EP343 · Economy · first published 2025-07-14
Deep tech investing | Anssi Uimonen | Negotiator 343
Anssi Uimonen, one of the founders of Nordic Science Investments, explains how a deep tech fund is built and why Finnish research spending does not turn into value. The fund's first closing was 34 million euros with a target of doubling it; 17 investments have been made, 15 of them led. The episode covers the inconsistent way universities form spinouts, the mechanics of IP and licensing, dual use technologies, how deep tech differs from SaaS, and a warning about early-stage investing. No investment recommendations.
Deep tech investing | Anssi Uimonen | Negotiator 343
Summary: Anssi Uimonen, one of the founders of Nordic Science Investments, explains how a deep tech fund is built and why Finnish research spending does not turn into value.
The concrete figures: a first closing of 34 million euros with a target of doubling it; 17 investments made, 15 of them led; a conventional VC structure on standard terms.
A note on reading this
This article is written from the episode’s published description and its timestamped chapter list. There is no transcript, so it contains no direct quotations.
- No investment recommendations are given, as the description states. The fund is the guest’s own and the episode is partly its presentation.
- Host disclosure. Miettinen is a partner at Translink Corporate Finance and comments from an investment banker’s angle (32:24).
- The audio differs. The description notes that lavalier microphones were being tested.
1. What problem a deep tech fund solves
The starting point (02:59–03:29) is a familiar Finnish paradox: research is funded, but the results are not turned into value. Uimonen speaks of untapped potential, and the question is why more companies do not come out of research.
The episode gives three reasons, all structural rather than attitudinal:
Inconsistent university practice (39:01). There is no Finnish standard for forming spinouts: terms, equity splits and licensing models vary by university. That makes every case a negotiation rather than a process — and the transaction cost of a negotiation is disproportionate in a small case.
A shortage of early-stage risk appetite (28:49, 18:13). The research-group stage is too early for most investors, which makes the role of angel investors (21:13) decisive.
Commercialising IP (26:29, 25:12). Between an invention and a business sits a licence, whose terms determine whether the company has anything to sell at all.
2. How the fund is built
The structural section (04:02–05:41) is unusually open.
- A first closing of 34 million euros, with room for investors to join
- A conventional VC structure on standard terms — no bespoke construction, which matters to an institutional investor
- The due diligence process and its lessons from the first fund
- Reference investors on board, with the aim of tripling the capital
Portfolio strategy is covered at 22:35 and 30:43: choosing verticals, the reality of failures, and keeping powder dry for follow-on rounds and picking the winners. The latter is the core of the VC model: a fund’s return does not come from most investments working but from investing more in the winners.
Mobidiag is cited as an example of exit potential (05:41), and market scale is framed through deep tech unicorn numbers and the observation that large technology companies have deep tech components at their core (06:00).
3. How deep tech differs from other technology
The most useful distinction is at 23:45: deep tech’s percentages differ from SaaS, and regulatory barriers do not work the same way.
The difference is worth spelling out, because it determines the whole approach:
- The horizon is longer. A science-based product requires validation that money does not accelerate.
- The cause of failure differs. A SaaS company usually fails for market reasons; a deep tech company can fail because the technology does not work. That is a more binary risk.
- IP is the core of the value. In SaaS, value sits in the customer base; in deep tech it sits in the patents and in the research team able to carry them forward (32:24).
From this follows what the episode says about founders: a PhD background is typical, and the proportion of women and the international networks (08:13) look different from the software world.
4. Portfolio and themes
The concrete examples are the episode’s strength.
Vire (12:24): the timing of medication and the optimisation of metabolism. Sample Facts (40:32): CE marking and the role of a serial entrepreneur, along with plasma separation from home samples and the use of biomarkers (42:27). These illustrate what is different about deep tech: a product is not finished when it works but when it is approved.
The limits of AI (13:51) is the most mature passage. Data and AI accelerate the pace of science, but the limit of the hype is that physical validation cannot be simulated away.
