ElevenLabs CEO/联合创始人 Mati Staniszewski：欧洲增长最快 AI 初创公司的未公开故事

So we crossed $200 million now.

It was 20 months to $100 million and then 10 months to $200 million.

Today, we have one of the fastest-growing AI companies in the world, ElevenLabs, and I'm so thrilled to welcome their co-founder, Mati, to the hot seat today.

Pre-seed was tough. Pre-seed was hard. I think we spoke with a good number of investors—between 30 and 50.

How much did you raise in the pre-seed?

We raised $2 million.

Do you remember the price?

$9 million.

For us, even from an investor and product perspective, especially now, the speed of execution and the speed of investing are the only things you have. Our biggest contract is around $2 million, and they're mostly in the call-center, customer-support, and personal-assistance space.

What do you believe that most people around you disbelieve?

That you can build a company from Europe at a global scale.

How do you answer the question, “Why won't OpenAI just do this?”

They definitely will, but I think they lack the ready-to-go...

Mati, dude, I cannot wait for this. I've wanted to do this one for a while, so thank you so much for joining me today.

Harry, thanks for having me. After working on a number of projects together, I'm so happy we can finally speak.

Dude, I am so thrilled we could do this. I spoke to so many people beforehand, and I have to say Luke was phenomenal on your team in prepping me so much. But I do want to start a little bit with the origins. You grew up in Poland. How did growing up in Poland impact your mindset toward the world and company-building?

You know, in Poland, it's a very different, much smaller world. Growing up, I was in a very lucky position: I was in the suburbs of Warsaw, then went to high school in Warsaw, and started seeing more of that world. That was, of course, incredible, because it opens your eyes to what's possible—to the fact that there's so much world beyond what you've seen in smaller cities.

I think, in a similar way, as you think about building ElevenLabs now, the scale of what we don't know is still ahead of us. This is what's exciting: we know that if we climb those additional hills, those additional mountains, we'll see increasingly more.

The second thing was that I saw the transition from being in the suburbs of Warsaw, in a public school—a school where you would have any kids who lived in the area—to being in high school, where I met my co-founder. There, you had a slightly different crowd: people who would win competitions and need to go through a few steps together. Suddenly, that density of talent was the most motivating factor to explore more, learn more, and do more.

Now, we're trying to replicate that at ElevenLabs. At any other company, I'm sure, too, you try to keep that density as high as possible because, at the end of the day, that's what's most motivating: being part of this incredible set of people.

I was most inspired—or given hunger—because my dad always said, when we went into Chelsea, “Oh, it's a better life here, and this is where people who've won live.” I didn't live there, and I wanted to fucking live there, and it fed me with this insane hunger. Where did your insane hunger come from in those early days?

The early days? Definitely, it was the combination of family. My older brother trailblazed going abroad to study, then motivated us: “You need to do the same. This is really helpful.”

The second thing was really this community of people in high school. Meeting my co-founder, who was extremely smart, meeting some of our close friends, and going through this cycle of motivating each other: Yes, we want to study at the best university; yes, we want to go deeper and crush this exam. I think this was a huge element for all of us—just going after it and knowing that maybe it's possible.

At the time, everything felt distant, and many of the things still do. But without that motivation through the people, it would not be possible.

Speaking of the motivation through the people, you have a very special relationship with your co-founder, Piotr.

Yes.

So I have to ask: you come together. How does the idea for ElevenLabs come to be? Was it truly inspired by bad movie dubbing?

There were 2 things that coincided for us in starting ElevenLabs. One was that, through the years prior, he worked at Google and I worked at Palantir. We would meet for hack-weekend projects together and try to explore a new technology.

One was in the recommendation space. During the crypto hype, we built a crypto risk analyzer, which wasn't very easy and didn't work very well. Then we did a project in audio. The idea was: can it analyze how you speak and give you tips on how to improve your speaking? That opened our eyes to what was possible in the technology space there.

Fast-forward later that year: we did the hack-weekend project in early 2021. In late 2021, the moment hit for me from Poland: wow, all movies are still dubbed. You have all the voices from the original, whether male or female, narrated by a single character. So you have one voice narrating all the characters in a flat, unemotional way. It's like an audiobook reading of a movie.

It's, as you can imagine, a terrible experience and something that we knew would change a few years from then. Combining the experience from working on audio with knowing that this was a problem, you could imagine that, with that kind of advancement, all the voices would have the original emotions and intonation and sound incredible. That kind of kick-started this idea, and, of course, then it expanded to, “Okay, we need to fix the research layer to actually make it happen.”

Then we started with a lot of our creative-platform work around ElevenLabs and expanded to agentic-platform work, where the interaction is now shifting and voice is this big interface for the technology around us. While it was very dubbing-specific originally, it expanded to everything voice today.

I mean, it must be the most boring movie experience, having one single voice for everything.

It's terrible. It's terrible.

So what do you do when you land on the idea, in terms of your next steps? You move into the research layer and determine whether it's possible to actually do this.

The way we approached this was to first try to do 2 things in parallel. My co-founder said, “Okay, can we use existing technology, stitch it together, and create a dub of a movie in a better way?” It quickly transpired that you get a good effect, but it wasn't brilliant. So, to actually fix it, we had to take a step back, fix one of those components, and make it great.

At the same time, when he was doing research, my mission was to figure out whether anybody actually wanted that dubbing product. I was emailing all the YouTubers, getting the emails, scraping them, and trying to message them: “Hey, if we had a dubbing product to make your movies available in all languages, would you be interested in that?”

There was roughly a 15% reply rate initially from the first batches that we sent. All of them were personalized. We sent thousands of them, but the interest was lackluster. All of them were like, “I don't fully believe this is possible. Can you send us a sample? It would be great, but how will I operationalize this? YouTube doesn't support it.”

So there was somewhat of an excitement, but not a burning problem that they needed to solve.

But then the second thing happened. We started sending the samples, started speaking more with the YouTubers, and it quickly became clear that what they actually wanted help with was much simpler. They wanted to post-produce and correct things if somebody said something incorrectly. They wanted to understand how the script would sound before they needed to produce it. They might not want to speak at all, and they wanted to voice-over the movie.

So, it was a super-simple problem that didn't include anything with language engineering. And, of course, Michael Vander[?], as he dived into the research, realized—and he's an incredibly smart researcher—that you can actually build a completely new text-to-speech model that would be a lot more emotional in nature and make that narration a lot better.

And that's only possible by creating your own models?

Yeah. The quick answer is yes. All the models at the time—you could tell immediately—were like the uncanny valley. They weren't very good. You couldn't replicate voices.

Dude, how should I know as an investor? Again, I use this show just to get better as an investor. How should I know as an investor whether a problem needs its own model or whether you can just leverage alternative existing models?

Wow. I think the answer will definitely change now. 2022 was the year when nobody was yet thinking about AI. This was the downfall of the metaverse and crypto days, so the attention to the models wasn't yet in the public eye.

You know, ChatGPT happened at the beginning of 2023. That's where it all started spiking. So, at the time, you didn't really have a choice. You knew, as a user and as an investor, that what existed out in the market just wasn't very good.

