Ask a scientist founder what the hard part of a first round was, and you rarely hear about the pitch. You hear about the 9 months before it: building a list of investors who could evaluate the technology, finding people who could introduce them, working out how much money the next technical milestone needs, and getting someone outside the lab to confirm the result holds.
Founders who raise well start early and go wide, then narrow. Tobias Grab, co-founder of Kipu Quantum, started from the top of that funnel and worked down:
"We started with all the VCs in an Excel sheet, 860 of them, and then played an erasing game until we were down to around 200. We began with the ones we could practice on and moved up the ladder to the ones we really wanted."
Two things make that approach work. The pitch improves measurably between conversation 5 and conversation 50, so the sequence matters. And a long list built months early gives you time to convert names into warm introductions, which is how most first rounds get done.
Hardware and lab-stage technology narrow the field of investors who can evaluate you. Felix Poernbacher, co-founder and co-CEO of DeepDrive, which builds dual-rotor electric motors and works with most of the world's largest carmakers:
"Hardware, automotive, that is not your average startup, and it is not what investors are normally used to. We knew we had to run a pretty broad process. Look for investors who did something in the space, who understand the business model."
DeepDrive's €4.3 million seed round in 2022 had 2 institutional investors with enough capital to follow on, plus 2 angels picked for specific gaps. Peter Mertens, former CTO at Audi, covered industry credibility. Jonas Rieke, co-Founder & COO of Personio, covered scaling. The company has since raised a €15 million Series A and a €30 million Series B.
Kipu Quantum solved the same problem differently, anchoring its €3 million first round on a lead investor who specializes in quantum, then adding a local investor who knew the team and a US investor positioned for the follow-on rounds the company expected to raise on the West Coast. Fifteen months later, it raised €10.5 million.
Both rounds follow the same logic. Pick a lead who can assess the physics, then fill the remaining slots with people who close potential gaps. A lead who has to be taught the field will slow the round down and add little afterward.
Martin Weber and Claire Hae-Min Gusko started one.five in 2020 after working at vertical-farming unicorn Infarm and built their first round mostly with angels. By their own account, they spoke with roughly 150 angels and brought more than 65 into a pool, with individual tickets starting at €10,000.
"A €10,000 ticket, which many startups would turn down, unlocked a much larger fundraising opportunity, simply because those angels then had skin in the game."
Martin describes making direct cap table entry dependent on a ≥€250,000 investment and routing everyone below that into a separate investment vehicle where the angels became limited partners. The cap table shows one entry. The company gets 60-plus people incentivized to make introductions. one.five raised over €10 million across those early closings and added a €14 million Series A in January 2026, including investors like the foundation behind Haribo.
Martin’s own summary of when to use angels: start with them to learn the process. Angels usually join only at the beginning, while VCs will be your funding source for years. Small single-GP micro funds behave much like angels while technically being institutional, and they are worth considering alongside traditional pre-seed funds.
Investors ask how far the money takes you, and for a lab-stage company the honest answer is rarely a revenue number.
When Jan Goetz first pitched IQM, his business plan was a line-by-line list of lab equipment prices, until a VC looked at it and asked, "That's nice, but what happens if I give you €12 million?" Every first round eventually comes down to having an answer to that question.
Tobias’ framing:
"The question I ask is what would de-risk the venture most. Does the technology work? Does it work in different applications? Does it work when the technical co-founder is not in the room? What kind of talents do we need to hire?"
For Kipu Quantum, product-market fit wasn't a plausible goal for a first round because the available quantum hardware wasn't good enough to deliver value yet. So the round was sized around technical de-risking instead, which produced a €3 million ask.
That approach matches what pre-seed deep tech looks like now. Our Deep Tech Napkin, drawn from more than 100 rounds across 20-plus European deep tech VCs, found 80% of pre-seed companies sitting at concept or lab-demonstration stage, up from 60% the year before. Exactly one company in the dataset had revenue at pre-seed. At Series B, 30% still had none.
On valuation, a rule of thumb Tobias used, is that early-stage valuation lands at roughly 3 to 6 times the capital you raise, with a priced round costing existing shareholders around 20% of the company. He is clear about the limits. Nobody can arrive at a defensible valuation this early, so benchmarks, comparable companies, and a credible market size do the work, and having more than one interested party does more than any spreadsheet. Terms other than price can matter more, especially liquidation preferences.
Market data lands in the same range. Carta's 2025 pre-seed data shows median post-money valuation caps around $10 million for rounds between $250,000 and $1 million, and around $15 million for rounds between $1 million and $2.5 million, with most early-stage rounds under $4 million still done on convertible instruments rather than priced equity. In Germany, that instrument is usually a convertible loan, which both Martin and Tobias used to bridge delays and move faster than a priced round allows.
