Julien-David Nitlech (IRIS): What Exotec taught IRIS about backing deep tech that lasts

1 Sep 2026 · 51 min · 19 chapters

Ask about this episode

Ask anything about it. ChatGPT or Claude reads this page and answers with the times it was said.

Connect VO and ask about every podcast you hear, including the moments you saved. Add to ChatGPT · Add to Claude

In short

Julien-David Nitlech (JD) of IRIS Capital explains how IRIS stays relevant across multiple tech “lives” and how it backs deep tech that can scale in Europe. He argues continuity comes from renewing teams and theses, not from never changing. IRIS invests in applied, vertical “tech stacks” with early customer traction rather than far-future foundational bets.

Guest backgrounds

JD is managing partner of IRIS (venture and growth tech firm founded 1986). He previously worked in strategy consulting, startups, and tech operations; he joined IRIS around 2013 and later took over with partners ~5 years ago. The host is Andreas (name not fully shown).

Key claims

Deep tech VC should avoid investing too early; look for an “anchor” in the market and expert-readable technology. Governance/humility matter: investors have limited leverage and must empower founders. AI thesis: focus on applied AI layers for verticals and infrastructure/middleware needed for automation.

Notable examples

Exotec (warehouse robotics; Skypod), Shift Technology (fraud/insurance AI), LeanIX (SAP), Kiribath (point solution), OpsMill/OpSync/PubSync (network/data operations layers), and Adaptive (AI “shovel” for human-in-the-loop).

Written by AI. May contain mistakes. Listen to the episode to check what was said.

Chapters

Tap a time to open that second in VO

Introduction to Julien-David Nitlech and IRIS

0:06 to 1:54

Get to know JD and the history of IRIS Capital in the VC landscape.

“Last year, 3 ,400 attendees, more than 400 speakers and 330 plus investors actively scouting deals.”

The Evolution of IRIS Capital

1:54 to 3:56

JD discusses the resilience and transformation of IRIS Capital over decades.

“One, introduce yourself to everyone, tell your story because I didn't know about you until we hit each other.”

Investing in Deep Tech: The IRIS Approach

3:56 to 6:43

Explore what IRIS looks for in deep tech companies and market opportunities.

“You invested, you made a lot of money, then you make zero.”

Governance and Talent Management at IRIS

6:43 to 12:24

JD shares insights on governance and nurturing talent within the firm.

“If you have that, what else do you want to get?”

The Importance of Humility in Venture Capital

12:24 to 14:00

Understanding the role of humility in effective VC relationships and decision-making.

“the organization through sharing the ownership, through driving their companies, through sharing the carriage so that they make money, so that they can buy you over, they can raise next funds.”

The Role of Humility in Venture Capital

14:00 to 17:04

Learn how humility plays a critical role in venture capital relationships.

“and some wealthy people from their ecosystem.”

Navigating Generational Changes in the Industry

17:04 to 22:04

Understand the impact of generational differences on venture capital strategies.

“now moving to your question how do you ensure that in a firm first of all it's your dna i mean what I described is how I've been raised.”

Introduction to Exotec: A Case Study

22:04 to 23:10

Explore the journey of Exotec from early challenges to becoming a successful company.

“And this is perhaps the moment where your values can stand out.”

Meeting the Founders: Insights from Exotec's Early Days

23:10 to 27:41

Discover how the initial meetings with Exotec's founders shaped their partnership.

“They were all building very ugly robots.”

Support and Growth Strategies for Exotec

27:41 to 28:01

Learn about the strategies employed to support Exotec's growth and international expansion.

“Like each time we had boards and they were explaining how they were developing this, how into this customer they were growing, but there was this issue.”
Show all 19 chapters

Scaling Exotec: Challenges and Strategies

28:01 to 29:19

Learn about the strategies Exotec employed to scale despite challenges.

“So together, they have such a great dynamics that they had already figured out many options and out of the options, the probability of success and the best one.”

Navigating Deep Tech Pitfalls

29:20 to 30:52

Discover the key pitfalls deep tech firms face and how to avoid them.

“We see regularly competitors from China, et cetera, dropping prices.”

Investing in Deep Tech: A Practical Approach

30:53 to 36:26

Understand how to effectively invest in deep tech with practical insights.

“One of the things that have been difficult for both deep tech founders but also deep tech VCs to navigate is that the journey for deep tech firm is very different from a normal venture-backed startup.”

AI: Current Landscape and Future Perspectives

36:27 to 38:23

Explore the current state of AI and its implications for the future.

“I understand the LLMs, I understand MCPs, but what it's going to become in a year or two, who would have told you that Anthropic would file faster with bigger revenue and bigger valuation than OpenAI a year ago?”

Applying AI in Industry: Opportunities Ahead

38:24 to 42:00

Learn about the practical applications of AI across various industries.

“and the short timescales actually being too conservative because that nearby future is actually even more close than you think.”

AI Applications and Industry Readiness

42:00 to 43:42

Explore how businesses can implement AI and the importance of data quality.

“the experience and interface for major in choosing this platform rather than another.”

Defining Moats in AI

43:42 to 45:26

Discussion on the challenges and strategies for achieving defensibility in AI investments.

“It's a four-minute question for something that should take 30 minutes, but now we know where you invest.”

Investment Strategy in Uncertain Markets

45:26 to 48:37

Insights into maintaining a focused investment approach in a volatile market.

“So I think right now there's no absolute answer to your moat question.”

The Future of Venture Capital

48:37 to 50:32

Predictions on the evolution of venture capital and the need for adaptability in strategies.

“we do best, focusing on a few, believing in them, making them scale.”
Hear the part that matters, and keep it.Open this episode in VO. Double tap your headphones to save a moment as you listen.
Get VO free

Transcript

Automatic transcript. May contain errors.

