Solving Data Fragmentation at Scale with Michel Tricot, Co-Founder & CEO of Airbyte | Inside Summit 2025 [LIVE]

when you're doing anything related to data, you have to think of it as a manufacturing process. You have raw information and you need to extract something that will be able to deliver value. And doing it at internet scale, not that many people can say that they had the ability to work with petabytes of data. Before that, it was just Google and Facebook that had this scale. And here we were able to power very strong value for our customers by just moving data from point A to point B.

Welcome back to Season 13 of The Room Podcast. Claudia, can you believe we've been doing this this long?

No, it's been 5 years. So many episodes, so many incredible conversations telling the stories of some of the world's most iconic founders and funders. And whether our listeners are new for the first time today or have been with us from Season 1, we're so grateful for the community we've built.

Completely. We've had the mission to open the door to the room where it happens for those entire 5 years. Been able to do so over 120 episodes with hard-earned lessons from some of the top technology and consumer founders of our time.

And we've been through a lot personally as well. We have indeed. I built, scaled, sold my startup Priv. You were our first investor, so you've been there through the entire journey. And you've also recently graduated from HBS. You're now building at Vercel. And through it all, we've brought the funder and founder perspective.

Yeah, it's really been an incredible journey, and I'm thrilled that I'm back full-time in San Francisco, and we're probably here to stay. Uh, that means that there's more for us in terms of growth, both inside and outside of the physical room. But what's important to have really out there is these stories. And we've been able to do so over incredible guests from the CEO and founders of Flexport, Perplexity, Particle, Zapier, Zillow, Vanta. I could go on. These guests and their enterprises have generated over $100 billion enterprise value and really, frankly, moved the markets.

And growing, especially in 2025 with everything that's going on in AI. So I'm so excited to share Season 13 stories that are maybe a little bit more AI-focused than past seasons.

Absolutely. And you can find us IRL and URL URL, especially with our events such as our annual conference, Inside Summit, or upcoming SF Tech Week events and beyond.

And if you want to get information around those events, subscribe to The Room Podcast and subscribe to our newsletter at theroompodcast.com.

Perkins Coie supports the most innovative entrepreneurs and investors in fast-moving and high-growth sectors, addressing their myriad of legal needs. But the firm doesn't just provide end-to-end legal and business counseling to its startup clients. It also facilitates introductions to key advisors and sources of capital. Perkins Coie's interactive website, Startup Percolator, offers access to programs, resources, and rich dynamic content designed to assist entrepreneurs on their startup journey. To learn more, go to startuppercolator.com and perkinscoie.com.

This podcast is brought to you by Mercury. The banking platform businesses like The Room Podcast use to simplify their finances. It's time banking did more than hold your money. Now it can. With Mercury, you can pay bills in seconds, close the books faster, and even send invoices. Not only does Mercury do away with a patchwork of tools, it eliminates guesswork, giving you complete and accurate visibility into your business's finances, all from one account. Apply in minutes at mercury.com. Welcome back to another episode of The Room Podcast. Today's episode is part of our Inside Summit special series, our way of bringing you inside the room, literally, with the founders and funders who joined us in San Francisco a few weeks ago. Inside Summit is our annual gathering where we bring a couple hundred of the sharpest operators, investors, and builders together for a full day of connection, real talk, and the kinds of conversations you rarely get anywhere else. It's intimate, energizing, and one of our favorite days of the year. One of the standout moments from the summit was our keynote with Michel Tricot, co-founder and CEO of Airbyte. Michel has helped shape the modern data stack and built one of the most influential open source companies of the decade. In this conversation, he opens up about scaling Airbyte, navigating hypergrowth, community-led product building, and the future of data infrastructure in an AI world. We're excited to share this session with you, so let's get into the room with Michel Tricot. Let's get started. And we love to start at the beginning. Michel, tell us, where did you grow up and how did that shape your view of the world?

You can probably tell by my accent that I was not born here. I was actually born in Switzerland. I lived for 9 years in Switzerland, in Lausanne, beautiful city, beautiful lake. Still have a lot of good memories from there. Then I moved to Strasbourg, east of France, and then Paris. And I did 999. And San Francisco is what? 14 years now. So that's the city that is bidding the Ozarks.

Incredible. Did you always think you were going to become a founder?