And then there’s more of a question: will this team be able to solve and create something better?

If you were creating ElevenLabs today, what would you do architecturally differently, given the architecture that we have today?

I think the current question is still to what extent you continue in that kind of single-modality space, where at one time you trained a dedicated model for speech, or a dedicated model for image, video, et cetera, versus what is now more of a theme, where you train a more multimodal approach, combining reasoning and speech together to create an even better speech experience.

Our most recent generation is effectively Eleven v3. If we had that, then I think we would have probably skipped a few steps, and our experience would have been even better.

So, off tangent, but I’m just concerned about, bluntly, the plateauing progression of models, and GPT-5 was the embodiment of that. You move from incredible feature descriptions to a discussion on cost, and when you have a discussion on cost and efficiency, it’s like, “Oh, we’ve hit that.” Right, do you agree, and are we in an element of incrementalism and plateauing in voice?

There are use cases like narration. In narration, we think it’s plateauing. The new model generations will not make narration of an audio look drastically different. It will still be in a similar quality, but I think in general, the point is right: if you just do research, eventually it will commoditize, and eventually that advantage you can deliver from research isn’t enough.

So you really need to build a product, and as you think about ElevenLabs, we combine both together.

Do you think we’re at a stage where scaling laws no longer equal an equivalent level of progression?

In a biased way, I feel like the progression is still just scratching the surface. The moment of seeing that S-curve of AI getting adopted everywhere is just getting started, while scaling does continue in the same way.

There’s a difference between adoption and development. So I agree with you there in terms of adoption, but actually, the development progression is different.

Got it. I think it’s probably a different answer in the LLM space and a different answer in the voice space. In the voice space, you are still seeing a pretty quick curve. There’s an interesting variation of, as they all combine, what does that mean for the world? It’s a slightly deeper understanding of everything around you. But on the LLMs, I agree it’s probably flattening to some extent.

I spoke to Andrew Reed before the show. I had a great chat with him last night, and he said, “Ask him about why OpenAI can’t do it.” I remember chatting to Kieran about this—Kieran, obviously, being my partner—and there were 2 reasons why we didn’t work together.

One is that small check, which you very kindly offered. The other was—no offense—it was super freaking early, and it was like, “Why wouldn’t OpenAI do this?” No offense, 2 very young guys in London are doing this. How is this not going to be done by them? How do you answer the question, “Well, duh, why won’t OpenAI just do this?”

They definitely will do something, but I think they lack the genius that my co-founder and team have.

The true, longer answer is, one, focus. In the early days, especially when we spoke then, there were so many different things you could do in the AI space, and I think we took this bet that we wanted to really own and win in the voice AI research and product space. All our work is directly tied to voice.

The second thing is, I actually think the number of researchers in the world working on voice and being exceptional is super small—probably 50 to 100 people at this top level. Piotr and the people around him were able to assemble one of the best teams in the space. I think we have 5 to 10 people who are in the top 100. It’s a mighty team, and I’m time and time again surprised by how incredible they are.

To make it specific, text-to-speech is one area that has blown everything out of the water. Speech-to-text is beating OpenAI and Gemini on benchmarks. Now music, which is again something that no big company has yet been able to crack or do. So it’s an incredibly mighty team.

Even if OpenAI does actually do something on research—and I think they are trying—then that product layer is where you really need that big advantage. If you’re in a creative space, if you do narration, if you do voice-over, if you’re a dev, you go through so many additional steps to really make it perfect. We do that well.

If you’re building a voice agent or conversational agent, you need to bring knowledge-base integrations and functions. You need to then deploy, test, evaluate, and monitor. All of these pieces are coming together in a platform, and I think OpenAI is not investing as much time. They could, but they aren’t.

So it’s a combination of a mighty research team, speed of execution, and actually focusing on the use case in a deeper way.

You mentioned that kind of mighty team. Those people are very valuable, and OpenAI, Anthropic, or Meta would pay a lot of money, it would seem, for people like that. Do you worry about the war for talent, and how do you think about retaining talent when there are hundreds of millions of dollars going across the table for them?

The talent, especially on the research side, and the impact that those people can create, is off the scale across any of those companies, especially in the early days.

Of course, to some extent, as I think about Meta and others, they are paying for the know-how as much as they are paying for the raw talent too. Getting those early people gives you some insight into the models and architecture that you can then bring across and use to accelerate. So in the early days, it’s more valuable than later.

But I think in our case, the reason it’s still valuable is, one, the upside for ElevenLabs is just getting started. So I hope we’ll be able to compete with any of those companies in the future at the same scale.

Two, what you do have, which I think is important as I think about any of those companies, is how close you are from developing research to actually deploying research. At ElevenLabs, if you are actually creating new models, they get to production. Some of the most important things that our products work on get there almost immediately. So the gap is super quick.

In bigger companies, you go through the usual corporate red tape to get any of that work done.

Three, I think we now have a very small and mighty team that can learn from each other and move pretty quickly. I don’t think there’s a guarantee you get that in some of the other companies. I think they are optimizing for a lot of people, but not necessarily the right people.

We’re going to go to team size later. I do want to stick with some chronology. So many questions that I want to ask you, but let’s stick with chronology. We mentioned the V1 and building to beta, and I think it was in January 2023 that you raised the pre-seed. How did the fundraising process go? You’re 2 young guys saying, “Hey, we’re going to do this.” How did the pre-seed fundraise go?

Pre-seed was tough. Pre-seed was hard. It was a similar question to what you asked: how will we fix research?

The second question, which was very interesting, was, “We think the market is very small for what you are solving,” which at the time was a combination of AI and voice—nobody was thinking about that. So that was surprising to us, and we strongly disagreed with it.

The third was defensibility: would it actually be better than what incumbents from the big companies would solve, or how would you outcompete in the long term? So those were the 3 questions.

We spoke with a good number of investors—

Double digits of investors, and they said no.

Yeah. It was between 30 and 50, and it was doubly hard at the time because, in early 2022, we got an offer from one of the accelerators in the US—not YC—and we were thinking, “Should we take it or not?”

We decided no. We rejected it because we thought we could create something more valuable, and we didn’t need that help at the time. So we decided to go independent, and this triggered a more stressful time, where suddenly you need to raise money.

We started spending a little bit more on GPUs. We hired the first few people. All of that was coming from our savings from Google and Palantir, which we were lucky to have. But I was like, “Okay, now it’s actually getting a little bit more risky, and we want to double down. We want to invest even more, so we need money.”

How much did you raise in the pre-seed?

$2 million.

$2 million. Do you remember the price?

$9 million.

$9 million post-money.

9 million post. The amount was exactly 11% of the equity that the first investor was buying, and then the other ones were layered in.

So it's like just over a million. And this was 2023?

That was 2022.

No, but dude, I'm so used to stories like this on the show where it's like HubSpot and you're like, “When was it?” “Oh, 2006, 2008, or 2009,” or whatever. But this was recently.