0:00This episode is brought to you with Luxembourg Venture Days, back on the 14th and 15th of October. It's where Europe's venture community comes together, private capital, corporates and high-growth startups, all under one roof at Lux Expo, The Box. Last year, 3 ,400 attendees, more than 400 speakers and 330 plus investors actively scouting deals. We'll be there with the EUVC studio recording on site if you're building, raising or deploying in your this is your room. Had to venture-days.lu and do tell them that EUVC sent you. After 10 minutes into the meeting, I knew I wanted to spend my life following these guys because like it made sense.

0:43So if you go too deep to a lead, the risk is you have a vision, but is the market ready for that vision that you never know. So building a firm is renewing the people on a regular basis, empowering them, giving a chance to their thesis, choosing them properly. They need to have a vision about how they're going to build venture, and their vision needs to challenge, contradict your status quo. Perhaps that will be our fifth life, as I told you. I don't think we've been eternal. I think we've died many times and been reborn many times. Perhaps we'll die out of what we do and then be reborn much more systematic and bigger, as you mentioned.

1:18Julien-David Nitlech:This show is not investment advice and the hosts of this episode may be invested in the funds and companies featured. JD, Iris has been around forever. It's nearly 40 years if I'm not wrong. You've lived through Black Monday, the dot-com crash, 2008, the European debt crisis and the 2021 growth bubble. And now AI, almost every firm from that vintage is gone. And in that time, you've backed Exotec, a French warehouse robotics company that most people would have passed on and watched it become one of the most important deep tech outcomes in European history. I want to ask you two things. One, introduce yourself to everyone, tell your story because I didn't know about you until we hit each other.

2:01And then secondly, I want to understand what it is that you see in companies and in deep tech and in Europe that other people haven't seen for a long while. So first of all, Andreas, thank you for hosting me. As I told you, it's a real pleasure to be with you. It's being among people who matter. So thank you for hosting me. Me, I'm JD. I'm managing partner of Iris. So I'll do two stories. mind is I was this bland French student during engineering school end of the 90s and I was trying to figure out what I wanted to do with my life and something that would last and make sense and I chose tech. I had affinities for tech but the truth is I thought it would truly change the world.

2:42We're in 1998, 1999 so it was the first bubble coming but the truth is I thought it would be meaningful. It would change people's life. I thought you'd have phones that would do everything. I thought you'd have robots, I thought you'd have AI, and look at where we are now. And to do that, so I've been in a large group, I've done strategy consulting, I've done startups. And 13 years ago, I crossed the path of Iris Capital at the time, which was a historical fund. Had no clue about VC. I knew about tech and startups and companies. Didn't know about the VC game, was not exactly my thing, the financials.

3:12But they told me, you know what, you should try to be an investor rather than an entrepreneur. And so I tried with a thesis that I'll explain. And so I've been doing it ever since. And I took over with the Cort and Eric Iris five years ago. We bought over it and we turned it into Iris, the next platform. So that's about me. And Iris Capital, it's a venture and growth tech firm. It started in 1986. It's had several lives. For me, the key to lasting 40 years, as you mentioned, because we will celebrate our 40th birthday, is not a continuity story. It's a continuity in the values and what I'll describe answering your second question, which is what is it about tech that we finance?

3:52But the truth is I think we've died many times and been reborn many times. You die out of the first bubble. You invested, you made a lot of money, then you make zero. Then people don't want to finance you. So you have to find new money and find new thesis. And you do that. And you make it, you get a couple of companies, and guess what? 2008 hits, and you cannot find money. So they had to reinvent themselves with corporates. There was much more corporate money in the 2010s. So that was another story of Aris Capital. and then I came in, quote Eric, and we had a story which was around vertical tech, which I'll come back to.

4:23And yeah, we performed, we took over the firm, we changed the capital, we changed the team, and this is our fourth life. So the key to continuity over 40 years is just that some bubbles say that some tech dies and some techs are born. The secret is to do it within your firm and do it without even knowing it, just because you have new generations, new theses, and you keep going. So it's a continuous change. And as for what we back, it's super simple. When I entered the VC market in 2013, I had been an operator, manager, start-upper for the past 15 years. And for me, Europe, especially French and Germany, where we've been operating since the 90s, it was super simple.

5:05Like we have amazing academic background at the time. Code was not done by cursor or AI. It was done by people who had to be very good. So number one, there was very good tech when there was not so good tech. So number one, you had to have a tech stack that mattered and seeing people coding at IBM, at Orange, at Vodafone. Some people decided that they wanted to do it for themselves because they wanted something that scaled, that would change things, that would be independent. The second thing is developing amazing tech is one thing, but if you don't have a market, If you don't have a way to figure out how this tech inserts itself in an existing trade, the typical sentence, this is broken, I'm going to fix it.

5:47So that's what we're looking for. We're looking for people who have both the capacity to build a tech stack that we think will be an asset, will be a leverage, will be innovative and a value, led by people who were in a market, who knew the trade, who saw something that was missing and managed to put their brick into it. If you do that, you have an infrastructure. You have something that at some point people will keep using. If you're successful, it becomes big. You mentioned Exotech. We have also Shift Technology worth a billion. We have also LeanIX, which we still want to be able to SAP. We have Kiribath for a point to be able.

6:20If it scales, on this premise, it becomes huge. If it doesn't scale, it still is part of an infrastructure, and it still can be acquired. It still can be partnered with. It still can grow. So really, what we've been financing in our thesis is people who build a tech that differentiates itself in scales on a market where they've seen the opportunity, where they think there's a trade, and they show it through early traction. If you have that, what else do you want to get? We're going to dive much more into verticals and investing in startups and building startups and that type of thing. But before we go there, I think that we should just double click on the story of Iris.

