Shadowing Practice: I Left a $100B Giant to Build dLocal | CEO Pedro Arnt - Learn English Speaking with Video

Choose a practice mode

Each mode opens a full-screen practice room.

604 sentences

If sentences are too short or too long, click Edit to adjust them.

Our guest today is Pedro Art, CEO of D-Local, a NASDAQ-lifted payments processor for emerging markets and Yerba's first ever unicorn.

Before D-Local, Pedro was also CFO of Mercado Libre, Latin America's $100 billion e-commerce giant.

There are few people as experienced in the art and science of scaling technology companies to massive scale across complex emerging markets.

Pedro, we're honored to have you on.

Welcome to the podcast.

Thanks for the invite.

Great to be here, guys.

So let's start in August 2023.

You're leaving the CFO role at a $100 billion company to start at D-Local.

What was it that you saw that made you interested or even excited about this massive bet?

When I started conversations with the founders at D-Local,

I was incredibly intrigued by a combination of things that to me made it one of these, You can't say no to this opportunity.

So I think the first one was one of the things I was thinking a lot about is,

you know, can I leverage what I saw at Mercado Libre, an early stage technology trend across the emerging world,

and somehow find something similar to that at an earlier stage and be able to ride a similar secular wave, right?

And so I thought, you know, payments is a clear example.

We're seeing that in Mercado Pago with the emergence of all the neobanks in LATAM.

But then when I actually started to understand what Delocal was about, I realized there was a whole second leg to this,

which is one of the most difficult things about picking emerging market winners is

that it's very hard to know who will be the next new bank

or the next Melly and who will go by the wayside.

I think death rate is even higher.

But DeLocal was a different take on riding the digital transformation and revolution of the emerging world,

which it's probably the single most precise proxy for how the Magnificent Seven

and most of the world's largest and most successful digital companies are doing across LATAM, across Africa, across the Middle East, across Asia.

Because at the end of the day, our business today grows if the businesses of our large digital global clients grow in these markets.

And so when I realized that there was almost like a double layer of making a bet on emerging market digital transformation,

which was the secular trend itself,

but the secular trend being able to ride the success of the companies you knew were going to be successful, right?

The Microsofts, the Netflix, the Amazons, the Spotify's, the Googles of the world.

and so that was really one of these I've been saying a lot, be careful what you throw out into the universe

because sometimes it throws something back at you that you can't say no to and that was exactly the case

That's so interesting, usually people will group, you know, Dlocal Newbag, Melly into like one group is,

I haven't heard this take where it's kind of a proxy for a bit more than

that I'd like to give some context for the audience though because I know D-local started LATAM, mostly LATAM, but increasingly Africa and Asia.

Something that you said before that actually sort of surprised me is that these emerging markets actually look quite similar.

That's why you're able to expand and do well in all these markets.

Can you help paint the picture for the people that have never lived in these countries?

Like if Netflix or Spotify wanted to expand to Peru and India and or Kenya, what is the thing that breaks?

What does that look like?

I think that is one of the key factors in understanding the potential of this business

because you don't see that quite often where there's some sort of industry structure and set of problems

that are similar enough across the entire global south that

if you've built a winning culture and a winning structure and you've understood key success factors in LATAM

that that actually ports over quite well to success in Africa or success in Asia.

I think that is something somewhat unique of what we've been seeing in financial infrastructure.

And it's a consequence of the state of play across the region, right?

I think it starts with the fact

that many of these regions are still going in transition from cash economies to digital economies.

that that transition and the drive by entrepreneurs,

by regulators to promote financial inclusion has generated quite unique, a lot of innovation across the global south in the fintech space.

If you look at most of the innovation over the last 10 to 15 years, a lot of it or most of it, I would say, has come out of the emerging world.

And so that generates significant fragmentation.

Unlike developed markets where almost everyone uses credit cards and credit card rails,

you see the emergence of digital wallets, real-time networks, mobile money.

That fragmentation means you need to build a company that solves for a fragmented ecosystem.

I think the final leg of this is that the existing pipes tend to be legacy pipes.

Their technology tends to not necessarily work that well, somewhat outdated but you still have to work with those pipes.

And so part of the success factor also is, how do you build that middle layer of technology that somehow serves as a translation almost,

or a middleware that solves for those disparate,

somewhat legacy, somewhat inefficient code bases and technology infrastructure bases that are present,

again, across, you know, whether you look at the Philippines, Nigeria, Ecuador, or you see a very similar situation.

So what emerged as we began to expand into these markets is

that a lot of what had made us successful in LATAM actually was very applicable in Bangladesh,

in Pakistan and in South Africa.

Yeah, that's really fascinating as someone that has not actually been to all these places

that the payment infrastructure could actually look similar that way.

Since you're already like touching on the technology side of it,

Something that really stuck with me is that you mentioned before that the locals note might not actually be primarily technological, but you said it might be operational.

The people on the ground in each market is really interesting.