Always. When I look at what I was doing when I was in high school and middle school, I was always trying to create side businesses and things like that. My first computer, I had to work the whole summer harvesting tobacco to just get the money. I did not have VC money at the time and That's what I did. So the grind of getting what you need to do what you want, selling games, selling movies to my friends in high school, that was a great business.

The answer to this question is always definitely not or definitely yes. So excited that you're on that side of the spectrum. You've lived in a bunch of different places. You were writing code in Paris, you were writing code at Siemens, you were at LiveRamp when it spun out from Rapleaf. Tell us a little bit about these early career milestones and how you made the choice to move to San Francisco, which now sounds like is home.

I've always written code since I got my first computer in the mid-'90s, and I've done that for my whole life so far. And 2007, I was fortunate to be able to do my internship in Princeton at Siemens, doing a lot of like pretty intense data science before people called that data science, working on like brain imaging, et cetera, et cetera. And for me, that was my first foot into the data world because when you work with researchers and everything, they have data everywhere. They have brain scans that are laying on like a random folder somewhere and suddenly you realize, wow, people keep asking questions about where can I find this data, where can I find this MRI? And that's when I started to build my first way of what does it mean to organize data at my scale as an intern. To me, that was really the first time I understood the power of being good with data was, I was at FactSet in Paris. So after Princeton, I went back to Paris and I wanted to do finance at the best possible time in 2008. I thought, yeah, I'm gonna write models for like finance and blah, blah, blah. And no, FactSet is a data company for, so basically it's like an equivalent of Bloomberg. That's when I really understood the power that you have the moment you can centralize all these little data silos. I was working like estimate databases. Analysts pulling their report out. But if you just read one report, you don't have the full picture. And that's when we started to just collect all these different reports across thousands of different agencies and just centralize it into our database. You realize by just having access to all these pipes, all this data, you can actually infer a lot of things from the market. And to me, that was really my first piece of, wow, this is really powerful. And I got there by luck. And then Lyreamp, like this one was working really at internet scale. That's when I, I understood that when you're doing anything related to data, you have to think of it as a manufacturing process. You have raw information and you need to extract something that will be able to deliver value. And doing it at internet scale, not that many people can say that they had the ability to work with petabytes of data. Before that, it was just Google and Facebook that had this scale. And here we were able to power very strong value for our customers by just moving data from point A to point B.

You spent a lot of time at LiveRamp and RydeOS, and to your point, having the scaled level of data, what were some of your insights there? How did that bring you to this feeling that there was needing to be something more in your data stack?

Data lives everywhere. And if you look at what the world looked like 15 years ago, it's very different. 15 years ago, you did not have that many SaaS applications. You did not have that many places where people could store data. It was generally like a flat file, a database, and an Excel spreadsheet. But as the world is moving more and more toward technology and like very specific services, then data lives everywhere. When I was at Lightroom, I remember, I think it was in 2012 or 2013, we did this hack week moment where I built a way for us to crank out a huge amount of connectors. And the product went to something that we were releasing to our customers. Where, well, we were able to just touch data everywhere they had some. So, but it was really built within a company that was building a product. And for me, WriteOS was the same thing and going into more like the, okay, every other company have the same problem. Whether you're doing analytics for yourself, whether you're building an AI platform, you need to have access to that data. How do we make that technology available for everyone?

Yeah, and you have made it available for everyone because the foundation of Airbyte is actually open source, which is really interesting. How did you go from where you were working full-time to launching an open source solution in building Airbyte?

Initially, we're not thinking of doing anything that was open source. When you start a company, you need to have some love for the domain that you are working because lots of ups and downs. If you don't believe in what you're building, or if you just say, I'm just gonna do something, make money, and I don't care about the product, it's not gonna work out. Like the first obstacle, you just give up. I deeply care about how good I can make the life for other people when it comes to removing this pain of integrating data. And the thing that happened at the time was we were thinking of a fully paid solution. And then I went back to what is the developer journey when they are actually building data pipelines? Like data pipelines is not a goal. No one wants to do it. The same way when you're building a city, no one wants to build sewer or like roads. Just, you have to do it because it allows you to move water and it allows you to move people. So in general, the way it works is that people just build it internally. And when they do that, they just put their finger in an engine that is going to rip them apart. How do you do when you don't know when projects are being created? And open source is a way of capturing that build motion. Like, I'm an engineer, in a way, in a good sense, I'm lazy. So if I can do something without having to build it, I will do it. And in general, like, open source is the solution. And that's, for us, it was really like, how do we distribute a product that has value and make sure that we are the first brick when people build systems?