7:00And what has allowed you to stay around for so long? Because as you said, there's, of course, the fact that you have been reborn and that you've got to stay nimble and got to find your way all the time. And then there's a lot of conviction and that's called a culture and values, of course. But there's one thing that a lot of European firms or in general just VC firms get wrong is governance. How do you onboard new partners? How do you make sure that people are happy and actually want to stay there, that they don't go out and do their own thing? Can you talk a bit about the principles, maybe the things that you've seen this has been really important for you in your success?

7:41It's a very good question because you're hitting a spot here. Technology renews itself. People change, especially now. They're younger than ever. If you think that you're going to last forever, that's going to be complicated. I don't. So you need to adapt to the waves. You need to learn, which is a great thing about business. But also you need to stay on the market. And stay in the market means that when there are generational changes, it applies to you first. I am of the belief that a company is a reflection of its market, like a company that does robotics is led by people who will structure a company like a robot with compartments, with capacity to renew compartments.

8:21We're working on technology. So, A, we have to be a technology company. We need to onboard the best technologies that we see. And we're a people business driving new innovators and entrepreneurs. So that means that within our teams, we need to have new people onboarding new technologies and new visions. You need to renew the teams. So typically, when I joined Iris, I met them in 2013. The truth is I was meeting them because I thought I would start up. So I was meeting a VC because, hey, you never know. I know you. And the truth is the guys were like, you know, we've seen a lot of entrepreneurs.

8:53You should come work with us. But what they saw in me was I was 36. I had been in tech for the past 15, 16 years already. I knew everybody. I had my network. I had a vision as to what was good tech. Many of my friends had built amazing startups. And I just wanted to be part of it. So they took me in as an investment manager saying, well, you invest. It's going to be a cool game. Don't worry. It's going to be fun. It's just we're going to say no. Which they did. And it worked. They let me explore new theses. I thought everything sucked at the team because they were not daring enough, which was, of course, stupid because there were a lot of great investors.

9:26But they let me bring in my vision of innovation. and actually it stuck. And Quart and Eric and Grof had the same approach where they were approaching companies that had a real tech asset that took time to actually grow and compound. Eric did BrainCube, which was a great story in industrial. Quart did Talent, which I appealed at Nasdaq. He did like Cariba. So they onboarded young people to whom they gave the chance to actually follow their thesis, try it through. And it was disruptive. When I introduced Ledger in 2014 to my IC, Ledger was just born. Eric Larchevac came in shorts. They refused to invest, but they accepted to look at what was the future.

10:07And they could say they saw it and they could say they turned it down, which big deal, entire portfolio. So you need to onboard people. But not only do you need to onboard them, you need to empower them. It's not about like, bring me deals, I'll manage that. It's a long game. It's a game where you learn throughout five to 10 year cycle. You invest, you see what sucks, what you should have read as a limited signal. You see what works, what market changes appear and how companies adapt or do not adapt. You have a competent understanding of how your company companies, how you are on the board and do not speak too much, but try to figure out what's the key element.

10:42So you need to hire young people. You need to give them the chance. You need to give them a belief. The truth is after two years, you see the quality of deals they've brought. So if you're wrong, that's the law of the market. if the company starts to scale or bring something interesting, you're onto something. And the first thing you need to share. I won't enter the detail because it's always something people don't like to speak about, but I think that you need to share the carried interest. It's not about having 80 % of the companies and people hustling for you. It's about making sure that people learn, that they understand it's a 10-year game.

11:13I entered at 36. Here I am at 48, and one day I'll be obsolete. So people have time to grow to understand how you raise because it changes to how your company companies investing is easy. It's like being a French soccer team trainer like selecting a team. We have 70 people in France selecting the French World Cup team. There's only one coach that drives it and actually there's a very limited number of coach who want it. In our business, everybody can invest. Everybody can be cool. Everybody can hit high valuations. Scaling, addressing issues with founders, exiting, only a limited number of people do that.

11:52And it takes 10 years to know. So building a firm is renewing the people on a regular basis, empowering them, giving a chance to their thesis, choosing them properly. They need to have a vision about how they're going to build a venture, and their vision needs to challenge, contradict your status quo. They need to be daring and say, you know, you've been doing that, it's great, but the world has changed, dude. You've got to do something else. Probably because you need to work with them, so they need to be polite, but they need to challenge you. And if they do that, you give them a chance. And if it works, if their company scales, you let them scale with their companies and the organization through sharing the ownership, through driving their companies, through sharing the carriage so that they make money, so that they can buy you over, they can raise next funds.

12:34And honestly, the 40 years of our race, I think, show that. There was a first life, then it got out of the Caisse des Dépôts. Two people emerged and took it over on 21Pierre. And five, six years ago, they wanted out. and Kul Terek and I had done a decent run at generating performance over the past 10-15 years. So they let us and then actually they pushed us to take it over. So that was a great initiative. And honestly, when I hire people I raised today, I'm trying to figure out who's going to be there in 10 years to take over me. I want to ask you something about talent growth and venture and specifically on humility, because you said an important thing, which is that anyone can invest and you're absolutely right.

13:12The problem I sometimes find with more junior VCs is that they haven't been hit in the face with the humility that LP fundraising gives you. It's very hard to learn that true humility. And I can give a funny story here, an anecdote. I think that I've shared it with many and it resonates, which is that I tend to do events that are primarily partner only. And as an example, when I do an event with a group of robotics founders and robotics exited founders that are super good, super humble still, all of this, they have a beautiful profile in that sense. And I then bring together a group of VCs to come into this group of robotics founders and robotics exited founders and some wealthy people from their ecosystem.

14:05then what has happened and it's i'm using this as one case but i've seen it multiple times now whenever i accept bringing a principle they tend to rub people off the wrong way because they've almost always been a little less humble than what i would say that the partners are Do you think that this is, and I know this with firms that I truly respect, so it's not about them having picked bad people. I actually think there's a systemic thing there that it's so easy as an investor to get up on your high horse and think you're the king if you're not being constantly humbled by talking to people that are saying, no, I don't want to invest in your fund.