Oh, yeah. It was only 4 years ago. In the end, we had Credo Ventures from CZ and Concept Ventures here in the UK. We also had one of our friends from Oxford, Peter Czaban, who invested; he was a co-founder of Polkadot. So we had a good set of early believers, but to get there, it took us a bit longer.

So you raised the 2 million. What happens then? You start building out the team and investing in GPUs?

Exactly that. The main reason we raised was to accelerate what we wanted to do. The first thing was building a small data center, so we built one. Back then, I think in Poland, we invested in a few GPUs, and then quickly after, we moved and started buying some in the US. On the team, we started hiring—not many, but it felt like a lot at the time. It was an additional 2 people, I think.

Don't go too far there. Mati, you did the pre-seed raise and the beta launch at the same time. Did you have product-market fit on the beta launch? Were there early signs that it was working?

Okay, so we got the good timeline piece there. We actually raised the round in mid—like Q3 of 2022—but announced it in Q1 of 2023. The reason we do that, and we do that all the time, is that we don't want to announce a round just for the sake of announcing a round.

Our philosophy was always that the round should have another purpose, which is to bring the product out and help you get the product into users' hands. Celebrate a set of customers to show that you've arrived in a specific sector. Bring a new research model into play. Every round that we did would always tie into a product announcement.

We held on to the announcement for a few months until we had our beta release, and then we triggered it in January 2023. But to your question, initially, when we worked on dubbing in those first early days and were emailing people, we didn't have product-market fit. It was very clear that people were slow to reply. We sent samples, and they weren't engaging. So we didn't have product-market fit through the 2022 period.

Then, when we shifted from dubbing into narration voice-overs, the product-market signal started to hit. I remember 3 things that happened. First, we did a blog post around the first AI that could laugh, and we sent samples. That got picked up by newsletters, and people were like, “Wow, this is incredible.” The next day, we had 1,000 people on our waiting list.

Then the second thing happened. We invited the first 100 or so users to ElevenLabs to test it out. We had this audiobook author who joined the platform. Our platform was this small text box where you could type in tweet-length text and narrate it, and he would copy-paste his entire book 500 times, download it, and stitch it together.

Then he released it on the platform. At the time, AI was banned; it passed through as human content, and then it started getting reviews that it was great. He came back saying, “I want my other book-author friends to do it too.” So I was like, “Okay, we're onto something.”

Then we launched publicly in January 2023. We knew we were starting to get more of that signal that creators and narrators loved the work. To be honest, I think from that moment onwards, we saw clear momentum. Then there were more events after the media picked it up, and then more creators picked it up. But at the moment of release, we knew that it was valuable.

Maybe the last thing on that is the product-market-fit concept. For us, we never fully knew what this actually meant. We knew users just loved our product at the time, but I wouldn't have called it product-market fit then, because we were thinking, “How do we make sure it's providing value for the next 5 to 10 years?”

Then we'll decide that this is something where we can say, “Okay, this is clear product-market fit.” We think this is self-sustaining for as long as we go into the future, so we wouldn't go by that definition. I think we are now closer to that, but we still know we can create so much more value.

There are so many things I want to unpack there. You said something about the timing of your announcements and aligning them with actual material news items. Do you have any big lessons on announcements—how to do launches—that you've found particularly work well and that other founders should know? So that's the first one.

100%. For us, the lesson is: make an announcement close to something big you want to announce—from the product, from the user, or from a hiring perspective—and tie it with that. I don't think celebrating the number itself is the right way. In a way, it's like you're giving away the company, so it's almost like, what are you doing it for? You want to show as much of that as possible.

Second, I think this is very easy to optimize for, and maybe I'll cover it up front. It depends, of course, on whether you are self-serve, where creators and developers are the users, or whether you're optimizing for enterprise and sales. You really want to focus on how you actually get the users rather than just the media PR element. I don't think the latter is actually as valuable as it seems to a first-time founder.

For us, I remember when we did the beta, we were speaking with some of the bigger publications. One rejected us, the other one accepted us, and we felt like it was a big deal. Then we did all the prep for the interview, did the actual news, and it was posted. We held everything around this post until the deadline they told us, and it had no impact.

I think probably a few more people from the investor scene read about us, but we didn't really care about it at the time. We were all about users. What actually worked was working with the newsletters that were talking about AI, working with our friends from the YouTube community, who then posted on Discord that we were opening this up.

Our Discord community was one of the most valuable. Reddit—those people picked it up quicker than anyone else. Hacker News, posting there—like, all of those were immensely more valuable. Since then, we always spend a lot more time on the actual forums where our users are rather than where they are not. I think we grossly overestimate the importance of traditional press.

Oh, 100%. I remember when I was first in TechCrunch, I was like, “I'm gonna go viral. Everything's gonna work.” And then it happens, and you're like, “You get, like, 3 followers.” This was when it was much bigger. So I completely agree with you there: the grassroots.

The other element tied to that is how do you think about fundraising? This is advice to founders fundraising pre- or post-launch.

What frequently happens if the launch goes well is that you'll get more interest. You'll get it from investors, from events, and from media, to a large extent. This is the time you should really focus on getting the product out, and all of that is a distraction to a large extent.

The first time we launched, that kind of happened. We started getting so much of that, and I'm sure I made the mistake then of picking up a lot of those opportunities and doing more events than we should have. The best course of action would have been to just spend even more time with users. Then more enterprises started getting interested.

On the investor side, I think you want to keep investors lined up. You do want to tell them, “Okay, I'm not raising now. I'm going to reconsider whether we need more capital in Q3 or Q2,” whatever the time period is. You want to line them up, and then when you actually need the money, that's the best time to reengage with them.

But I think it's a waste of time to be in this continuous fundraising mode. It's distracting. You need to have conversations. It's not useful.

Do you choose 3 that you'd like to have and engage with in between cycles, or do you not even do that? Do you just say, “Hey, I'm heads down on building. I'll come to you when I'm ready”?

Today, less so. At the time, no. Good question. We do have a few investors that we think would be valuable that we might not have today in the space. I mean, I'm very happy we are working together.

But no, that's the truth. Over time, we would have a few investors that we wouldn't proactively reach out to, but if they reached out through any of the conversations, we would try to engage. The engagement there wouldn't be, “Hey, we might be raising.” It would be more like, “Hey, can you help me with an introduction to X?” or, “Hey, can you help me with this hiring problem?”

This has worked tremendously well. I find that investors are genuinely keen to help if you find the right person, have a very clear problem, and don't try to overuse their time. You come with 1 or 2 things that are concrete and useful, and they know as well that if you're growing as a company, it's good for them to show that they helped too.

So we've done that.

It's also a good litmus test to see if they genuinely are interested in being there or if it's just platitudes of niceness.

Oh yeah, 100%.

Totally agree with you.

Before, it's kind of the second thing where you do get the interest from investors. I think this is the best time to actually test whether they can be helpful. Before you accept any term sheet or any money, it's like, “Now, can you help me with the angels that you want on the cap table? Do you want the introduction?” That's the best time.

Any advice on angel selection for founders in the early days?

