Klarna CEO Sebastian Siemiatkowski on 20 years building from Burger King to 100M users and the AI moment in fintech

brands. I guess Trump has a liquor brand. And so this is this is a proven proven strategy. Um anyway, we have our first guest of the show, our second guest technically since we already had that call-in. Uh the CEO of Clara. Welcome to the stream. How are you? What's going on? Welcome to the show. Thanks for having me.

I just realized the light in here is very very odd, but I hope this is fine. I think you look great. You're good. You're good. Great. Uh thank you so much. I imagine you had to stay up late for this, so we always appreciate it when someone can do that. Um uh would you mind getting last man in the office?

It looks like leading by example. Exactly. This is what your team said. There's some more people here than me. Yeah. You No, no, your team said this about you. They said that you're you're you're obsessed with artificial intelligence right now. Obsessed with digging in the weeds, tinkering with stuff.

Uh probably goes back to the the early history of the company. Maybe that'd be a good place to start. Uh h how did you actually like what was the inciting element or the inciting incident that led you to start the company? Wow. Well, so as you may know, I've been at this for uh longer than many others.

So I'm at the 20th year of this venture. An overnight success. Overnight success. We love an overnight success. It's been easy. Exactly. U But I was 23 back then, so I'm not that old yet. Um 23. But did you ever have a real job? What? Did you ever have a real job or did you go straight into entrepreneurship?

I had many jobs. I had many jobs. I uh you know I started working when I was a 15year-old at Burger King. Nice. Flipping burgers. That's great. Uh were you any but were you were you any good? Uh flipping burgers. Yeah.

I think it was like we actually had a competition for like making war who warpers really fast with quality with quality. I think I I think I scored pretty well. Good. So yeah, walk me through history of greatness. Uh so 15 you're flipping burgers.

23, you're starting a a massive tech company that would go on a 20-y year run. Uh what happened in between? Uh very shortly. Well, first look, you know, for whatever reason, I always dreamt to be an entrepreneur. Kind of odd thing to do as a kid.

My my father really wanted me to be a doctor, but I was inspired by Richard Branson and Ingva Campra, the founder of IKEA and stuff. Yeah. And I went to business school. Bad idea. I should have learned to code the first thing I did, but that was a bad idea.

But anyways, um and then you know after a few years of just like testing a lot of different jobs, I ended up by very odd story going around the world without flying coming back too late to continue my university studies and as a consequence ending up as a sales guy at a factoring firm that was doing like account receivables.

Last thing in the world I ever thought I would work with. Yeah. And at that point of time realized there was a lot of these very small newcomer e-commerce companies that were selling stuff online and they needed a better payment solution because their customers didn't trust them. Didn't trust that they would ship.

So they preferred the idea of buy now pay later where you would first receive the product and then kind of pay for it once you had it. And that was pretty much the beginning of Clona. That makes sense. Yeah, it's awesome. Um uh yeah, walk me through, uh the the more recent developments. How big is the company today?

Um so we're about 100 million users worldwide. Mhm. Uh we do about a hundred billion dollars worth of payments volume annually. Uh there's about half a million merchants, but a lot of them are really large. So there's like Nike, Macy's, Sephora, uh Walmart nowadays.

And um and yeah, what else on on number of employees about 3,000 people? Yeah. Wow. Yeah. What was the uh what was the initial go to market? I seem to remember um there were specific dynamics in Europe around uh cash on delivery being a unique kind of vector into growth.

Can you walk me through kind of like um what the early go to market was and how um and and it like it seems like you almost like leapfrogged the technology curve a little bit.

Well, the truth is funny enough even in the US if you look at the beginning of credit bureaus they actually a lot of them started as mail order companies where the first credit was provided actually by you shipping a product and asking for cash on delivery and then the third time you shipped to that same address you would say well this customer seems to have money every time I ship to them so I'm going to provide them a small credit credit line that was actually the beginning of a lot of credit bureaus even in the US back in the early 20th century.

So, and kind of same thing happened across Europe, across the US.

So, the idea of doing like cash on delivery and then turning that into credit line based on the performance of that customer always having cash at hand when they were buying those product or paying for those products that they ordered at that point of time mail order catalog orders, right? Um that that actually worked.

So, in the end to us when online came along, most people were offering debit card payments. And the problem with debit card payments is the customer ships money and then maybe receives what was on that picture that they ordered.

And so basically offering credit uh not necessarily for the lend for purpose of lending but more for that kind of security of getting and touching the product before you paid for it made a lot of sense especially in markets where credit cards were not as prevalent as they were in the US. Right. Yeah.

