What "capex-light"means in practice for industrial startups - the Edonia case
In deep tech, the orthodoxy says capital buys speed. One portfolio company spent two years proving that speed can be engineered, not bought.
Vous souhaitez lire en français ? → Cliquez ici
Deep food tech runs on one assumption: that only a large early raise can buy the speed to industrialise before the money runs out. In twenty-four months, on €1.1M of equity, Edonia reached a risk profile its well-funded peers take four to eight years and tens of millions to approach. What that says about capital efficiency, risk sequencing, and the work of an early-stage fund.
Every deep food tech founder is handed the same map on day one, and it is a seductive one. Fermentation is slow to master, industrialisation punishes every shortcut, food safety leaves no room for improvisation, and the infrastructure that carries a process from bench to commercial volume rarely costs less than tens of millions of euros. From those four hard truths the map draws a single road: raise early, raise large, and build fast enough that the market is still waiting when the production line finally turns over.
The logic is airtight on a slide. The trouble is what it does to a company - it converts the entire enterprise into a single bet that the first factory works on the first attempt, and quietly spends, on steel, the very capital that a second attempt would require.
This continues an argument from an earlier issue: that tooling and scale should follow market validation, not run ahead of it, and that building lean reads better as a competitive edge than as a constraint imposed by scarcity. Edonia carries that argument onto the hardest ground available - physical industry, food production, where prevailing wisdom holds that only heavy capital buys speed.
Edonia is a refutation of that logic. By April 2026 it had moved from a process that worked only on a laboratory bench to a 50-tonne-per-year pilot running in the exact configuration of its future factory. A patent filed, a second drafting. The factory designed and budgeted, buildings secured, equipment tested and ready to order. Fifteen deals signed with demanding industrial buyers, Newrest among them. Recurring revenue beginning.
It did this on €1.1M of equity.
To see why that number matters, hold it against the benchmarks the industry trusts. The Climate Brick framework - built by EQT Ventures and Contrarian Ventures with McKinsey research support, on data from more than three thousand companies - places top-quartile Series A businesses in Edonia’s category at TRL 4 to 6, pre-revenue, having raised €35M to €60M. The HTGF Deep-Tech Matrix sets Series A entry at TRL 5 to 6, first proof-of-concept revenue, and a pipeline above €10M; one late-stage investor quoted in that study judged that a deeptech company at TRL 5 with no revenue still belonged in seed. Edonia arrives at the Series A conversation at TRL 7, with recurring revenue, on €1.1M of equity. The orthodox cohort is two TRL levels behind, pre-revenue, and roughly forty times more capitalised. What follows is an attempt to explain how, and what it says about the work of an early-stage fund.
Why the orthodox map kills companies
The GFI’s industry data shows demonstration or commercial facilities running €15M to €250M, with meaningful first revenue arriving five to eight years after founding. More than one well-capitalised European player has since discovered that the size of its raise offered no protection against the sequence of its risk. The most visible French casualty was forced to hand control to new owners months after inaugurating its first factory in early 2024 - not for want of capital or engineering talent, but because it committed to heavy infrastructure while the core technical and commercial risk was still unresolved. The plant had to work on the first attempt. When it didn’t, the capital that might have paid for a second attempt was already gone.
None of these were careless bets. They were the disciplined application of a playbook that treats capital as the only lever capable of compressing time.
Where the company started
Edonia turns microalgae into a clean, high-protein textured ingredient for food manufacturers - a B2B input that goes into plant-based and hybrid products without the long additive lists those products usually carry. When Asterion led its seed round in March 2024, the company stood at the very beginning of its execution curve: a technology that worked at lab scale and nowhere else, a product not yet stable or reproducible, no patent, no contracted supply, not a single binding letter of intent on price or volume, no sales. Three co-founders, the third barely onboarded, no employees.
What it had was a sharply differentiated prototype, an enormous addressable market in protein ingredients, a founding team surrounded by a serious bench of industry operators, and a research collaboration with AgroParisTech. The signal was real. The entire execution risk lay ahead. That is the honest baseline for the two years that followed.