The way we approach this is that we had the standard venture capital considerations, of course, whether it was the help of their network, the brand, or making sure we arrived in a specific region. But with angels, we would usually optimize for whether they had domain expertise that we didn't have. That's one category.

Second, could they help us validate ourselves in specific circles that we might not have access to? If you are an AI founder, it may be valuable to have a set of AI founders in there, so you are part of the same events and the same conversations. Then, of course, in our case, it was: could we have some of the go-to-market expertise where we didn't have it? We had a lot of self-serve and some volunteer experience, but with sales-led, how could we bring that in as well?

You mentioned starting with the API, and suddenly it was going well. You were seeing this great groundswell and great adoption, and this was from January to June 2023, I believe. Then, in June 2023, you raised $19 million from Brian Kim and Andreessen, and then Nat and Daniel from—I can't remember what it's called—NFDG or something.

Yeah, NFDG. It's Nat Friedman and Daniel Gross.

The NFDG is now—I feel like it's a good name, but I was pushing to change it to a different name for a while. Maybe now it's not possible anymore. The interest started around March 2023. We had a lot of investors who had been in close conversation with us in the past reapproach us.

We would effectively be a little bit more in waiting mode: let's see how we get through, and let's optimize for the partners that we truly want. What we truly wanted was a combination of making sure that people globally in San Francisco trusted and knew that we had arrived, that we were a company that could be trusted and was building something ambitious. Then, of course, we wanted people we admired who had created something special. Brian was definitely in that latter category.

He approached us. We spoke with all the classic tier-one funds, and a16z was one of the most—

All the funds in London?

We spoke with a few funds in London, yes.

And then Brian flew to London?

Brian flew to London. a16z was phenomenal. They did exactly what we spoke about: they showed us that they cared before they invested. They introduced us to incredible people, including some celebrities, to work on the voice licensing. They were on it for 2 weeks prior, already helping out in any way they could.

Then Bryan flew to London, met with us, and we signed a preliminary term sheet. It was interesting. He flew in, and I had a phone call with my 2 co-founders: “Hey, what about these terms? Can we accept these?” Piotr would give guidance on where the hard line was and what we could do. Then, on speakerphone, we signed together and finalized.

Do you think speed really is a differentiator for an investor in winning?

Oh yeah. Speed, for us, even from an investor and product perspective, is something you need. Especially now, the speed of execution and the speed of investing are the only things you have.

This is what pisses me off. I'm happy in that world. I can move supremely fast and invest a lot of money. But what I find more and more with founders is that they want to run a roadshow and do a 10-day process, meeting investors, and then have another week to decide which term sheets they want to take. I'm willing to give you a term sheet tonight—save 17 days. But it's more and more common that founders actually want to run roadshows. How do you think about that?

I'm now in a lot of great companies that get built on ElevenLabs, so I get to be in these conversations more frequently when they raise money from our investors. Recently, we had a company raise a Series A that had built a healthcare voice bot, which would effectively help patients calm themselves through conversations in an incredible way. Then there's another company building for healthcare customer support. So there are quite a few, and now I'm in the conversations about them raising money, which is an interesting one.

Usually, it's clear that some people don't fully know what they are actually optimizing for, and it's good to talk through whether they are optimizing for the valuation, the dilution, the right brand name to help them out, the network that this brand name has, the specific partner, or something different. I think it's pretty distributed. When I speak with founders, or our investors ask me to speak with founders, it's usually to ask: what is it that you actually care about for the thing? What do you miss?

Then it transpires pretty quickly whether the roadshow is there just to bump it up because they think the term sheet they got is not valuable, or whether they actually want someone else but want to keep the optionality of having the first one while running the process. But, you know, all the traditional, simple things.

I'm a romantic. I hate this idea of a roadshow where it's like, “I'm going to meet everyone, then I'm going to compare term sheets, and then I'm going to decide.” It's like, what happened to the partnership?

The truth is, the other way is also true: a lot of investors want to give the term sheet, but that is before any show of partnership. So how do you know that this is a good partner? How do you know whether they are giving you good terms?

Especially if you're a first-time founder, you've never run this process and you don't know how much you are valued. You need to have some relative comparison. This can happen through conversations with other companies and other founders, comparing yourself to other companies in the field, or running some form of a process.

“Roadshow” probably sounds pretty bad because it sounds like a lot. The way we approached fundraising was always, especially in the early days, to queue up a few investors we really cared about and first speak with maybe 1 or 2 who weren't a priority, so we understood whether we were perceived in the way we wanted and whether the messaging we wanted to put across made sense. If it did, we moved straight to the top ones.

How do you feel about exploding term sheets? I don't like them, but we got a few. We had a few come through.

Do you understand them?

I do. I think I do. At the time, I didn't, but I do now. I still don't like them. I won't name the people, but one example was a smaller fund, and they felt that if they got us the term sheet, we would use it to bump up other term sheets. It happens very often.

And that does happen.

So I got it. They came from the approach that if it goes too high, they won’t be able to invest, and they don’t want to be part of that. So that happened, and of course, being an investor, being the first term sheet is the worst, because you’re used as the stalking horse. They then go, “I’ve got a term sheet. I’ve got a term sheet from a great fund.” They never named the fund, and you get no credit for being the first. Then you’re just trampled on by everyone else who comes in.

That’s what I don’t like: founders trying to do that too much, trying to get the first term sheet to bump the other ones. I think it’s the wrong approach. If it’s an order of magnitude off, it’s probably wrong, or if it’s 5x, then yes. But if it’s plus or minus 20%, that’s the wrong optimization. There’s so much more value that investor partners can give than the money itself, and I think that’s what should be optimized for.

You’ve got Andrew introducing you to celebrities, and you’ve got Nat and Daniel, two of the brightest minds in AI, who are querying APIs that no one else does. Are the Americans just playing a different game from the European VCs?

We love our partners. I mean, our team is so incredible, and the network is incredible. Now a16z with Andrew is just another level, too. The ICONIQ team also provides incredible help. NEA is just bringing—

All U.S.

They’re all U.S. So I think they’re playing a different game. From our experience so far, they’re a lot more keen to take the risk. Even our conversations are always like, “How do we bet bigger?” rather than, “How do we optimize for the negatives?”

The one check I always try to do with P for any of the partners that we bring, beyond the capital, network, and brand, is always: how will the person you partner with behave if it’s not going well? We usually try to get reference checks for how they behaved when the company was failing. All of the ones that we had were very positive. They were really on the side of the founder, helpful when it was not going well, even more so than when it was going well.

I was doing the same check with other companies, and maybe this is because in the U.S., the number of companies that have gone up and down is so high. But with some of the investors we spoke to in Europe, that check didn’t yield good results. I had one with Brian Kim that didn’t go well, and I have to say Bryan was the best. He was the lead investor in the round, and he got the money back for everyone. That’s rare, and he really put in the work to do it well and to do it quickly. I was really impressed with him in that way.