What was the uh like as you were looking at the global map of expansion? Uh what what what informed like where you go next?

Like like walk me through the different markets that you that you had at the top of the list and how you chose to kind of rank those as you expanded the How soon did you know the American consumer was undefeated and you had to be you had to had you had to be here. That took a long like stuff. Yeah. We like to buy stuff.

Exactly. funny enough took a lot of years before we realized you mark. I mean first and foremost like when we started you have to remember at that point of time there was one one globally successful European tech startup and that was Skype. That was it. Like Skype was the thing.

Everything else had been launched out of Silicon Valley. There was no real European tech success of that size that had like global consumer you know recognition right and hence to us it was like okay where are we going to go from Sweden?

Let's go to the massive market called Finland that is half the size of Sweden because it's very close and we kind of know what it means, right? So, we kind of started with expansion in the Nordics, then went to like Germany which was a huge thing for us.

Figure that out and then went to UK and we ever we always thought that we never like figured out how are we going to come into the US market because credit cards are so prevalent.

But the funny thing that we hadn't realized was due to the financial crisis in '07, uh there was a lot of regulation introduced that made it much harder to get a credit card.

One thing that kind of sounds sane is that credit card companies couldn't advertise in campuses anymore, which is one of their like most common places to sign up credit card holders really. And so there was a lot of that regulation.

So it meant that between 2010 and 20, debit cards in the US grew like a credit cards volume grew about 2x, debit card volume grew 10x, right? And so actually rather than us figuring out how to be successful in the US, it was the other way around.

the US market became more like the European markets with much higher debit card penetration where buy now pay later and the kind of products that we offered became much more relevant suddenly for this new uh consumers that didn't have credit cards to the same extent or didn't have as much credit limit as they used to have historically right yeah what was the the roll out like in Finland Germany the UK the US um from a regulatory perspective do you have to get like banking licenses there approval is it mostly like regulatory or are you more focused on engineering challenges, setting up servers, localization, all of them.

All of the above. All of the above. But I think from a regulatory perspective, um, we quickly identified that it would be a benefit to be fully licensed over time.

So about 10 years ago, we did get our bank license and we were fortunate that EU had implemented legislation that meant that you could take your eur uh Swedish bank license and passport it across all the European countries. Right?

So we run on the on the Swedish bank license, but it's applicable for all and it included UK up until Brexit, right? So yeah. Um so that worked um for a long period of time, but obviously technology was an important challenge for us. Yeah. As well. What are the knock-on effects of having a bank license?

Uh I I I talk to fintech founders all the time where they it feels like they're in the money moving business, but then they you find out they have all this float and they're making a ton of money on on treasury bills or something like that.

Is there an aspect of your business where uh you can actually drive more profit and bottom line on the basis of having that bank license or is that less relevant and then I want to go from there into stable coins and what's going on uh and how that might be changing the market.

So I think that the in our case what you're referencing is kind of funny enough something we have underutilized the clona which is the ability to generate the deposits and kind of earn interest of those deposits. Right. Yeah, funny enough.

And that's something we're addressing right now where we're introducing more products and services in that area.

Um, but where the big advantage for us was, if you look at like the history over those 20 years of a lot of fintexs who were credit providers, a lot of them offloaded their balance sheet and did so effectively up until at some point of time something went wrong in the underwriting model and it turned out that the losses were higher than they expected.

At which point of time the people that had taken on their balance sheets realized that they wanted to shift the liability back to the fintech but the fintech was not capital-wise equipped to take that on meaning that like their valuations either collapsed or the whole thing went bankrupt. Mhm.

So, and it's happened over and over again if you look at a lot of the like fintex over the last 20 years. In our case, the benefit of having deposits is we are a bank. We're fairly capitalized as a bank. We need to hold the same capital that, you know, any large bank needs to do in in comparison to our balance sheet.

And it means that we're much better equipped. Like even if our losses go up and down or whatever, it doesn't change the sentiment of our balance sheet. Doesn't mean that we can dry up funding. We're not dependent on the funding in the market. the whole the whole thing that went down in 07, right, which was the problem.

We're basically we only have deposits and that are insured by the government and and so forth, right? So, it gives us a much more stable way of funding funding the bank basically. Then on stable coins, is there any benefit to you there?

We've heard stories about lower fees and faster payments and obviously there's a massive, you know, the frequency of payments across the Clarin network is massive. Uh are you looking at it? Have you already looked at it? What's your takeaway on on stable coins or uh the uh the genius act that is seems on the customer?