You know, you're a tech company in a world where a software is really rapidly becoming cheaper and easier to create.

Can you explain your moats and what do people and investors generally underappreciate about it?

Wonderful.

And it's actually a perfect segue from the previous question, right?

Because the question then becomes if you're dealing with fragmented ecosystems, right?

Multiple payment methods, many of those local or alternative.

And remember that not every payment method has the model that the credit cards have, where they've delegated the acquiring to something called an acquirer.

And so you can go and build an acquirer across the markets you want to expand into, and therefore you can own the technology stack.

And you can say, you know what, I'm going to, like a Stripe or an Adyen would typically think of, I'm just going to own the stack and my stack is going to be so much better than everyone else's.

That's my moat, that's my competitive advantage.

But all of a sudden, you're running into these places where you can't own the digital wallet of choice in a market.

You can't own some of these local payment methods and they don't run through acquiring businesses.

You want to offer that digital wallet, you connect into that specific digital wallet, and you live with whatever that digital wallet's feature set is,

stack is, and how their infrastructure works.

And so by definition, success across these fragmented markets, and the digital wallets vary many times from one country to the next,

so you have to solve for that over and over and over again.

What you begin to realize is you can't, which is typically what a tech company will try to do,

code base or build vertically your way into solving this for your clients.

So I always say, whereas the traditional approach of successful tech companies, that's kind of where I came from in the past is we wanted to control everything and build everything.

When you're dealing with this payment reality I just walked you through, you essentially have to build this layer,

which is that translation layer between global merchants, right, on one end, emerging market consumers on the other,

but in the middle you have this really complex ecosystem of payment infrastructure that you can't avoid,

you can't bypass it, you can't go straight from the global merchant to the consumer.

We can talk about the blockchain after and that's interesting,

but the existing infrastructure requires dealing with this mess, right?

And that's why when I say that the moat isn't really technological.

I can't improve the technology stack of the majority of my payment partners.

What I can do is help them improve with feet on the ground,

but most importantly understand what drives performance there and build the relationships,

100

the technology, the models that then translate that into improved performance for each one of my global merchants.

101

And that varies by vertical, by specific merchant use cases, by merchant best practices.

102

So you're constantly having to optimize

103

and think through how do I get this merchant need and use case to best work on this specific payment methods infrastructure.

104

And at the end of the day, a lot of that is not code, it's people.

105

I had a software company in 2021

106

and we then wanted to take payments from Indian customers via Stripe

107

and I was so mind blown by how difficult it was.

108

I was like, I thought we could just plug in Stripe, take payments and there we go.

109

And then obviously I learned about how Indians don't use credit cards at nearly the penetrations of US customers,

110

etc. But then there were so many follow-on problems and regulations and stuff.

111

I was like, oh my god, this is a nightmare.

112

And conclusion was basically that we just, honestly, we just abandoned that market for a while because you're just like, no, I just want an API, I want to connect.

113

And that was 2021.

114

And it was so difficult.

115

And that's exactly what we would be today, right?

116

Here's the API, just connect.

117

My API will be exactly what you're accustomed to.

118

And then on the other end, I'll solve for all of that complexity.

119

So that's exactly the pain point that we arrived to solve yeah

120

so my my my exact feeling back then was i don't care

121

if we have a service provider i don't care if it's

122

if there's like a super dirty manual solution i don't care

123

how it's solved on my end i'm willing to pay more

124

i just want a simple fast code-based solution from my end i want an api

125

and then how they solve it at the end of the

126

day i don't care i'm even happy to pay more to get access to these massive markets

127

but stripe did not provide that and i remember we had

128

so many calls with stripe to just to find out like they didn't have a solution we wanted to

129

that nothing was simple basically but anyways um

130

so i'm glad you did mention stripe and eddie and um

131

so now i'm not the bad guy for bringing them up

132

but where do you where do you believe the local is structurally better positioned than they are.

133

Yeah.

134

And the flow of the conversation is perfect because look at how one thing is leading into the others.

135

I brought them up because I actually, I am a huge admirer of both of those companies, right?

136

But in admiring them and understanding them, I think it also becomes very evident to me

137

when I'm dealing with the day-to-day complexities and pain points and friction of the emerging world financial infrastructure that in a way, and I've just mentioned it, right?

138

Their design is extremely effective and optimized for how you solve the payments and financial infrastructure stack for the developed world.

139

And it has to do with this vertical integration of owning

140

as much of it to make it as elegantly written code and sophisticated code.

141

And most importantly, friendly to merchants and efficient to merchants.

142

That doesn't really port that well.

143

The key success factors, when I look at our company, they're different.

144

It's how do you build a middleware that injects features where features are not present in a payment method.

145

It's about optimizing for very, very variants, for extreme variants around what drives conversion rate in one digital wallet to the other

146

or in a real-time payment method.

147

It's working with stakeholders, whether that be regulators to promote the appropriate regulatory framework, or many times the work,

148

the stakeholder work is with the local acquirer, with the local payment method to help their engineers fix something that isn't working well.

