portfolio construction for non-professional investors might involve:

Stage diversification: Mixing seed-stage deep tech investments (higher risk, higher potential returns) with growth-stage opportunities (more validation, lower risk) balances portfolio exposure.

Sector diversification: Avoid concentrating entirely in biotech or entirely in AI. Climate tech, advanced materials, robotics, and quantum computing each face different risk profiles and timeframes.

Geographic diversification within Switzerland: Zurich-focused AI companies face different challenges than Lausanne-based life sciences ventures. This geographic spread provides exposure to different research strengths and corporate partnership networks.

Validation stage diversification: Some portfolio companies should have completed pilot projects or clinical trials. Others might be pre-pilot but have secured competitive Innosuisse grants. This mix reduces correlation between failures.

For family offices, deep tech investing often fits within a broader private capital strategy that includes real estate and lending. These asset classes involve different cashflows, risks, and rewards. Deep tech’s high-risk, high-return profile complements more stable alternative investments.

Conclusion: Precision Over Intuition

Swiss deep tech startups solve hard problems. Carbon capture, quantum computing, advanced therapeutics, autonomous robotics. These challenges require years of research, substantial capital, and specialized expertise. Evaluating these companies demands frameworks suited to their complexity.

You cannot test deep tech products in an afternoon. You must verify that credible experts have validated the technology. You must understand regulatory pathways and capital requirements specific to each sector. You must recognize that extended development timelines and capital intensity change the investment mathematics.

Switzerland’s ecosystem provides structural advantages. The 90% five-year survival rate of ETH and EPFL spin-offs reflects systematic support, rigorous selection, and access to world-class research. Validation programs like Innosuisse grants and Venture Kick stages create observable checkpoints. Geographic concentrations in Zurich, Lausanne, and Basel provide access to specialized expertise and corporate partnerships.

For investors building alternative capital portfolios, Swiss deep tech offers exposure to breakthrough innovations with global market potential. The combination of academic excellence, capital availability, and a proven track record creates an ecosystem where patient capital can generate exceptional returns.

Whether evaluating your first deep tech investment or building a mature portfolio, these frameworks help translate scientific innovation into informed investment decisions. The opportunity exists. The challenge is evaluation.

CapiWell serves investors seeking structured access to vetted growth-stage Swiss deep tech opportunities. By combining comprehensive deal evaluation with regulatory compliance and transparent reporting, it aims to make sophisticated alternative capital investments accessible to qualified investors who understand the unique characteristics of deep tech as an asset class.