That makes a ton of sense. I now work at Vercel and we also had our roots in open source, so there's just such like a profound community around this and it seems like it was able for you to build this wedge with Airbyte as people were starting to build data pipelines for themselves. Taking a quick step back, we've had Sriram Ramaswamy, the CEO of Snowflake. We've had Kashish Gupta, the founder and CEO of Hightouch. So we've talked to a lot of people in the ETL space and the reverse ETL space. But one of the unique qualities about Airbyte is that it's ELT, extract, load, and transform. Tell us a little bit more about why you that important architectural shift for a nomenclature that's so profound? Was it something to do with what you heard from customers, or was it like a philosophical direction of where you thought data was going?

We can take any letters, put them in any orders. At the end of the day, what matters is like what value we provide to people, whether it's ETL, ELT. For me, it's just about what does the manufacturing of data look like? And in a way, it's a way of taking advantage of the fact that technology It's cheaper, storage is cheaper, compute is cheaper. And so you can afford to have all the data sitting around. It gives you more agility when you want to process it, when you want to make mistakes with it, when you want to analyze it. So for me, at the end of the day, data needs to move from point A to point B. That's all. EL, ETL, reverse ETL, it doesn't matter.

How is this important architectural decision informed subsequent decisions with building the company and scaling?

I think it makes the whole thinking about data a lot easier because there is a ton of complexity when you're trying to break down data silos. You are trying to extract data from a system that you have no control over. And so anything can fail. It can be like Salesforce rate limiting you, it can be a database going down, anything can go wrong. And so when you go for this first approach of just, let's just load the data, let's make this data a mirror of what exists at the source level, you remove that complexity, your system is more reliable. And you're building on strong foundations. And even for our software, for the product, we're not just limiting ourselves to extract and load work. We're doing the activation piece, what you call reverse ETL. We're doing, we're going more into the realm of transformation, but we're building on very strong foundation, the same way people that have been adopting Airbyte have done.

Awesome. Today you're at over 600+ connectors. What was the first connector?

Postgres, pulling data from one of the most known database. And to be clear, it was terrible.

You're at 600 today. Tell us about the 1 to 50 and then 600. What was the harder part?

When you go from 1 to 50, you can afford to focus very carefully on every single one of them. When you go from 50 to 600, and actually we have more than 12,000 that have been built with low-code and no-code, but—

Oh, wow.

You have to create and to make it an actual process. That to me was the big change of going from 1 to 50, from 50 to 600.

Yes.

It's an industry at that point.

The classic do things that don't scale, but then once you have to hit scale. I wanna get these numbers right. Airbyte has raised over $100 million at a $1.5+ billion valuation. With lead investors like Ultimeter Capital. I saw our Ultimeter investor sneak in the back there. Benchmark, Thrive, Accel, and more incredible names. Congratulations.

Thanks.

Who was the first person to say yes?

Among the names? Because fundraising starts before the big names.

That's so true. No, who was the first person ever?

The first person ever, I would say, was the former CEO and founder of LiveRamp, Oren Hoffman, who, he loves data.

Yeah. Naturally.

Showed him an idea, at the time it was just an idea, and said, yeah, let's go for it. So that was the first.

And tell us a little bit more about those traction metrics that you felt really caught the hearts and eyes of these investors that helped convince them it was a billion-dollar-plus market.

The product was very immature when we released it, but people went above and beyond to make it work. And when you have this type of behavior, you know that you're actually solving a real problem. People are willing to go through the pain of integrating a product that is not yet mature at the time to make their life easier. So in 2021, when we released the product, the community adoption was really fast and we started to get like the first support contract, et cetera, et cetera. And to me, it's more like, like the period of time and the acceleration that happened.

There are a ton of early-stage founders in the room, folks that are just thinking about gearing up for their first fundraise. What's a piece of advice or hot take you might have for them?