We've had fintech CEOs on the show that have said actually traditional banking rails can be pretty awesome if you utilize them properly.

So curious how what your I think that the I have been one of the crypto skeptics for a long period of time and um and the reason for that is when like I started playing around with this stuff in 10 years ago I felt that you know I always ask myself like how's this technology going to make my mom's life better like how is it going to make our products and our customers experience better and I didn't really see any any benefits of the technology apart from the fact like people have been speculating that it's an interesting asset an alternative asset to hold right like Bitcoin or whatever.

Um but then I gave it a second chance, a second look now one about a year ago started looking at the next generation of stable coin implications and implementations and I very much agree with the fact that like now the technology is showing strong promise both in its speed in its cost uh etc etc versus the traditional bank rates especially especially obviously in international transfers h in corridors that are more like you know you're moving mainly from like UK to US you're probably not going to make that big of a But if you're moving money from, you know, Brazil to, let's say, Greece, uh, oh, sorry, that was a Euro country.

Stupid idea, a stupid suggestion, but another one which, uh, with a smaller currency, you're going to have a a very real benefit of it, right? So, I think that stable coins definitely are going to be a major thing.

I think also like crypto is here to stay and the bitcoins of the world are here to stay as assets, right, that people find are interesting assets in their portfolio. So, I don't think that's going to change either anytime soon. Yeah, I'm always interested in in where the value will acrue with these technological shifts.

Specifically, like with stable coins, you can imagine if you're going from two-day settlement to instant settlement.

Well, if the customer would accept one day settlement, all of a sudden you have one day of float that's going to sit there that you could kind of keep as your own margin and maybe put in T bills and maybe it gets to be a big number.

And so I've been I' I've been kind of trying to understand their like when there time is money and when you know when time decreases and and systems get faster you can pass all of that along.

You can pass half of it along um and and and there's different things that different companies can do um based on the consumer practice. I I think you're totally right.

That also means that in in in some degree which is something I tell my investors right it's a race to the bottom like this payments industry has had a you know combined revenue globally addressable market of like $2 trillion or whatever it's not going to stay that way.

There's no there's no real good reason why it's more costly to send money than send an email, right? There's a little bit more risk involved.

there's a little bit more like you know security involved but in essence you're just like shifting data between two accounts and so like in in some to some degree that like that those kind of solutions will drive and that also means that the revenue generated from these services will be significantly lower than it is today right what's your broad uh not not to go too broad but what's I'm curious to get your broad market outlook right now you have this interesting view into consumers companies all over the world it's It's a very interesting time right now.

We have geopolitical instability. We have, you know, and and effectively war. We have uh tariffs, trade wars, all happening at once. Um yet at the same time, you know, we've seen Circle Public and is now trading higher, I guess, higher than than Coinbase, you know.

So, it seems like the the general excitement around um capital marches on. Yeah, capital is marching on. But but yeah, how are how are you thinking about uh markets broadly from your view?

To me, um from my perspective because of kind of the deep um interest that I've taken into AI and tried to understand the implications on how to apply the technology. I am very much convinced that it will have implications on white collar jobs.

Effectively you can have like economical swings that you know recession non-recession it may you know higher inflation higher cost pressure and so forth may have some implications for consumers sentiment of spending you know I like I'm going to spend a bit more I'm going to spend a little bit less and so forth but the but the true life-changing moment that affects people's true ability to pay back right is to a large degree uh loss of jobs right if you if you lose your job then it severely impacts your ability to do So I think that the key question now is to what degree we're going to see an acceleration of loss of jobs among white collar uh people which in itself then potentially leads to a recession and less spending in that group of people which is obviously a group of people that have had a lot of money.

Yeah. Um, so I think that's the area that I'm like mostly because the tariffs so far obviously have been announced and kind of changed and then announced again and changed and so forth.

And now it almost feels like if I look at our merchants when the first tariffs were announced on Labor Day, they really did adjust their marketing spending. We saw changes in the transaction volumes.

We saw for example some of the larger Chinese companies that we work with would shift their marketing spend from US to Europe and would try to offset more of their inventory in the European market as opposed to the US market etc.

But as now also like there's been more and more like turbulence around these things people start also discounting it saying like yeah it was announced this week but who knows what's going to happen next week and so we're not going to draw too early conclusions from this.

we're not going to change, you know, the business of what we're doing until we know it's for real, right? And until it's actually going to happen. So, I don't actually see currently in the trends the direct implications of the tariffs themselves in our business.