No, exactly. Brian—yeah, Brian was one of our checks that came up strongly. Jennifer, too. All of them. It’s just crazy how—

Okay, $19 million in the account. Most people suddenly go on a hiring spree—not a 2-person hiring spree, but a proper hiring spree. But you favor very small teams. You’ve said this to me before, but Luke said it to me. Why do you favor really small teams?

I think the first piece is that more people frequently doesn’t fix the problem. You don’t need that many people to do something special. That’s the first thing.

Second, by keeping teams small, as an organization today we are 250 people, but really it’s more like 20 teams going to market with 5 to 10 people. They’re just executing on specific projects where they have higher ownership. They can move extremely quickly, see the results, and iterate with reality a lot quicker to improve that. I think this just works in such a beautiful way. Of course, there are other challenges with this.

Are those 20 teams organized by function or by project?

It depends a little bit. In product, it’s by product area. We have a team working on our Studio interface that’s responsible for all the core experience when you log in. We have a team responsible for the entire voice-agent suite, and within that there’s a team working on some of the more enterprise components and another team working on some of the self-serve elements.

All of the teams are organized around the product area, and then in the other parts we try to shard it pretty quickly. We’ll have a separate team for talent, of course, and then a team for people. They have a high degree of independence in how they actually execute, which helps.

You mentioned 250 there. It’s small given the size of the company in many ways, but it’s also still 250 people. When was the company culture the worst, and what did you learn from that?

There were moments where I could feel this tension between go-to-market, research, and engineering. There’s a specific story. It’s a super lucky position where we develop our models and bring them to customers, which of course makes it very special. We frequently show the world, for the very first time, something that was not possible and is possible now.

In late 2023, we did text-to-speech. We then did voices, so you could recreate a voice. Then we created a basic way for speech-to-text internally. We had all the components that we originally talked about to create dubbing, but we hadn’t done dubbing publicly yet. We gave all the components to some of our customers so they could use text-to-speech, the voices, and speech-to-text.

There was a lot of pressure because it was valuable technology that we could give to our enterprise clients. We told one of the enterprise clients that we were planning to launch our dubbing solution, combining those components, later that month. I think it was September. They took the components and released dubbing 2 weeks before we did.

That was one of the low moments for me, my co-founder, and the entire team, because dubbing was our story. We told that to everyone. We knew for almost 2 years that this was something we wanted to solve first. We had all the components to be able to do it; we were just waiting to optimize it and make it perfect. Then suddenly this partner released it and got all the attention.

All the media, Twitter, and all the users were like, “Wow, this is the most incredible thing that has happened. How amazing that you can speak in another language and still sound the same.” You could feel the morale in the company was low. People were like, “We spoke about this. We knew for almost 2 years that this was something we wanted to solve. How did it happen that the customer had this and solved it first?”

Research and engineering weren’t happy: “How did they do it before with the components we gave them?” Go-to-market wasn’t happy: “How did we sell this to the customer? Now all the potential clients we could have had aren’t ours.” Of course, Piotr and I were like, “Hey, this was our idea. Why would you do any of the partnerships if all of that is given to another company?”

Who was the partner?

I don’t know if I can mention them. To their credit, I don’t think the timing—or the fact that we told them—was actually a factor. I think they were just trying to do something quickly, and we had a lot of calls with them in those days because we were thinking, “What do we do? Does the contract give us flexibility not to continue?”

So we talked with them about it: “Why did you launch?” I don’t think it was dictated by our timeline, but the timelines were very close to each other. They told us that they thought it was a good hack-weekend idea. They gave it to their intern, and the intern built the project. Then it exploded. It did give them tens of millions in revenue over that period of time, so it was significant.

What’s your biggest piece of advice to a founder who has a moment like that, where you just feel the air come out of the company? How do you inflate the company again after such a damaging blow?

I don’t think the first reaction should be, “Hey, everything is fine.” You should be authentic and tell the team what you’re feeling: “How has this happened?” The important thing is to go through what we’ve done wrong. That frustration was clear. I think it was valuable for us to talk through it: “We are angry at ourselves. This is wrong.”

Then, as we move from that stage—yes, we are angry—what are we actually going to do about this? You need to move to that second stage very quickly. I think I’ve made the mistake sometimes of just going to the second stage, as if it’s not a problem and we should just solve it. It’s actually super valuable to talk through what has happened.

But then you need to go into the second stage, and in the long term that will work out. The relentless execution will prove itself and show it to the world. The worst is if you repeat the mistake: then someone needs to feel the repercussions. But if you’ve learned from that and haven’t repeated the mistake, it’s fine.

Given the commoditization of a lot of technology today, do you feel speed of execution is the core differentiator between those that will win versus those that won’t? Or is it quality of research, access to GPUs? How do you think about that?

I think it’s both.

I think the way we approach this in the company is that we have research, product, and the ecosystem that we've built, which is a combination of distribution and brand. But research for us is a head start. We are investing in research, and we'll continue investing in research. We want to be the best across voice technologies. But all this gives us is an advantage over the competition for the next 1, 2, maybe 3 years.

How far ahead are you compared to the competition, do you think?

I think it depends on the use case, but 6 to 12 months, I would say, depending on the space that we're in.

Do you think that's a lot or not?

I think that's a lot. That research piece of 6 to 12 months is enough for us to then do the second thing well, which we do in parallel from the start: build a phenomenal product experience.

What percentage of your revenue do you think you spend on compute?

We've built our own data centers for training, and then for inference we, of course, use some of our great partners across the traditional cloud providers.

Why build your own data centers? Most people will just use CoreWeave or NVIDIA, or whatever that is. Why build your own data centers?

We did the math. In our case, if we assume we continue training the models the way we want to—continuously, and a lot of those models—then, second, for data transfers, and as we think about continuously bringing in more data, we will likely, on a 2-year horizon, break even by having our own and not renting.

Assuming, of course, that you make some improvement in the GPU infrastructure—but assuming that wasn't too high of an improvement—we would have been successful, and we were. I think the ROI made sense there, and now it pays dividends because we can do more experiments more quickly.

This can still be wrong at some point. There might be innovation that breaks that equation, but for the current time, it just makes more sense: more control.

A lot of investors are talking about the poor unit economics of a lot of companies that we see today, whether it's your Replits or your Lovables or any companies like this—well, any application AI companies, period, really. They generally have pretty tough unit economics, generally speaking. Do you think that's a fair criticism, or do you think it is simply shortsighted given the changes we will likely see in the cost structure?

Of course, the unit economics in most of those cases that you mentioned are pretty poor, but I think the strategy is that, first, the models will optimize for cost, and second, they will be the brands that customers trust, and then they can actually use a lot of the signal back.

In the ElevenLabs example, our unit economics are much healthier than most of those companies. We control our research, our product, and our distribution, but still, if we had a new model that we needed to release ourselves, we would optimize less for the cost structure so we could get that magic out more quickly than other competitors can even think about creating their own model.

But to be clear, you think the concerns around margin are overblown and not justified?

Well, it's a risky business where there will be a winner among them, whether it's—I love what Lovable is doing with Anton—but Replit and Vercel are incredible companies too, or products too, that are doing some great work. I do think at least 1, maybe more than 1, will create something special. The market is so big that all of them can create something special.