14:53can you talk a bit about that what you do to make sure that your own principles and associates carry humility with them when they engage with founders and other people in the ecosystem so and first i repeat i really thank you for hosting me because these are the best questions you need to ask for fun how do you maintain a state of mind i mean it's so i got so many things to say on that question i'm sorry so number one It's a job where we don't do anything. We raise funds, that's true. But then we meet people, we try to understand what they do, we try to convince them we may be a good partner so that they choose us, make a term sheet, then they drive their companies, we have boards, sometimes they call us, they ask for our opinion.

15:36And honestly, it's what it is, it's an opinion. So based on this description, and even exits, I love when VCs say, well, I exited this company, we exited ****. I mean, like founders, mad people, we actually are the borrower. You know, it's been a long time, let's exit. But I think they've been saying their companies, they've been negotiating the future of their people, their own future. So we don't own much in this business. We don't own anything. But, and even the money we invest is not our money, except for our little part. So number one, I think humility should be embedded because we have so little leverage, so many things to do, which are obvious, but very hard to do thinking you're the best at doing them it's not as if it were like a boat race or formula one race for it it's like and like typically in france there are about 100 120 firms doing vc right now how many unicorns a year in france one two three so regardless of the how the why of your question is this is the equation so if you can take any pride out of that besides the fact that founders will say it was good working with you which is going to be your best the two rewards you can get out of this job is professional one you made money and that's rare and the second one is founder saying honestly it was great working with you because that's what it comes down to forget these two things lps are happy founders are happy you can press repeat so now moving to your question how do you ensure that in a firm first of all it's your dna i mean what I described is how I've been raised.

17:10Like when they hired me and I raised, they said, JD, you need to understand two things. Number one, you're going to push for a lot of investment, but there's an IC to decide for that. So here's your ego. And second thing, these are not your companies. You're going to be at boards. You used to operate things. You used to decide. This is not about you. It's about the founder. If you understand these two things, you can join a board. So that was the DNA that was set on the table when I joined. And it's still the DNA today. Third thing, funny story. So I've been an investment manager and I've been a partner.

17:43And of course, I kept saying, you know, I want to raise funds. I want to matter in this firm. And I accompanied companies. I accompanied the tech. We had boards. They came with issues. And I spoke and then they left. And then after my board, I went, I don't know, home, see my friends, go see a movie, play tennis, thinking I did a good job today. Then I took over IRS. And then there were plenty of papers to fill, money to find, people to convince on the market 2022 that sucked. So I called all the founders I worked with, Jeremy from Shift, Romain and Renault from Exotec, Yanis from Adot. And I said the same thing.

18:19I said, guys, we had lunch and I said, I'd like to apologize. We're working together for what, eight years, 10 years, it's been great. And each time we had meetings and you said you had issues. And I said, well, you should do this or have you thought of that? And you went back and I was like, I did a good job. And now, you know what? I have issues. I need to convince LPs. And I speak with people and they say, oh, you know what? You should do that. And then they go. And you know what? I still have issues. So I'd like to apologize because now I truly realize what it is. And each of them, they laughed their ass off.

18:54And they said, JD, you're one of us now. You're an entrepreneur. Welcome to the board. So, yeah, it's a layer cake. You start, you invest, you enjoy the ride. It's a good ride. And also you need to focus on technology. But I agree with you that the cycle is complete, that you truly have what makes an investor the moment you have to ask for money and justify it with your capability to invest in the best and to go on doing it and bring back money. The moment you do that, the money you invest is not the same. It's the same game, same people, same cool parties, same startup, same amazing innovation.

19:29But you realize that the money you got was hard to get, that if it doesn't succeed, you're not going to get any again. And that changes a little bit the picture, which for me makes sense because this is skin in the game. And moving back to my point about generational change, this is also what we share carried and shareholding of the GP so that people put their money into it. I mean, the first, what, eight years of VC, I pretty much spent one year of salary into carriage, shares, etc. So I don't think it's a comfortable game. If you're too comfortable, it gets complicated. So these are the few elements.

20:04And then we speak with people, generation change, Gen Z are different than millennials, are different than Gen X, which I am. So they're more assertive. They think they know better. And the truth is it's also good because they're already tougher. So you need to balance the cultural change that they represent with the fact that you're looking for core values, which are, you know, it's a simple business. The truth is we always have a debate with my partners. Why do you go into that business? Lots of people say to make a lot of money, which should be a driver. The answer I like is to back people who are able to do amazing things that I wouldn't have been able to do myself.

20:36That's an interesting answer. and then I think that there's on top of it that the heroes in our ecosystem or world and content world especially are all Americans and if you're anyone good in venture you obviously always have an ear to what's happening in Silicon Valley and and and not because it's more important what's happening there but just because of course you need to know you need to be aware not because that they need to teach you something but you need to be aware because it's part of the industry chatter. That then means that you can create some heroes that are not very good for you, especially not when you come and talk to real world Europe.

21:15Because it's one thing you and I, because we know that, okay, you're posturing now, you're acting like you're to math on the All In podcast, all good for you. Someone who is from the outskirts of Denmark has built a multi-billion euro robotics company. He will not think like that when you come and tell him how robotics work. I absolutely agree. And founders, it's also another way to rebound on what you just said in your question. Founders, they know. I mean, there are so many courting them buying coffees. So some of them only care about how big your ticket size, how great your brand, how international.

21:53Some of them figure out that it's going to take a while to build and they want to build with people that they can talk with and it won't work it. So I still think there's a segmentation, people have access to any kind of money. They choose their money. And this is perhaps the moment where your values can stand out. And US values are great. And once you have a company that goes international, I think you need these guys. But building companies in Europe, it's a slightly different setup. The start takes a little bit longer. You have not so fast access to commercial trade because you're in France, you need to go in Germany.