By April 2026: TRL 4 to TRL 7, the pilot testing critical unit operations on real production equipment rather than a lab approximation. The product finalised, clean-label regulatory cleared, client quality audits passed, on the market including an organic version. Fifteen signed deals representing €30M - and these are pre-contractual commitments with real detail and real buyer engagement, not light signals - covering 87% of projected 2029 volume. First recurring revenue in.
Three heterodox moves
None of this was luck. It rests on three decisions, each cutting against accepted practice.
Industrial philosophy. The conventional route engineers custom equipment around the biological process. Edonia did the reverse, forcing its process to adapt to standard machinery already on the market, some of it built for other industries. The consequences compound: a capex-to-capacity ratio among the lowest documented in food, equipment that qualified for leasing rather than purchase, a pilot already running at 90% conformity with overall equipment effectiveness at 60% and climbing toward an 80% target - all without Industry 4.0 spending. Scale-up reduces to replicating identical lines rather than commissioning bespoke infrastructure that has never run at volume. Structurally, the opposite of what destabilised the orthodox cohort.
Commercial intensity against headcount. That entire pipeline - fifteen deals, €30M, buyers of that calibre - was built by the equivalent of one and a half full-time commercial staff. The 50,000 meals served through the Newrest proof-of-concept were not a marketing flourish; they were an industrialisation test run in school catering, among the least forgiving segments on budget, logistics and the expectations of the people actually eating the food.
Non-dilutive capital. Edonia has won every public funding application it has filed in France, including the i-Lab Grand Prize - roughly twenty winners a year out of several hundred - and mobilised more than €3M non-dilutive on top of the €1.1M equity. That money financed execution directly, without dilution. It is the lever most teams leave untouched.
What a fund actually does at this stage
Capital alone does not move a company from TRL 4 to TRL 7 in twenty-four months. The conditions around it do, and building those conditions is the job — which is why it tends to be underestimated, because none of it appears on a term sheet.
In Edonia’s case that meant an advisor bench of operators, not fellow investors: people who have commissioned plants and survived industrial scale-ups, among them Laurent Cardinali, former VP Global Engineering at Danone and Mondelez, and Alexis Angot, former CFO of Ÿnsect - profiles a team of three at TRL 4 cannot recruit and could not otherwise reach. It meant support on the public funding applications that turned non-dilutive capital into a real execution lever. And it meant introductions that compressed a path to industrial partners which usually takes years of cold prospection.
The posture is simple: Asterion does not look for companies that have already removed their risk, because by then the return is largely priced. It looks for teams with the capacity to remove it, and puts its operational weight behind making that happen.
The objection worth naming
There is an objection here, and it deserves naming rather than skirting. The companies that failed did so after building their factories. Edonia has not built one yet. Comparing the two means comparing the orthodox path at its point of maximum mortality with the heterodox path at its point of minimum exposure. Pre-contracts are not revenue at scale; a 50-tonne pilot is not an industrial ramp; 87% of 2029 production under letters of intent says nothing about the ability to produce in 2029. The real test of Edonia lies ahead of it.
So the claim is narrower, and more interesting, than a victory lap: when the large cheque finally arrives, the residual risk it has to absorb is structurally lower than what the orthodox sequence carries at the same point. It is a judgment about where the remaining risk sits, and who has shown they can retire it.
The lesson worth keeping
In physical industry, capital efficiency is not a polite word for scarcity. Applied deliberately and early, it is the discipline that retires the structural risk most companies need tens of millions to absorb. By the time the money for a factory arrives, the unknowns should already be known, the pilot should already be running, the buyers should already be waiting. The risk doesn’t vanish because a team got lucky; it vanishes because the team, and the fund alongside it, hunted it down line by line.
That is the heterodoxy, and it is not really about food. It is about what European deeptech looks like once we stop mistaking the size of a round for the strength of a company.
A new format for this V4V edition: two perspectives on the same trajectory: Hugo Valentin, co-founder and CEO of Edonia, and Marine Reygrobellet, the Asterion partner who led the investment.
On the strategy
Hugo, two years compressed like that can’t have been linear. You didn’t start out planning to do it this way.