But, and I'm more just like trying to second guess if I'm going to see a major change in the uh in job.

and what are you trying the other thing is you're kind of hedged in some way and that that I think people would be more likely to use CLA if they have if if they you know uh in in some ways have reduced spending capacity or or you know you know more more likely to leverage a product.

I'm curious just getting getting you know more specific around this what is your up updated hiring strategy throughout the the rest of this year and next year? what what are you um you know hiring what areas have you are you hiring in or or kind of shifting away from?

So we um about almost now two years ago um after experimenting with Chachet and so forth for a few months, our conclusion was that this will lead um this will change our ability to get more done with less people and that we believed that in an AI world, revenue per employee will be a very key metric because the higher revenue you have per employee, uh the more profit you make per employee means that you can spend more on marketing, means you can acquire more customers, which you can scale more.

So if you have a low revenue per customer, you'll be less competitive basically. And so uh for us the question then was how do we catch those efficiency gains? And so we decided to stop hiring about 2 years ago. At that point of time we were about 5,500 people.

And since then we shrank to about 3,000 employees that we are currently which is basically through natural attrition. Uh because you know in a tech company about 20% will leave on an annual basis voluntarily because they're going for other jobs, other opportunities.

And so by just being very strict in not rehiring, uh we have been able to shrink down the company and and and continue to do so. Now, we've hired a little bit. Doesn't mean that we haven't hired at all.

We've hired in some areas and so forth, but if you look at the net, um you can see that we've shrank from about 5,500 to 3,000.

And we uh and then what we also done is internally we've been very honest and open about this and we've communicated to our employees that um what's going to happen is the gains because obviously this means our labor cost is shrinking uh while our revenue actually has been growing right so revenue as a consequence of that our revenue currently went from about 400k up uh to about over a million now and which is kind of equivalent to where Goldman Sachs is now next yeah thanks next step is um is kind of you know Apple Netflix being $2 million.

I sometimes joke because like Only Fans is at like $30 million for Yeah. Don't go there. That's Don't look at Tether. Don't look at Te te Tether will have you looking at uh net income per employee. Yeah, exactly. So, I'm not sure that's a great the comparison you want to have.

The the thing is that like um the anyway so we are continuing down that path and and the point is that um uh what we've communicated internally however is that as the total labor cost shrinks um we are reinvesting that in an acceleration of compensation for our existing employees.

So as they participate in the efficiency gains we have been able to accelerate both equity and cash compensation at CLA to our existing employees more than we had in previous years. And that's kind of a way for them to share the benefits with the shareholders of the efficiency gains that this creates. Right.

So if you're a new if you're if you're a new grad and you want to get a job at CLA and make a career out of it, maybe be there for 20 years and earn a lot of stock and cash over come for your job. Come for your job. Uh what should a new grad do? Should they learn prompt engineering?

Should they learn to vibe code or program? or should they be focused all on networking and management skills and people skills and the things that the robots seem to be mediocre at? I think very clearly is there's one area where we have not reduced um number of people at all and that's sales.

In our case, I mean like B2B enterprise sales. Um so we have people locally on the ground across the world in over 50 offices.

you know, whether they sit in in China next to to to Shien and Tik Tok and so forth or whether they sit in Portland next to Nike or whether they are in, you know, uh somewhere on the east coast or in parts of Europe.

Um and those obviously there's no like immediate expectations that this those are affected by this, right? Rather the opposite like that human connection in in customer relationship is super important. I've never I've never seen a a Chad GPT have a power lunch and a steak dinner.

Yeah, it's impossible till we get better humanoids. Um, but on the rest of jobs actually it's been quite interesting because for a period of time now we said okay let's stop hiring all other competences than engineering. So if you look at that 5,500 of them 1,500 were engineers. Today of the 3,000 1,500 are engineers.

So actually the number of engineers have stepped kept the same while it's all the other like marketing you know analytics um whatever it might be that shrank.

Um now with that said however in the last like months because of the acceleration of uh claude code and and you know and cursor and all these tools what you're seeing is now is a group internally at Lana who are like a few hundred business people that in the last two years have you know learned how to code not necessarily like the most complex things but inspect code understand code interact with code uh but doing so with having all the business knowledge understanding what the customer wants understanding a lot of things that engineers sometimes have struggled with like they can't articulate uh why we're building the product for what purpose for how it's going to solve customer needs and so forth and they've been maybe sometimes more focused on just like latency or technical implications or you know etc.

So I kind of see a convergence now of the truly kind of renaissance people that are kind of you know both on both sides that can understand the code but can also understand the business those are the ones that are now truly becoming the most productive in my opinion.

what is your framework for buying AI tools versus building AI tools internally? Right.