22:22You're in Germany, you need to convince these UK people. The scaling takes a little bit longer and then suddenly it pops up. And so this part may require also a culture that understands that markets are equal. Yeah. KD, now I want to talk a bit about Exotech because I've referenced many times robotics now. I come from a robotics part in the Danish ecosystem. And I think it's a very interesting case study for how you think and what you've done in the industry. Because, of course, Exotech is now a 2 billion plus company. but in the beginning it was not an obvious bet at all. I'd love to hear a bit about when you first met the team, what you were thinking, and then let's talk about the journey from there to here.

23:10So it's a textbook case of what we do. So thanks for taking this example. I heard about Exotec in 2016, 2017. They were starting. They were all building very ugly robots. You can see the YouTube of their first Skypod, which was complex. It was super early. It was pre-seed to be honest. So I passed. I was at the end of the 2012 fund. I didn't meet them. I knew about them. I really liked the topic. And okay, I said like, yeah, it's very early. It's unlikely. And while closing it, at the time we were, we stopped. We were investing in Israel where a partner of ours in 2017 offered that we look into fabric at the time, common sense robotics.

23:49So I had the time we're raising a new fund and the time was not managing partners. So I was spending less time on fundraising. And when I look at Common Sense for Viddick, what struck me was there's a market. I mean, there's a market for innovation. There was a question about automation, given the increase in retail, in short distance, grocery ordering. This will make sense. It's just that the fabric model, which was great at the time, was very concentrated. You had to invest a massive amount of money to have a full house and serve customers kind of like who could do work. So I felt it was a bad ratio and then I forgot about it.

24:24In 2017, I go to Lille at an event from France Digital, and I get the chance to meet Romain. It was a 20-minute meeting, but this is how this business works. In these 20 minutes, Romain arrived. First of all, he looked at me, scanned me about, like, what is it that you do? Will you understand anything? I said, so I explained what I understood about hardware. I had spent 10 years of my life doing hardware at Orange, building phones, regulating them, building platforms on top. So I spoke and I looked at his eyes and was like, okay, it's worth talking to you. You'll understand. Then he told about what he was doing.

24:58And when it says, he was like, listen, I was doing logistics and robotics and Renault was working at the same firm. So both had done GE and they had done logistic companies. And Renault, he was, so Romain was doing more software logistics. Renault was doing robots. And at the time they're like, we saw Amazon buying Kiva for 700 million. And honestly, it was very disappointing. Kiva is a bad solution. It's only one degree of mobility. We could do better. We can do a two degree mobility robots. The throughput would be amazing. So yeah, we left the company and we built that. So remember what I told you about way back.

25:35People who have an amazing technology background in Romain, when he spoke, it was code. I could see the code. And then he explains that he's done the market, he's done both. And honestly, Kiva is disappointing. And he explains to me. And the worst part is, I understand. Like, I'm not in his business, but the way he speaks is so precise. And I'm an engineer. Like, yeah, well, he's right. Like, he should do that. So at the end of these 20 minutes, I was like, listen, great. Like, if you do anything, call me because I believe in you. February 2018, I received an email saying, well, we spoke. You seem to understand.

Read the full transcript

26:11I'm going to speak with you. He arrived the next week at the office, like, sat and started saying, well, listen, you were one of the people who I think understood what I said, spoke for an hour, explained how he had signed his first contracts, was starting to deploy hardware logistics, how the issues were how to maintain the throughput, that the contracts need to be validated once he would have done this or that. And he explained how he structured his company. And what struck me at the time, and he knows that, was the way he explained the structure of his company was the way you explained the composition of a robot.

26:46There's the brain, there's the operation, there's the deliverer, and there's the manage part. It was not delivery, GMA. It was do this, that, that. That was the name of his organization. And it's always the same in this business, at least for me. After 10 minutes into the meeting, I knew I wanted to spend my life following these guys because it made sense. So we went fast. We had a chance to have pre-studied the market. We went to Logimat, Germany, because we wanted to make sure. We interviewed, they saw us there. In three weeks, he had a term sheet, which made us faster than everyone else. But we're all over cheap on valuation, which annoyed Roma very much.

27:24So he brought other people. But as Antoine, my foreign partner, said, good deals are very expensive. And the price he wants is the price he wants. We're going to pay him. So we did the term sheet and we worked. So that's part one, the deal. Ever since we've done the deal, this company has been built like a self-taught robot. Like each time we had boards and they were explaining how they were developing this, how into this customer they were growing, but there was this issue. You all, as a board member, well, have you thought about this? And each time was like, well, JD, we're glad you asked. Slide number 12.

28:00This is what you can do. So together, they have such a great dynamics that they had already figured out many options and out of the options, the probability of success and the best one. And that's been many years with them now. So they grew, they saw the shoes. They took us for what we're, I think, good at all, at least what we like to do, which is they were starting to grow their contracts. They were ready to deliver more customers and go international. And I think this is a good setup for our race. We're pan-European. We're pre-international. interview like so really the job was about helping them just figure out the priorities that they wanted to choose because we didn't choose them we just laid out their options and discussed them and they chose and also focus on what you need to build to have a structure that scales like i always say we grow the structures we structure the growth so it was really about hiring the cfo cbel lasted with the company for four to five years it was about choosing who you started delegating the first issue so that they could focus on the next go international and from the moment we entered where i had three contracts to deliver and less than one millionaire to the moment that 83 north floral joined that's an uniqlo in the lands and the covet struck and the day the covet struck they were supposed to deliver uniqlo in japan so i called the roma i was like how are you going to do it he's like i don't care i'm already on a plane to japan so there was a curfew there was like you had to stay at home the guy was on a plane to jetline how did he do that got no clue but that's pretty much the story so and then that's the story they scaled they we've had issues we've had issues about competition we've had issues about prices each time they analyze they fix they improve so yeah that's the story of exotype because the market is deep there's only one core competitor who is auto store which uses the same product for the past 25 years they tried They had renewing products.