But yes, there will definitely be a loser in that mix too, and that company and the margins that it is carrying just do not make long-term sense, and probably the same for coding apps. But I think it's a cool bet—not even cool in the sense that it's a nice application, but it's an ambitious bet that they are trying to win with the biggest companies in the world that will try, or are trying, with Google and Firebase.

Sorry, I don't remember the name of their product, but the Lovables of the world are winning.

That goes to your point on brand. Neither of us can remember Google's product name, so it's just interesting.

Anton is a phenomenal marketing person too. The way the founder is doing it is amazing.

One thing that's also really interesting is that you were very horizontal in your customer base from day 1, which, respectfully, I would advise you very much not to be. I think Philippe was telling me he was advising you the same: be much more targeted in your ICP and have a clear mind of who that customer and user is.

You were much broader and more horizontal. If you're advising founders, how do you tell them, when they're launching from day 1, whether to be horizontal versus vertically specific?

I think if you're launching something very new, very different, and you don't yet fully know a subset of the customer base, but you know that there's a bigger one, I think horizontal is completely fine. If you know and you have domain expertise and this is the category you're trying to win, I think go vertical.

You say go vertical. Normally, when you have verticals, you have a maturing of the organization; you have titles. Titles are something that you decided to get rid of. This is very counter-traditional in terms of structure. Why is it better not to have titles?

There are some incredible companies that, of course, did something similar, Stripe being one example. But in our case, it was especially then a super-small team. We had a team of 5, and a lot of people joined. What we wanted to optimize for was that the main thing that matters is impact. You can join, and you can be the most impactful person from day 0 in any of those teams.

The title shouldn't define your level of decision-making. Second, with the small teams, what happens is that you have a lot of those small teams. If you start looking at who gets the title or not, it just becomes a distraction, given that teams are those small units that just keep executing.

The third thing we wanted to make very clear—and still do—is that if you're joining ElevenLabs today, you can transition to being a leader of any team, of any function, super quickly, and titles felt limiting to that. We felt like people joining and seeing the wider organization already having a set of roles defined for them would make this a little bit harder.

You can have lower tenure and be a manager of people with much longer tenure if you're the right person. So we felt titles were taking away from that.

At the same time, we have a good structure internally within the subteams, such as who the current lead of that team is who will make the decision if people cannot agree on the right path. But that person can change, and it's not guaranteed that the person will remain a lead forever. In any other company, when you give a title, it usually stays forever.

Speaking of titles, a common criticism of European tech and scaling is that we don't have these titled people who've seen significant scale: your VPs of sales who've seen $1 billion in ARR, you name it. How did you think about broaching that? Do you want to get top US talent here, or do you want to grow talent here?

We want to grow talent here. We match a lot of our current talent with the advisors from our network of US investors, and the goal is to grow people. We love growing people. Our approach is almost that, if we can, we want to take a bet on the person growing rather than bring them in externally.

But we know that to grow, you sometimes need mentorship from someone who has done it. So we try to tap into all the people who have done it and match them to people across the organization.

Anton said on a show with me recently, from Lovable, that building in Europe is building on hard mode. Do you agree with that?

I think it is. I think it is building on hard mode, but there are some great advantages too.

What are the advantages?

I think the first one is that the talent here is incredible. You just need to know how to get it.

What do people get wrong, then?

In general, there is a good subset of people that really want to work hard and create something special, and they just don't have that opportunity because there aren't ambitious European companies trying to do that. The only way they can do that is to work for US companies.

I think now there are, like you mentioned, Lovable. I think they're showing that ambition too, and there are just so many more, like Sana recently in Sweden. We spoke about Synthesia. So there are quite a few companies that are ambitious; they want to show it.

I think the people joining want to be part of the ambitious company that is competing on a global scale. I think the thing that—and maybe in the early days, I was even worried to a large extent—I think you want to be able to build from Europe, but not build only for Europe.

I think those 2 get conflated, where sometimes building from Europe meant, okay, you are building for the European ecosystem, which isn't the right thing. You still want the global aspiration.

So even now, when I think about describing ElevenLabs, I think about us as a global company. We are a global company. We want to win in the US, win in Europe, and build in Asia. We have the best team, and most of the team is in Europe because the talent is incredible.

So you think it is wrong when our American friends say, “Hey, the Europeans just don’t work as hard as we do”?

I think no.

Yeah. Is that quick?

I think you can find people who want to work harder. We’ve had it, actually, at some point: a team where we hired some people from the West Coast of the US, and our people—of whom we have quite a few from Central and Eastern Europe—were like, “Oh, yeah, they don’t actually work as hard as we do.”

I mean, we have such a—credit to the team—true missionaries who are there on the weekend all the time and really care. They really care about the success of the company beyond just working hard. They feel part of the company. So I do think you can find people in Europe who are phenomenal.

Is it bad to hire people who come because you’re a glossy name?

It’s an interesting one because, in the early days, we didn’t have much inbound. All of that was outbound, and it was easier, in a way, to find people we felt were right because you didn’t have to filter through the noise.

Now we have so many people coming because of the AI buzzword, or because we’re now at scale, which makes sense. But no, I think it’s fine. Not everybody is the same amount of a risk-taker, so I’m not in any way discarding them.

What’s been your biggest hiring mistake, and what did you learn from it?

I think this is probably a traditional one, but as I think about bringing people in and growing them into the company through the full hiring process, you hire a person and, of course, you frequently have very little time to assess how they are. Then, as they are in the company, you have a little bit more signal.

I would have taken some decisions quicker than I did when separating from people. Usually, if you’re unsure in the interview, let’s say you bring them in and give them a chance, and you’re not sure in the first weeks or months, you should separate straight away rather than keep giving them a chance. I think that’s happened a few times.

What’s the hardest role to hire for?

Researchers.

Easy. When should a founder no longer be involved in every hire?

Well, I hope I’ll be involved throughout, so we interview everyone.

You still do now, at 250?

Yes, we do. We hope to interview everyone for as long as possible because it’s a good signal of who we bring into the company. We get to meet them.

But, of course, we still think the company is shifting in terms of how we’re approaching that. Now, even more, we are optimizing for engineering and technical skill sets in all parts of the company than we would have probably 6 to 12 months ago. So it’s increasing in some aspects.

I hope we’ll interview 1,000 people. The thing isn’t so much when to be involved in the interview; it’s more how many people you’re trying to hire within a specific period of time. If I try to hire 1,000 people in a month, it’s impossible because it would be more than the time we have. But if it’s—

In a year, how many people will you have?

I think we’ll have 400 by the end of the year.

By the end of this year?

By the end of this year. I think we—

That’s a lot. You’re almost doubling. You’re adding 40% in.

That’s almost all in 3 to 4 months. We have 50 offers or so out, or people joining. So, 250 now, with roughly 30 to 50 people already scheduled to join.

Respectfully, small and mighty, that’s not 150 in 3 to 4 months.

Still small and still mighty. But we have a pretty global team now, where we are trying to bring a lot of our go-to-market and engineering into every location we’re in.