When you start, it's just about survival. We always say that only 1% of startups make it, so You need to do really whatever is necessary to survive, like optimize for survival rather than the big vision. Like big vision is important, but what matters, like what you need to do today to actually execute on that vision. We raised in COVID, we had to do some deals that were not the best, but they allowed us to survive. And then we course correct in the future. So that to me is very important.

Outside of fundraising, which always has its twists and turns, we know that a journey of a founder is not always a straight line. Tell us a little bit more about something that might have gone wrong and how you adapted to that?

I cannot count them, but at least two main events that were very, very painful is first one, we started in January 2020, boom, March 2020, everything falls through the window. It was actually a good forcing function for us to just redefine what product we wanted to build, because in a way, very extreme, we don't want to build a product that cannot survive a global pandemic. So very extreme, but—

Good bar to have, yeah.

The second one was the war in Ukraine because we had a big team in Ukraine, helping us on like building more connectors. And from one day to the next, no people anymore. They were obviously very distracted. So they seem small now, these problems, but when you live through them, it's, it's painful.

Well, let's talk a little bit about the future and scaling Airbyte in the age of AI. A few years ago, everyone was talking about the modern data stack, whether it was Snowflake, Fivetran, Hightouch, so on and so forth. Now it's all about AI data infrastructure and AI ready pipelines with nearly a million jobs running daily. You've really had a front row seat to this transition. Tell us, what is AI data infrastructure? Is it just rebranding?

I think it's more about the maturity of the market. When you're thinking about data teams in general, you already have a very strong product that exists. When you're talking about AI product, et cetera, et cetera, data teams are involved, but it's mostly software engineers actually building this thing. So for me, it's more like a different type of audience that you have to talk to, but but they don't always know like all the practices of what does it mean to manage data at scale. So for me, it's not really a rebranding, but it's also how do you appeal to a different audience that is facing the same problem? And how do you make sure that we don't reinvent the wheel for a problem that is quote unquote solved?

As we look at the ecosystem and how it's evolved, we have our big providers, Azure, AWS, Google, and then we have a lot of the foundational model companies, shout out OpenAI, who has some folks here, Anthropic, so on and so forth. But the ecosystem has really changed. Tell us a little bit more about how you've navigated those relationships and where Airbyte really fits in.

Talk about OpenAI, you talk about Anthropic, you talk about all the more public models, they're all trained on public data, so they're amazing there. But the moment you need to make these models actually work with your own internal data, that doesn't work. Or you need to train your own model, but no one wants to do that. And at that point it becomes, how do we work with this company to make sure that We don't just have intelligence on top of public data, but we can actually touch and inject like the goldmine of data that all these companies are sitting on. So for me, it's more about working with this company and making sure that we can offer these connectors as a way to enrich any kind of requests that come into this product and inject that very necessary internal data.

Yeah.

As the ecosystem has changed, I'm sure your competitor set has also changed as these giants are building their own integration tools. How does Airbyte think about navigating that evolution?

I think there is a rule of 80/20 here, which is you have 20 connectors that everyone needs and all these companies are gonna build them and it's okay. When things fall apart is when you have either your own internal product or when you have a more localized version of Salesforce, like Salesforce of India, or when you have to touch data that you don't want these big companies to have access to. That's when things start to fall apart because you still need a way to bring that data in. And that's where Airbytes comes into play. Whenever we're thinking about data, and that also comes from my background, is the amount of scrutiny that companies put on data now that they've understood the value of it is huge. So we need to give them a guarantee that they can control where the data is going, how it's being used, etc., etc. So to me, that's where our fundamentals is around sovereignty. And that's what we've been pushing, and I would say we're probably one of the only ones really pushing that.

Absolutely.

That's incredible. Part of what makes Airbyte such a success is obviously your customer base, and you have plenty of Fortune 500 companies and have a front seat to how they're scaling, growing, and really how they're thinking about innovation. A lot of these companies are thinking about how to become AI-first or transition into this world of AI. What do you think sets apart AI-native companies and some of these bigger giants that are trying to cross the chasm there?