So we we have been fairly vocal about the fact that you know as part of acceleration of AI clana one of the first thing that we remember chat GPT came along there was this like hey I can now connect all my PDFs to chatd you know and like and people were like oh that's awesome.

Uh the only problem with that is that like the truth in data modeling and data science has always been like [ __ ] in [ __ ] out, right? So if you feed it tons of like internal stuff that is contradicting, inconsequential, not truthful, whatever, it's just going to get lost, right?

Just like some of our MPs get lost when they start navigating the internal workings of CLA, right? So so the first thing that we realized we can't silo the data anymore. We have to have a a common data model. And that meant that we started removing software as a service uh providers, you know.

So we um we closed down Salesforce, Workday. Uh I think the next big one that would go is Atlash and Gyra and and and so forth. Uh but then there's also like a long amount of smaller software as a service vendors that we have utilized.

So actually if you kind of count it in total, there's over 1,500 that we have uh closed down in the last two years.

Now obviously some of them are like super small like some plugin to gyro or whatever like it's not like every one of them is like a tremendously big software as a service application and then we've kind of harmonized and standardized our data layers so that we have one data that we can then let the LLM interpret and run on top of uh allowing us to uh to do more with less right um so currently there are a few things like we will use 11 labs for voice uh we will use like specific people that are really really proficient.

We use a company called Full Story that's fantastic to kind of catch user interactions. So there's some some places where we just found like this software is so strong that there's really no, you know, meaning in like trying to replace it.

But a lot of traditional SAS may have just been a database with, you know, some nice UI and some data and some process logic on top of and that is easier to replicate today. and and it's not easy, but it's easier. And the problem is if it's spread out, it's inconsistent.

And we need it to be consistent for both humans and LM to make sense out of it. Well, this is fantastic. This is super interesting. I feel like uh we should talk regularly about the the situation with white collar workers and just the advance of AI tooling.

There's so many interesting things where you have uh a unique insight because of your position in the industry. Last question from from me before we jump. Do you think uh cookies hurt conversion rates across Europe policies? Yeah. Just the fact just generally on this topic well well generally half an hour. Yeah. Yeah.

Yeah. We could spend half an hour but no but generally like the more clicks you need on a website to get to the eventual purchase you know the worse the conversion rate and uh maybe that's why the American consumer is outperforming just purely technicality.

I think the the challenge of GDPR which is the legislation that led to those that cookie uh sorry [ __ ] show is that um the problem was it it came from a very good principle and idea and that is that we as users should own and control our data not the companies right and if you think about it when when a few years ago WhatsApp there was some rumors that they have updated the terms of use and now there was so much more privacy infringing and everyone should shift Right now, how many people really shifted to Telegram or whatever?

Not enough. Right? Why?

Because at the at the center, the way it should have worked is you should have been able to go to WhatsApp with a click of a button, extract all the data, all your contacts, all your messages, and be able to then feed them into Telegram or, you know, Signal or whatever your choice of preference would have been.

that would have been true customer mobility and that was the purpose of GDPR is to allow that customer mobility but because there wasn't the legislation wasn't implemented thoughtfully enough it it that hasn't materialized because in that world where you had 100% customer mobility the point is if you lost trust in WhatsApp as a provider of your primary chat messages you would have instantly been able to move to something else and the fear of losing all your users in such such a very mobile world where everyone could move would have forced WhatsApp to be concerned about privacy and not just speak that they are concerned about privacy right so like I think that and that was the original intent but then it was implemented with these cookies which actually creates the opposite which is that like we're just tired of accepting cookies I don't want to go to new website because then I have to click again on 55 like the people that the people that wanted these kind of laws had the right intentions but the implementations are horrendous and actually cause the opposite which is less customer mobility.

People less willing to shift services and try different things and funny enough are hence playing into the hands of the same companies that they were annoyed with for infringing on our privacy because as that creates just like the bigger success of these big tech companies.

full customer mobility, the being able to take your data with you and go wherever you want is the key to creating a better competitive internet where uh companies that actually have to compete on creating more value for you as a customer rather than locking you up and locking your data up and then you know creating models that just uh forces you to use the same thing.

Makes a ton of sense. Makes sense. Thank you for joining. It's fantastic. Talk more. Thank you for an amazing podcast. Yeah, we'll talk to you soon. We'll talk soon. Bye, Sebastian. Let me tell you about public. com. Investing for those who take it seriously. Multi-asset investing, industryleading yields.

They're trusted by millions, folks. And they're the sponsor of Aston Martin F1.