We think we can parallelize that: building in Brazil, Japan, India, and Mexico. We are really going for those local nuances, too, where we are building small outposts everywhere, and I think we can make it work.

Do you care about revenue per head in the long term?

In the long term? Yes. We want to be an efficient company. I think now we have a very good metric for revenue per head. It’s a good way to show whether you are an efficient company.

But if we think there’s a path for us to get there over time, we would easily take the new hires that can help us. You know it extremely well as an investor, but one of the key metrics is retention, or how you think about NRR for any of the clients.

Bringing people in now to help us get distribution before the competition does might decrease the revenue I get. But if the NRR keeps growing and I don’t hire more people in the next 5 years, that metric will of course change.

Can I ask what revenue you’re at now?

When will this air, plus or minus—

2 weeks?

2 weeks. Two weeks. Okay, so we crossed $200 million now.

Whoa.

So it’s a good number.

That’s amazing. Yeah, that is a great number for 250 if you’ve got that now.

Yeah, it’s pretty good.

What was it at the end of 2023?

At the end of 2023, I think we were $35 million.

$35 million.

$35 million.

You went from the beta launch in January to $35 million?

To $35 million, we did. So we did 20 months to $100 million, and then—

20 months to $100 million.

Yes, I think something like that.

And then around 10 months to $200 million, a bit longer—15 months, maybe. I’m just crying as I write these numbers. Oh, it’s so depressing. I mean, you know, I’m so happy for you, and—

Well, the thing is that, of course, it goes quickly, but it can also go quickly down.

But is your revenue not relatively sticky?

I think it’s relatively sticky. Large enterprises are the majority.

Our biggest part of the business now, and what we are obsessing over, is building an effective conversational-agent platform. Most of our biggest customers are building on it.

Who’s your biggest customer? Not the name, but what size?

Our biggest contract is around $2 million, and they are mostly in call centers, customer support, and personal-assistance businesses. All of those companies are orchestrating a combination of the stack: speech-to-text, text-to-speech, and then bringing in a lot of the integrations that we now create.

That’s on the enterprise side. Cisco and Twilio—these are not the contract sizes—but Cisco, Twilio, and recently working with Epic Games are some of our biggest deployments of the work.

Then, of course, we’re in the lucky position that we still have a huge self-serve distribution of creators and developers building all the time.

The one I really want to ask, if it’s okay, is this: you got to $200 million in 10 months.

Yeah.

What’s that to $300 million, then? You did 20 months to $100 million and 10 months to $200 million.

Well, we are an ambitious company, so we hope to—

Can we do 5 months to $300 million?

We would love to break the record. If it’s healthy revenue and we are creating good work, we would love to break the record.

What was the price of the last round?

$3.3 billion.

All rounds divisible by 11.

Okay, $3.3 billion.

And then you did that when you were at $150 million in revenue.

We did that when—no, lower. We were somewhere between $100 million and $120 million, I think.

Okay, $100 million to $120 million. But I’m just looking at that going, you did that in January 2025.

We announced it in January 2025, and we did that towards the end of 2024.

I’m just looking at it thinking, okay, so they’re doing end-of-year revenues for 2025 at $250 million to $300 million. It’s quite a cheap deal—11% or 12% of your revenues.

Oh, like 4 or 12 months.

Yeah.

Well, it was—you know, I think the first—so, we did the round in October of 2024. We were probably at $80 million when we got the term sheet, and then it was signed.

You need the money?

The way we approached any of the fundraises was: can we bring some of the bets forward? In that case, it was spending even more on models, expanding to multimodal; bringing our work internationally, so expanding into other regions; and doubling down on the agentic platform.

That meant starting to build even more of the true enterprise functionality, whether it’s the reliability you need, integrations with Salesforce and ServiceNow, or SIP trunking. Investing in those things was the most important.

At the time, it was 30 times revenue—30 times current revenue—so pretty good. Then, how do you balance between focusing on and executing according to plan versus being able to do more? I understand the benefits of being able to do more, but you can probably always do more with more money. It doesn’t mean it’s the right thing.

I think the variation on this is: can you parallelize a new effort without distracting from the core work? That’s roughly why we are trying to approach this like: can we bring our work to a new place? Can we create a new product experience without affecting the core thing that’s actually important for our users? If we can, then we’ll likely invest, bring in the people, and do it.

If it's distracting, then it becomes a question of whether it's worth the upside.

When we think about doing more, the question that I think we forgot earlier was: what's the biggest line of business in the future that is not a very big line, or is nonexistent, today?

It's interesting because, on a relative basis, I think our agents work is already huge, but I think it's just scratching the surface. I think if we play it right, it's a multibillion-dollar, revenue-generating business just from creating voice agents and going deeper.

These would be voice agents that you sell to companies to manage their customer support.

That's right.

Gotcha.

That's right. I think then you can go deeper. You can, of course, expand from voice into conversational agents. By that, I mean you start building an omnichannel solution with email integration and WhatsApp integration, which is that more classic customer support.

Do you sell then to Intercom and Decagon today?

Today? We have a public partnership with Decagon. So we do, but of course, as you start going deeper, it depends a little bit on where Decagon goes and where we go. Are they going to verticalize, or will they continue horizontally and go down? We verticalize more, so there might be some areas where we overlap a little bit more.

But today, given that we have a very partner-friendly, horizontal approach, we treat all of them as good partners.

Speaking of good partners, right now we're doing the human-plus-agent. We're so friendly, and we make each other better—and then there's a time when human-plus-agent just becomes agent. Do you think we're seeing a lot of resistance from employees within companies toward agents coming in?

We do, but the way we've seen this transition happen now is that you will have more specialized humans, where the agents are taking more of the manual parts that nobody really wanted to do, or that were effectively very easy to do or didn't require domain expertise.

A good example is appointment scheduling. If you are doing a refund, all of those, of course assuming you have the safeguards and authentication in place, are pretty easily done with AI. But then suddenly, if you need to help a patient navigate the outpatient flow after leaving a hospital or understand an analysis, that cannot be done with AI. There's too much at stake, and you need deep domain expertise.

So we've seen the transition where that side of people helping in that last mile is even more valuable. Of course, AI can help with those easier tasks across the board. I think this will continue. The percentages will shift, where you will have even more automation as it goes deeper, and even more value assigned to the people doing that domain expertise, because ultimately it will help automate the task at hand.

Well, does taking money from Sequoia meaningfully move the needle in a way that it doesn't from other funds?

I think so. Our first round was with a16z, and it did meaningfully move the needle for us. It was very clear that people respected you in a way that they didn't before.

Yeah, I agree.

And then Sequoia came in in Series B, and that doubled down on that perception, where it's like, okay, a16z and Sequoia are part of the company. It's also very rare to have both of these investors in any of the companies. So it did help. A lot of investors—sorry, a lot of clients—would respect that, and now ICONIQ on top of that is just an incredible mix.

NFDG, too. I think NFDG is, of course, something clients will usually not have that perception of, but investors do.

Investors do.

But investors do, and some of our engineers or users really admire and trust Nat, and I do too. So he's great.