That's very interesting. I was actually just talking about that with the team on Wednesday. We have this big internal initiative, like we were born before all the LLMs become so powerful. And so we're also, over the past 2 years, we had to completely adapt. And this initiative, we call it Buy AI, meaning how do you incorporate AI into your workflows? Basically the big difference when you were born in the AI age versus when you were born before is that you have habits and you need to break these habits. That's the real difference. It's like when you build in the LLM world, all the habits you have are involving LLM somewhere. When you haven't, then you have to change these habits. And what I see is that a lot of people are very good at changing habits on the things that they don't like doing. I hate to write long-form documents. Generally, I'm more like a bullet point person. I just— ChatGPT and Claude are amazing for that. The place where I see a lot of struggle is getting people to change their habits in the things that they love doing because they don't feel the pain. And some people might say, "Oh, I love to write this piece of code." Maybe you should also learn how to change these habits here and get your satisfaction in a different way. And that's to me is the big difference between these companies. They've built on different habits.

We won't hold you to these predictions, but we're curious when everything is solved, when data infrastructure and integrations are just done and dusted, we don't even think about it. What does that mean for us?

I would say you don't think about the fact that you have roads, but the roads still need to exist. They still need to be built. They still need to be maintained. And I think whenever you're talking about infrastructure, you can never call it done because it is a living organism. It has to evolve over time. It has to innovate in places. And there are new places where data can live. You need to build these new roads to this data. So In a way, it's like a constantly evolving organism.

We have a lot of potholes on our roads, so that makes a lot of sense. We need to not have them. Okay, so inside the room today is predominantly under-35 founders or future founders. What would be your like general Michel advice for them? And if anything you've observed in the last 5 years of being a founder yourself, you know, how that's shifted, what tidbits you would have and wisdom you can impart?

The first thing is don't do it halfway. Talking about like always wanting to be an entrepreneur, I've always had side projects, trying to build side products. You burn yourself out when you have to have your day job and your night job. It hurts you when you need to talk to customers. So yes, you might not have a safety net, but I think if you want to be a founder, you also, you probably are pretty skilled already. So if it fails, it's okay. You will find a new place, a new home. But for me, it's really, if you do it, you have to do it 100%. Now, in terms of things I've learned is that every decision when you need to make them in the moment, they seem huge, but at the end of the day, when you look back, you say, oh yeah, that was such an easy decision. And so don't be afraid to make hard decisions when you need to, because if you believe that this is the right thing to do, that will just help your business and that will allow you to build on strong fundamentals.

Wonderful. That was great advice. We have a closing question on our end, and we like to ask every single one of our guests, and it is someone that is important in your life that has been a woman that's had an impact on you and your career.

Honestly, the CEO of Epic, not the game company, like the medical company, Jodie Faulkner. It's an engineer looking at a problem, going into their home office or garage. I don't know what it was at the time. And just building a solution that just has completely transformed how the information is moving in the healthcare world. And in a way, yes, I'm not in the healthcare industry, but I relate a lot to the, like, how she's built the system and the problem that she was solving, which is there is silos everywhere. If this data is siloed, the experience is not good for the patient, not good for the doctor, not good for the hospital. And she just said, okay, data just needs to move from point A to point B. You put a nice little UI on top of it so that people can actually interact with that data and boom, you've solved a massive problem. And she's what, 80-something and she's still the CEO at the company.

Very cool. Thank you so much for joining us at The Room Podcast. If you want more from The Room every week, subscribe to our newsletter at theroompodcast.com/newsletter. We'll be back next week. Next week with a new episode and inspirational guest. Tuesday, 10:00 AM Eastern, 7:00 AM Pacific. See you in the room. Perkins Coie supports the most innovative entrepreneurs and investors in fast-moving and high-growth sectors, addressing their myriad of legal needs. But the firm doesn't just provide end-to-end legal and business counseling to its startup clients. It also facilitates introductions to key advisors and sources of capital. Perkins Coie's interactive website, Startup Percolator, offers access to programs, resources, and rich dynamic content designed to assist entrepreneurs on their startup journey. To learn more, go to startuppercolator.com and perkinscoie.com.

This podcast is brought to you by Mercury, the banking platform businesses like The Room Podcast use to simplify their finances. It's time banking did more than hold your money. Now it can. With Mercury, you can pay bills in seconds, close the books faster, and even send invoices. Not only does Mercury do away with a patchwork of tools, it eliminates guesswork, giving you complete and accurate visibility into your business's finances, all from one account. Apply in minutes at mercury.com.