Hugo Valentin. No. When we started Edonia at the end of 2022, our first plans were largely inspired by the dominant foodtech playbook: raise as much as possible on early signals, then use that capital to build a technological and industrial moat. We even envisioned a much more vertically integrated company, including the microalgae production we now fully outsource. But as the market evolved we realised the next generation of winners would look different. The correction in venture markets wasn’t rewarding scale at all costs anymore, but focus, capital efficiency and disciplined risk removal. Our vision sharpened around one idea: focus obsessively on the biggest bottleneck in the value chain, the consumer experience of microalgae. That led to choices that were often the opposite of the conventional path - a B2B ingredient company rather than consumer brands, lean on sales instead of scaling teams early, sourcing biomass rather than integrating its production. Even on industrialisation the obsession became: how do we get the same outcome with less capital? That’s what led us to design our process around standard food equipment rather than custom-built infrastructure.
Marine, was there a decision where the fund and the team didn’t see it the same way?
Marine Reygrobellet. At one point the team was weighing a different sequencing: a smaller intermediate industrial site first, then a larger one later. The logic was understandable. But in practice that means setting up twice - two rounds of operational disruption, two moments of intense pressure on the co-founders and the factory teams, and a set of delays and frictions that get heavily underestimated when you’re drawing it on a whiteboard. We worked through it together, the three co-founders and the board, and landed on a strategy that goes straight for the right industrial scale rather than staging it. I’m now convinced it was the right call.
On what comes next
Hugo, the factory is the next chapter. What keeps you on your toes?
H.V. Every scale-up milestone has been emotional for the team — I still remember our first few grams in the lab, our first pilot batches, the first meals served to kids in a Newrest school canteen. Seeing something that existed only as an idea become a real product, used by real people, never gets old. What keeps me on my toes is that industrial scale-up is ultimately a synchronisation challenge. Building and financing a factory is one thing; making financing, operations, supply chain and commercial ramp-up all move at the same pace is another. A factory launched at the wrong moment creates problems, and so does demand without the supply to meet it. The next challenge is making sure every part of the system arrives in the right place, at the right time.
Marine, what would you tell investors who see a company that fits none of their templates?
M.R. The main concern I hear is market risk: B2B ingredients means a long sales cycle, and industrial buyers are conservative. My answer is always the same. Look at the quality of the commercial traction built with almost no capital deployed and a very lean sales team. And there is LOI and there is LOI: some are light signals, others are deep pre-contractual commitments with real detail and real engagement from the buyer. Edonia’s pipeline sits firmly in the second category, across both foodservice and food-industry clients, before the factory even exists. In my experience that is rare. It tells you something about the product, and even more about how this team sells.
Go deeper: the conversation with Hugo
This issue looked at Edonia from the outside, through benchmarks and the fund’s vantage point. The podcast takes you inside the founder’s head. Hugo Valentin sat down with me for the latest episode of Humans of Asterion.
They get into why the alternative-protein market crashed and why the recovery looks nothing like Beyond Meat, how Edonia bent standard, off-the-shelf equipment to a patented process instead of building custom infrastructure, and what it actually takes to scale a food deeptech without burning millions.
Along the way: the Newrest school-canteen test, raising a Series A in a sector most investors now avoid, and how Hugo would judge a climate deeptech if he were the one writing the cheques.
A closer look at one of our portfolio companies currently raising, at the intersection of energy, industrial data, and grid flexibility.
Companion.energy turns energy from a cost line industrial sites absorb into an asset they actively pilot. Its platform links contracts, site software and physical assets in real time, then steers flexibility every fifteen minutes without touching the industrial process. One client logged €300-450k in yearly savings on a €45k licence, a 3.6-month sales cycle, zero churn, and 2+ TWh piloted per year; ARR doubled in three months to €1.2M. Its moat is invisible: connecting to industrial control systems (SCADA) takes months - Companion does it in hours.
This round, led by Realyze Ventures and Pi Labs, is nearly closed. We’re holding the last few tickets for operators and investors with deep sector expertise. Contact: antonin@asterionventures.com