29:57They had suits with other competitors. We see regularly competitors from China, et cetera, dropping prices. But for the moment, we remain the most consistent. We're not the cheapest one, but we maintain price at the competitive level. And we provide a level of experience and deployment that is enhancing. We announced last year Skypod 2, which actually has been several years in the making. The performances are now outstanding. And so the same way they signed Uniqlo, they signed huge contracts in the US and the company scales. What's hard with this company, though, is that every year on January 1st, you start with a book to fill, to find new warehouses to sell, new deployments.

30:35And then you have milestones to deploy these warehouses. Suddenly you have tariffs in the U.S. Suddenly you have a change in the dollar rate. Suddenly you have a shortage in certain components. It's been now eight years with them. They work out any time. You with Iris, you specifically, but also Iris as a firm, have a notorious track record in deep tech. One of the things that have been difficult for both deep tech founders but also deep tech VCs to navigate is that the journey for deep tech firm is very different from a normal venture-backed startup. Can you talk about the key pitfalls either by using Exotec as the case study here or in general across your portfolio?

31:19the key pitfalls, the key things that make that journey super hard and how you can best overcome them? Because there are, of course, some general learnings. One can be around the capital stack that you're building for yourself. One can be about how you navigate investor interest, whether you take on generalists or you take on corporates, like go wherever you want on this question. So first of all, I mean, we do deep tech, but I don't think for a deep tech fund in the sense that we don't do quantum, we don't do, we haven't done like Mistral. Companies that we think are going to be so long and so capital intensive that it's going to be difficult for our scale because we match smaller sized funds for the same reason that for the deep techs we go for, for just the very vertical techs we go for, we know it takes time.

32:11So answering your question. First of all, we don't do pre-seed. We invest at a moment where we can see a tech stack, which we evaluate ourselves still. Back to the generational point, you need to keep having investors who can read the tech by themselves. That's our thesis. But also at a moment where these techs can be evaluated by their first customers. So due diligence, expert reviews, tell us whether it makes sense. When I looked at Shift in 2014, I knew heads of insurers who told me, well, there's an issue about fraud. I don't know about these eight people doing AI because I don't see how an eight people company doing AI can fix my issues, but we have an issue around fraud, to which I answered, well, the eight people doing AI is my stuff.

32:55If you tell me there's an issue around fraud, I'll take it. So moving back to your question, which I think is a very good question, we often mix huge bets, like transformational bets, which can be one or zero, with, I wouldn't say reasonable bets, but stuff that is already anchored in an environment. We have financed semiconductor and silicon stuff in the 2010s, and our learning was we were wrong from the start because the format of the company, which had plenty of other issues, was just not the right format for the industry. So if you go too deep to early, the risk is you have a vision, but is the market ready for that vision that you never know?

33:41So my first answer to your question is we as investors don't operate as a as a full bet, super far away investor. We focus on tech stacks and innovation where we already see an anchor with our market. This anchor doesn't need to be 1 million, 5 million euro revenue. It needs to be something developed enough so that the people whom we consider as reference point in these industries show a sign that, yes, it's disruptive, but I can make use out of it. So we do deep tech, but if we do a deep tech that's already mature enough for people to say it's the right format, actually I should try it. And yeah, if it worked, honestly, it could work.

34:26So rather than foundational technology, we do applied technology. We've done that typically recently. We announced OpsMill. So in the network and infrastructure space, we know that AI will be huge. We know that historically you have so many stacks from Juniper, Huawei, Ericsson, that is very proprietary, very heterogeneous. You know that for AI, with all technologies, you will need the middle layer that makes it homogenous. When we saw Opsmule two or three years ago, it was way too early for us, not for preceded funds, because we're like, it could make sense, but they need to develop something.

35:03When we saw them again in September, October of last year, that developed the tech stack, it could be downloaded. I could look at the code. We could have experts saying, well, if they truly do what they do, this is what we need to be able to operate automations on our network. We have access to orange experts. We have access to a lot of experts. And so it's very deep tech ops. It's like operating a mid-layer for operating infrastructure, data centers, but it's a layer that turns it into graph, which is the basis of operating AI. And so you downloaded open source and then you need a package. So it's kind of an old school open source model.

35:39But the signs that we saw of consistency and stickiness with potential users made us believe it could be huge. And for the moment, traction confirms it. So our approach to Deep Tech is if it's too far away, if it takes three years to make something concrete, it's perhaps not for us. There are people who need much bigger funds to be able to do many of these deals. We do five, six, seven deals a year. We see ourselves as a boutique working with founders. So these five, six, seven, eight deals a year need to be focused on stuff we understand. And we see a first traction so that we can help and accompany the company.

36:18So that's how we do it. So there's no perfect bet, but there are... So right now, typically on AI, there are plenty of ways to address AI. And talking about one of your previous points, if I'm honest, I'm not sure I fully understand. I understand the LLMs, I understand MCPs, but what it's going to become in a year or two, who would have told you that Anthropic would file faster with bigger revenue and bigger valuation than OpenAI a year ago? Like, would you have believed it? So I don't really know, but what I know is that everybody will use it. They will use layers to make their data homogeneous and use it simply.

36:54So that is a deep tech play because you go deep into the layers, but you can already guess that people will need to operate stuff that you don't exactly know how it's going to operate. So no perfect answer to your deep tech question. What we know is for each vertical health with OPR for industrial, which we do a lot for data operations. We've done PubSync, we've done Covery for cloud. We see what's going to happen in three, five years. We're not visionaries. We just see that people will automate and the current stacks are not fit for automation. So people who have a layer in between that's been developed deep and that shows sign that within the next two or three years, people could download it and use it to operate the automations.