You're a very strategic asset when you look at what you have and what you've built. Have you had acquisition offers?

We did have acquisition offers.

Did you contemplate any of them?

Honestly, we always do the basic diligence and let our investors know that we had this, and usually—

What was the largest one?

Well, the largest one didn't have a monetary figure. The ones that we did have were interesting conversations because we would frequently try to go into a strategic partnership, and then they would be like, “Oh, can we consider—are we open to M&A activity?” But it would be a while back, and I prefer also not to share.

Was it tempting?

It was a tiny bit tempting, but in a way where we— in the first one or two examples, the first one was when we were doing Series A. It was going really well, and then, of course, we were unclear how this would continue. It's like the very beginning of the curve.

The first one was like, okay, this is the first time we have ever done this. Let's understand what they are and what they're offering. But we were more interested in seeing more about the process itself, how it happens, rather than actually giving the company.

Because then it was clear: okay, they don't actually want to acquire us. This is what we understand, and we need to be part of the company. So we were a flat no. Since then, through any of the conversations, now that we know that this is even approaching M&A, I think it's even more so now. I think we are more confident.

Did you do secondary?

We did secondary.

And that helps?

So almost every round, we do secondary and a tender offer for all employees who have vested stock, so they all feel that there's actually liquidity for them. I think it's valuable because we are betting on something huge, and you want to know that you can take the risk to bet on some big outcome.

I think it's very easy to say that financial gain, per se, is not important, and I think it isn't, but there's this basic layer that people want to have covered.

I think it is: paying for childcare and paying for a house.

No, of course. Exactly. That's what I mean. It's not the goal in itself, but you want to have this basic level of a good life covered, which we are extremely lucky as a company to have and to be able to offer.

But now the aspiration is much bigger, especially now that this basic layer is covered.

The thing I often think is, how many great European companies of the last 20 years would not have sold had we had secondary and liquidity options available at the time? So many did sell because we didn't have that, and it was so meaningful.

I think it does help with the perception, where, to some extent, I think you can put away any of the greediness that comes with money just by—

Taking some of the risk equation away.

Dude, we're going to do a quick-fire round. I'm going to say a short statement, and you're going to give me an immediate thought. Does that sound okay?

Sounds great.

What do you believe that most around you disbelieve?

That you can build a company from Europe at global scale.

You don't think people still think that? You don't think we're moving the needle a little bit?

I think you are helping, and I think many of the people in our ecosystem are, but I don't think most people do.

Maybe another one, which is not that specific to company-building: I do think voice will be the interface to the technology around us. It will be the primary interface for a lot of technology around us, which I think most people would not agree with.

I'm not going to let you wiggle on this one. You can buy OpenAI at $300 billion, Anthropic at $170 billion, or xAI at $120 billion. Which one do you buy and which one do you sell?

I don't know if I like this question.

Anton answered it. Anton, for context, answered that he'd buy Grok and sell OpenAI.

Okay, let's keep it positive. I would buy OpenAI, but I love Anthropic. If I was on the coding side, I think I would be buying—

Mandate Cursor?

No, we don't mandate. You use what you think is most valuable.

What do most people use?

Most people do use Cursor. Yeah, some people use Claude Code, but still more Cursor. I think it's shifting a little bit.

To your other question, I don't remember which investor it was, but they said something that I think about when making that decision: I would happily invest in a lot of the products I use. I use ChatGPT very often. Every so often, I will use Anthropic for testing where we are with the space, but I think they're investing more in coding rather than Claude, the consumer product, and OpenAI is clearly investing a lot in ChatGPT, the consumer solution.

What have you changed your mind on most in the last 12 months?

I think that we—and I do know it's a rapid one, that's why I'm giving a quick answer—previously would not do any product innovation if we knew that we were doing any research initiative ourselves. I think now it's shifted, and we will sometimes explore a product with outside research, even if we don't build it internally.

Is speed of ARR growth a [__] metric?

It depends on the time horizon, but in general, yes. I think the speed of ARR growth doesn't matter.

What's your favorite consumer brand today, and why?

I really like Eight Sleep. I don't know if it's a consumer brand, but I do like it; it has changed quite significantly recently. I don't know if that's the brand. I am a huge user of Google Maps and love Google Maps. Lovable—I really like it among consumer-ish applications.

You can be CEO of any company for a day. What company would you be CEO of?

I would choose Google or OpenAI. I would know the know-how of some of the incredible models. I think more Google. I would say Google's Gemini 3 models are incredible innovations, and at the same time, the scale of operation would be interesting.

Super bullish on Google. People question them with the golden.

Hey, you didn't give me Google in the previous question. I would invest in 300. I would invest in Google, but that's not an option.

But you're super bullish on Google, even with the ads model potentially—

No, no, it's not like super bullish—definitely not super bullish—but I think Google has a good future. Especially recently, they are catching up in many places. So I think they have—I think they are definitely in the race.

I'm creating a title for you. You're going to be the president of Europe. I'm aware, for all of our American listeners, that Europe is not a country, even though you like to collectivize it and make it one. President of Europe: what's one thing you would do to give the European ecosystem a higher chance of success?

And my word goes into law immediately.

Yeah.

I would proxy a lot of AI law to U.S. law. I know this has a huge set of repercussions, but I would try to follow exactly how the U.S. is approaching a lot of AI-related regulation and just implement the same, or create another state in the European Union, in Europe, that people can opt into, which follows that law, to not make it too hard given all the repercussions.

How important is founder brand?

That's an interesting one. My answer here would be, "I don't know," because in many ways, as we thought about ElevenLabs, especially in the early days, it's the people that build the company. In some ways, we still don't know to what extent having some people who are very much out there—let's say I'm now on this podcast—takes away from that, and we want to make it complementary.

The reason we are successful, I think, is to a large extent because we've created incredible research. The engineers are just grinding and creating the best product experience. The go-to-market team is inventing new ways to combine self-serve and sales. So it's all those parts, supplemented by operations scaling the company from fewer than 100 to now 250 in the span of 7 months while keeping the culture intact.

These are all such hard things, and with hypergrowth, I think there's less time to be able to appreciate all those individuals than you would otherwise. I sometimes worry that having too much of the founder brand takes away from that. But my mind is changing a little bit. I think maybe you can elevate that by having a founder brand out there.

Final one, and it may be a little bit of a thought, but what's the single best piece of advice you've been given that you think of most often?

Well, I don't think it's something I think of most often. In recent times, I like what Peter Thiel said about the biggest risk being not taking the risk, and staying—not taking a decision or staying—

What risk did you not take that haunts you most?

You know, it's shifting, because if you ask me what risk I should have taken, I usually hope or try to take quick action on top of that. But one that we are considering is an acquisition of another company, which is a big company, and that company is in the hundreds of millions of dollars. So it would be a huge risk to bring them in, and we think we can do better internally. So I think we want to take that risk.

Mati, this has been so much fun. Thank you so much for putting up with my meandering, and you've been a fantastic guest.

Thank you, Harry. It's a pleasure to finally be able to speak together.