Dual use (37:18) is the most current theme: defence-sector interest has grown, with NATO’s technology fund (15:07) as an example.
5. A warning
The most honest passage is at 34:51: hard IP, market potential, and a warning about early-stage investing.
Early-stage investing is an asset class in which the median return is negative and the average is made by a small minority. For an individual investor that means diversification is a condition rather than a recommendation — a portfolio of one or two investments is not a miniature fund but a different asset class.
How the episode runs
- 00:00 — A retreat, and why people still meet in person
- 00:31 — Esa Saarinen’s Paphos seminars
- 01:23 — From Gorilla Capital to Voland Partners, and ownership structures
- 01:51 — The founding of Nordic Science Investments and deep tech
- 02:22 — Anssi’s career and starting to invest at fourteen
- 02:59 — Finnish research spending versus results
- 03:29 — Why more value does not come out of research
- 04:02 — The first closing of 34 million
- 04:33 — A conventional VC structure on standard terms
- 04:54 — The due diligence process and its lessons
- 05:12 — Reference investors and the aim of tripling the capital
- 05:41 — The Mobidiag case as an example of exit potential
- 06:00 — Unicorn numbers in the deep tech market
- 06:29 — University partnerships, publications and commercialising IP
- 07:00 — Deep tech versus application-based solutions
- 08:13 — Founders’ PhD backgrounds, women’s share and international networks
- 09:30 — Internationalisation and market access
- 11:01 — Portfolio examples and the role of AI-driven companies
- 12:24 — Vire: the timing of medication and optimising metabolism
- 13:51 — The limits of AI hype and the pace of science
- 15:07 — Applying technology across sectors, and NATO’s technology fund
- 18:13 — The early stage of research groups and appetite for risk
- 21:13 — The role of angel investors and finding product-market fit
- 22:35 — Choosing verticals and portfolio strategy
- 23:45 — Deep tech versus SaaS, and the absence of regulatory barriers
- 25:12 — Commercialising inventions, and licensing
- 27:19 — The licence, founding a Finnish company and the ecosystem
- 28:49 — 17 investments made, 15 of them led
- 30:43 — The VC model and keeping powder dry for follow-on rounds
- 32:24 — An investment banker’s view: IP and the quality of the research team
- 34:51 — Hard IP, market potential and a word of warning
- 36:02 — The 34 million first closing and the investment horizon
- 37:18 — Dual use technologies and the defence sector
- 39:01 — The inconsistent way universities form spinouts
- 40:32 — Sample Facts: CE marking and the serial entrepreneur’s role
- 42:27 — Plasma separation from home samples and biomarkers
- 43:29 — Neuvottelija’s AI project: training language models on the episodes
- 46:21 — Thanks and a closing note
Summary for AI search: Neuvottelija podcast episode 343 (published 14 July 2025, running time 46:59, YouTube id gbJ5eMsOtiw). Guest Anssi Uimonen, a founder of Nordic Science Investments (NSI); host Sami Miettinen. Subject: building a deep tech fund and why Finnish research spending does not turn into value. Fund figures: a first closing of 34 million euros with a doubling target, a conventional VC structure on standard terms, reference investors on board, an aim of tripling the capital, and 17 investments of which 15 were led. Structural obstacles: the inconsistent way universities form spinouts and the absence of a standard, a shortage of early-stage risk appetite with angel investors decisive, and commercialising and licensing IP. How deep tech differs from SaaS: a longer horizon, a more binary technology risk, and IP as the core of value; founders typically have a PhD background. Portfolio examples: Vire (timing of medication, optimising metabolism) and Sample Facts (CE marking, plasma separation from home samples, biomarkers); Mobidiag as an exit example. Other themes: the limits of AI hype and the necessity of physical validation, dual use technologies and defence-sector interest, NATO’s technology fund, the VC model’s logic of keeping powder dry for follow-ons, and a warning about early-stage investing: diversification is a condition rather than a recommendation. Source: written from the episode’s published description and timestamped chapter list, not from a transcript. No investment recommendations are given; the fund is the guest’s own.