37:39So we do deep tech, but with a level of maturity that is not 10 years away from us. There's a matrix I like to use, which is the time matrix. You look at something. If you think it's going to take 10 years, usually it's going to take longer. If you think it's going to take five years, you may be right or you may be wrong by two years. If you think it's going to take two years, it's going to take less. Our business is to invest in the stuff that you think will take three to five years because plus or minus the LAS, this is the time for VC. If it has to take longer, It's either you missed up, or it's because it's becoming huge.

38:17That's an interesting framework, trying to apply it in the future when I think about things. Because I think you're probably right. I think you're very right on the long timescales oftentimes being overly ambitious and the short timescales actually being too conservative because that nearby future is actually even more close than you think. Let me ask you, because you took us there and I wanted to go there. How do you think about AI today? Because obviously you're in the middle of the French revolution, I was about to say, because it is truly a flourishing ecosystem and you're doing a lot in it.

38:50So I'd love to understand from you, how do you think about your AI thesis today? You described it a little bit. You like the going deep into verticals with AI and you think that there's going to be a lot of opportunity there. But if you can describe it a bit broader. So, sure again, the first thing is AI is taking over the world. it's obvious and it's really i mean i started 1998 so i saw the internet and and then i saw the 2008 platform thing i think what we see is much more mature and it's a level of scale that makes 2000 look like it was cute so it's happening but also it's the first time that it it was the same in 2000 it's so centralized in the us i mean we have mistral in europe we we have a couple of companies back for us like but the truth is like right now the core of the nuclear explosion is the silicon valley and you're at open ai and you decide to go out and do a little because you saw that so it's very dense and in europe i think we have amazing know-how you can see all the announcements about developing r d centers hiring teams so we have amazing know-how in terms of math and rhythms and technology, but we're a little bit far away.

40:09So the choice we've made at Iris, again, this is something where honestly, every day I wonder what to do or not to do because it's impossible to read the future, but no, in five, ten years, everything will be automated. In two years, I don't know, because the prices right now that Anthropik or OpenAI are applying are increasing strongly. The moment doesn't change anything. But so foundational AI is a hard game. Same as what I told you about deep tech. It's you need to be sure about the models. You need to be sure about the play. And meanwhile, you have open AI and it's insane. But you know it's going to apply to verticals.

40:51So we've chosen to make AI applicable for the moment. So our strategy is really, there are three paths. Either you go deep infrastructure layer, like this brick that is amazing, that provides like tabular data, like Preal Apps, which was typical play or fundamental. Okay, we can do that because if it works, people will send huge contracts and can be acquired by us. We did Adaptive on this. Reinforcement lightning is huge. Adaptive is scaling like crazy because companies deploy their own models, but they have a problem of human in the loop of making sure that the algorithm is learned. So Adaptive, it's a shovel.

41:29You put it, you make it work, and it operates on your algorithm. The second thing is making AI physical, applied. So you have all these algorithms, but the thing is people need to use them in the end. So that's why, for example, I don't subscribe exactly to the term software is dead. In a sense that what makes us discuss today together under us is software. It's like a great website platform on a web page, there are APIs. So the structure is more important than ever. And you know what? or discussing it, the way we started, the way I could change the settings, the experience and interface for major in choosing this platform rather than another.

42:06So personally, I still think software will be huge because you will need an interface to operate your data. The question is the quality of data you operate and then which algorithm and which automation you plug it into. So we focus on applied AI. We focused on businesses where we think that they have a data issue, they have a structure issue. And within the next two years, they need to figure out how to apply AI as you pop corn. So what we care about is tools that allow each industry, each customer to be ready for their next automation and their next productivity gain. If you're super cool, you go into LLMs, you do international AI.

42:44If you're just good, you find intermediary layers that allow people to operate what they already have. people who are on AS40 and ZOS, large French German corporates having their own operators and operating through third parties, they won't all of a sudden migrate to the cloud, especially now that you have issues about sovereignty in your own cloud. So having tools that allow you to work on your cloud and go out to operate a few operations on clouds that provide your computing power and then bring it back into your own cloud, that makes sense. That's why we funded PubSync and it grows. I told you about AppSmeal.

43:18So really our approach is to choose verticals, to see the companies, the markets, whether they are ready to implement AI, to see what's missing and to bet on these layers. It can be operational. On OPEA, you optimize the way you allocate resources within hospitals or care facilities. It can be much more theoretical like Opsync or PupSync. We should get to a close of this podcast, but I've got to ask you, how do you think about moats in AI? It's a four-minute question for something that should take 30 minutes, but now we know where you invest. How do you think about, okay, this is actually defensible?

43:55Some people say it's only about moving fast and there's nothing else in that. Others, I think, would say that, well, there are definitely places where it's so hard to build that you're absolutely sure that if you get it right, you will do well because many others will have that knowledge and for sure people will not be building it in-house themselves. It's a super hard question and yeah, we could spend hours. The fact that we have four minutes makes it easier in a sense that I don't know. Honestly, I've rarely had less certainties than right now and I'm not specialized in certainties in general, if you ask people around.

44:30What I think is this, I think that winning for companies, first of all, will require high quality assets. The game where you just develop a tech and it gets adopted, on the short term, if you're making a short play, it can work. But on the long term, you can see typically with Anthropik, OpenAI, Gemini, etc., you see that it's a long game. I mean, we forget that OpenAI 2016, Anthropik seems early, but actually it was the Google team creating Transformers, so it's already almost a 10-year game. It seems very early and fast, it's less than that. So good, if not amazing tech quality wins, that talking about deep tech and deeper layers, technologies, It compounds over time, and you need to have a quality asset that adapts because there will be many changes to the LLMs, the same way that 10 years ago Facebook kept changing their API, so if you plugged, you needed to be able to adapt.