Revenue is one of the strongest proof points, and occasionally a deep tech company has some before the first round. Hyperdrives, which we backed in 2025, went at the same market as DeepDrive from a different angle. Instead of a new motor layout, its patented hollow-conductor cooling raises continuous current density by up to 3 times and can be industrialized on existing hairpin production lines. The founders came out of competitive motorsport and volume automotive engineering, and passed €1 million in revenue in 2024 with no external funding.
Paying customers, plus a board seat filled by a former McKinsey senior partner in automotive, answered most of what a first-round investor would otherwise have to take on trust. The €3 million pre-seed followed.
Without revenue, you need something else to point at. Martin calls it corporate inventory, an explicit list of what the company has delivered, down to items that sound trivial in isolation: signed NDAs, letters of intent, or completed test campaigns. Read together, they show movement toward a market.
Felix’ version is external validation:
"It is very hard for investors to judge whether you have an edge over other technologies in the market. So we spent a couple of months validating on the test bench before we raised, and then went to an independent third party to confirm that what we claim is achievable."
DeepDrive built prototypes, measured efficiency and cost against incumbent motors, and had a Fraunhofer institute confirm the results. The claim then travels without the founder having to defend it in every meeting.
External validation not only makes fundraising easier, but also hiring and market entry later on. A test report is one route to that, early commercial signals including NDAs, LOIs, design partnerships are another, and most first rounds lean on several at once.
Non-dilutive funding is genuinely attractive for research spinouts with several in instruments, including exist, research allowance, or the prestigious EIC Accelerator.
Tobias’ warning is worth taking seriously anyway. He has watched companies become dependent on public funding to the point where they could never break even without it, and he suggests talking to entrepreneurs and investors about the business you are exploring before you apply, rather than after. The sequence usually runs the other way, which leaves the founder close to their doctoral supervisor and far from customers. Heavy grant funding also comes with fixed milestones, which is a poor fit for a company whose plan should change when the evidence does.
Incubators need the same scrutiny. Martin’s line on dilution is blunt: incubators that take 30% to 40% make you difficult to fund later, and in the earliest stages founders should still hold 90% to 95% of the business. Programs offering a desk, some introductions and no real capital in exchange for equity are a different proposition entirely. Structured support, access to facilities and genuine networks are worth paying for. Shares for free are not.
Incubation can also run the other way. First Momentum assembled Telura's founding team and built the company around electric impulse drilling, and it has since raised a $5 million pre-seed from outside investors. That last part tests any hands-on model. Done properly, it avoids the usual venture studio outcome, where the builder holds so much of the cap table that the next investor has no room to invest at reasonable terms anymore.
Fundraising timelines drift, and they mostly drift in the investor's favor. Daniel Reese, a former VC investor at UVC Partners:
"Don't give away the driver’s seat of your fundraising process. They will always try to take more time and see if somebody else jumps first. Two weeks for first calls. Weeks three to four, everyone who wants to proceed gets the data room. No exceptions."
Deep tech complicates this, because technical diligence takes time. Investors need to understand the physics, call references, and talk to experts. So start the relationships earlier, and still hold the process itself to a schedule. Even a tightly run process stays a numbers game, and the last stretch is unpredictable. Felix’ read on the odds:
"You talk to 80 investors and you get 75 or 78 no's. That is just how fundraising works, and it is never going to go exactly as you planned it."
The most transferable practice from any of these companies is Martin’s reporting habit. one.five sends its investor updates to prospective investors as well as current ones, quarter after quarter, before anyone has committed. When the next round opens, those conversations resume from last quarter rather than from two years ago. It also changes what goes into the report: updates written for future investors track where the market and the pipeline are heading, alongside what already happened.
That includes what went wrong, which is where the incentives are less obvious than they look. Daniel's point is that investors are comfortable taking a risk they know about. The failure case is an LP asking two years later why they invested, and having to admit they didn't see it. Naming your risks and showing how you plan to retire them is a stronger position than presenting a plan with no visible holes. Investors who do this for a living will find the holes anyway, and finding them late costs you the round.
If you are still deciding whether your research should become a company, the sequencing looks like this. Build the list months before you need it. Find the 20 investors who could actually assess your technology, and route to them through people who know them. Work out which technical risk the money retires, and size the round to match it. Line up at least one outside verdict on your central claim, whether that is a test report, a published result, a fabrication partner, or a customer paying for a prototype. Then run a tight process and keep reporting to the people who said no, because a first no at pre-seed is often a yes 12 months later.
There's still structural work to do around all of this. IP negotiations between founders and their universities average 1.5 years before a proper round can close. Worth kicking this off early.
For the parts you can control, the European Spinoff Summit exists precisely for this stage. Co-hosted by First Momentum and SPRIND, it includes workshops on fundraising and IP transfer, founder talks from people who have raised these rounds, and a help desk where you can bring your own challenge. The waitlist for 2027 is open.