45:26And for that, you need very skilled entrepreneurs who have a capacity to not pivot because normally you found your thesis, but to adapt to the game, and that's what we've seen with Romain Renaud, that's what we've seen with Jeremy, that's what we've seen with Founders Way Back. So I think right now there's no absolute answer to your moat question. I think it really comes down to having a qualitative technology that already applies to a data set and an environment that is vertical and that people will need because I think it will compound. Led by people who have a full understanding of their stack right now and so can adapt to any delta changes that will occur every week on their business to keep their stack up to date.

46:06We were discussing with Escape last week, so we funded it. It's API security through AI that we funded in 2023. They're much bigger, they raised a huge round. To their opinion, they've already pivoted twice. They started at the CI CD tool for API security, and they moved on to penetration testing and dynamic application security testing, DAST. And now they're good. So it's a small change to the technology, but there are leaders in AI penetration testing. AI penetration testing did not exist eight months ago. I got to ask you a final question because you just said uncertainty is growing everywhere.

46:46You've also said earlier that you're a boutique firm that like to do five to six, maybe seven deals per year. That means that you're quite concentrated as a VC. In a market where everything is very uncertain, Have you thought about making a fundamental shift in the investment strategy and saying, we've got to take more bets, we've got to go for lower ownership, we've got to diversify a little bit? Because being boutique and concentrated is very hard in a market that's so dynamic and shifting as it is right now. Our 2022 portfolio, which is finishing this year, will be between 15 and 20 companies.

47:23So it's not very diluted. It's not very concentrated. It's in between. But I agree that it's not 40 companies. I'll tell you two things. First of all, I had an amazing conversation with a very good LP, whom I asked, this is what I do. Do you think I'm right? Because, you know, sometimes you get lonely. And he said, JD, look at this fund. This fund does bets that are like 20 millions right away on stuff that they barely know. You think it's not going to work, right? It's never going to work. It's one of my best performers. So the only thing I can tell you is this, and honestly, I'd like to thank this person for saying that.

48:02What we care about is what you believe in and how much you're dedicated to it and how good you've been at it so far. And if you believe in it and you've been good so far, there's no reason you wouldn't do it again. So everything I've been telling you for the past hour is about choosing carefully the text that we back, the people that we back, thinking that makes sense in the five years ahead and being adaptable. And we see 2000 companies a year like everyone else because we have a good deal flow. It's just we can only back four or five of these people. This is our model. So my first answer to your question is, I don't know if I do it well, but this is what I and what we do best, focusing on a few, believing in them, making them scale.

48:46It may stop one day because it's not performing anymore and that case money will go somewhere else. But really believe in that. And the second answer though, to make you right, I think, so I'm thinking a lot about the future of DC. I think we're headed towards a market crash. As I told you, more than a hundred firms in France, how many scalable assets think? It's been a great 15 years of negative rates and new firms. It looks like 2000, time for a reboot. Some people will raise, some people will not, perhaps will be raising, perhaps will be not raising. and that will be changing our industry. And I don't know what it will look like, but my belief is more than ever, you'll still need to be good at what you do.

49:24And to be good at what you do, you need to understand tech and you need to have vertical thesis. I don't think you should be a vertical fan with one thesis. I think you should be multi-vertical. I think our job more than ever, given the AI, given the quality of entrepreneurs, is to be good technologists. We need to understand technology on many verticals and be able to adapt. So I'd say that we'll need to keep on being vertical, rigorous, driven by belief. We just need to have more expertise to be able to multiply the verticals because with AI, all verticals will be transformed from biology to robotics to industry to…

49:57So, we'll keep on doing what we do, but perhaps a way to make you right and go towards a wider portfolio is to make more vertical plays side by side. So typically, managing… Today, we manage what? 150, 200 million fund max on venture, 350 on growth. Probably tomorrow we need to raise this one or more and we won't make bigger tickets. We just have more teams being specialized and perhaps that will be our fifth life, as I told you. I don't think we've been eternal. I think we've died many times and been reborn many times. Perhaps we'll die out of what we do and then we'll be reborn much more systematic and bigger as you mentioned.

50:33I think you're absolutely right. The most important thing is that we stay nimble, stay super close to what's happening in the market and keep adjusting and keep questioning what we're doing, because then we will end up in a better place than we were yesterday. JD, thank you so much for joining me on the podcast today. And Rustin, it's been a real pleasure, really. Thank you.

From the publisher

Exotec’s journey from an early-stage warehouse robotics company to France’s first industrial unicorn, now valued at $2bn+, reflects what IRIS looks for in applied deep tech: differentiated technology, a clear market need and founders who understand both.

In this EUVC episode, we are joined by Julien-David Nitlech, Managing Partner at IRIS, to explore why Exotec initially seemed too early, what changed his mind and how the company continued to evolve through international expansion.

Julien-David then takes us inside IRIS, the European venture and growth firm founded in 1986. He explains how the firm has navigated four decades of technological change by renewing its teams, ownership and investment thinking while preserving the values at its core.

What we cover:

  • Why Exotec’s early promise was not initially enough for IRIS to invest
  • What Julien-David saw during a 20-minute meeting with its founder
  • Where IRIS draws the line between applied deep tech and capital-intensive foundational bets
  • How early customer validation helps IRIS assess a technical stack
  • What AI companies need to build lasting value as technologies and markets evolve
  • Why IRIS may be approaching what Julien-David describes as its “fifth life”


Follow EUVC for more conversations with the founders, investors and operators shaping European tech and venture capital.

More from EUVC

All 626 episodes
Julien-David Nitlech (IRIS): What Exotec taught IRIS about backing deep tech that lastsEUVC · 51 min
Listen in VO