149

And a lot of it, quite frankly, is incident management.

150

Like the reality of emerging market payments is you can try to be as proactive as you want to.

151

You can try to build that middleware that avoids problems.

152

But so much more of what we do is how do you very quickly respond,

153

react, have sufficient redundancy, and then help your partners fix when things do break, right?

154

Your acquirers in the developed world or your payment transfer rails, they don't break that often.

155

That's not the case across most of our market.

156

So when you begin to understand that, you begin to understand that the company design, the company structure, and the things that you're solving for,

157

they're actually quite different to be successful in more developed worlds than in the emerging world.

158

Yeah, I like that.

159

I also think that a lot of people, when they talk about emerging markets, they always talk about the risks and the problems and the challenges.

160

But obviously, for companies like D-Local and people like yourself who like the building, every challenge is obviously an opportunity, right?

161

Because the challenge is the opportunity.

162

Right.

163

So I think the Silicon Valley slang I've been using lately is the volatility

164

and the inherent complexity is not a bug like you're seeing it.

165

It's the feature.

166

That's why we exist, right?

167

That's exactly what we're built to solve for.

168

Were it not for what you're seeing as a bug,

169

then what I'm doing wouldn't really generate the value to merchants that it does.

170

And you see that in our net retention rates.

171

You see that in the kind of consistent growth we've had across our merchant portfolio.

172

Yeah, and even as an investor, it's not the Silicon Valley version of it,

173

but one thing I've recently heard that I completely agree with is that to traditional institutional investors, volatility is risk.

174

Volatility is always bad.

175

But to smart and flexible investors, volatility is opportunity, right?

176

I mean, obviously, I like my favorite stock going down.

177

I want like if I want to enter position I want

178

that stock to go down not up like but yeah anyways um

179

so let's get over the quest you probably like the least

180

because for you it must be very repetitive I think I've heard you talk about this every single earnings call

181

and it's like a boring question but I feel like we have to cover this

182

so skeptics argue that payments are in general so competitive

183

that margins will inevitably shrink in the long run and are always under attack.

184

Can you maybe once again, just for our listeners, explain how you think about margins in the long term and how some effects are,

185

for example, that you actually expand with bigger clients and then they might get like a volume discount.

186

So as you get a massive win to expand globally with a huge client,

187

this could mean that like basically volume discounts are a thing that is obviously can lower margin for a customer, but it's actually a good thing, right?

188

So look, I think it's critical, critical, critical that we do address this issue as many times as necessary, because the more I spend time with investors,

189

I do think this is what they're struggling with the most to underwrite their investment.

190

Now, I sit in a slightly different, if you want more privileged position, right?

191

I get to worry both more long term than a lot of our investors,

192

but more importantly, I don't have a valuation entry point that I need to time, right?

193

I don't need to say, okay, if I wait it out another two quarters until take rate actually bottoms out, then that's when I come in.

194

I need to think of our strategy.

195

And we're extremely convinced that we're managing the whole take rate issue the right way.

196

And let me separate take rate from margin, right?

197

first thing is there's a lot of operational leverage in this business going forward now that we're exiting our investment cycle.

198

And I think you're demonstrating that.

199

That's in the numbers.

200

That's not a claim you're making.

201

That's fact.

202

Correct.

203

Yeah.

204

And there's more to come.

205

So gross profit, revenue, TPV should all be able to grow more than OPEX as we leave the investment cycle further and further back.

206

Now, there is a monetization issue, which is for every dollar we process, we're making less and less sense.

207

But some of that is by strategic design.

208

We're optimizing for TPV growth.

209

We're telling the commercial teams to be both aggressive in the tiering that they offer merchants,

210

so that merchants really drive more traffic to us to gain those volume discounts.

211

But we're also trying to make sure that we're not losing deals on price.

212

And the logic is in large part driven by why I think

213

that this is not a race to zero as the bears will try to position it, right?

214

I'd rather have the merchant relationship be processing his payments,

215

adding value for him, because I trust that there are a couple of things that will begin to change going forward

216

that certainly will allow take rates to bottom, but I can even theorize why they can raise if I have those merchant relationships, right?

217

So let me walk you through some of those.

218

But just to be clear, so the strategy is bring the merchants through the door, build the trust, build the relationship,

219

help them grow their emerging market businesses, even if that takes lower prices now,

220

because that will give you the volume platform to then work on the monetization levers.

221

First of all, I think consolidation eventually will occur in this space.

222

We still have an excessively fragmented fintech world as a consequence of the fintech boom.

223

And so there's a lot of people out there competing solely on price, right?

224

That will consolidate and that will change market structure.

225

And the pay facts who today I've been quite frank about this, we are still by

226

and large price takers we will start becoming price influencers I don't want to say price setters

227

because I think that's too optimistic

228

but it'll no longer be I kind of have to take the price

229

because there's 60 other people throwing just pricing as a competitive

230

differentiator at them the second issue is as a lot of things happen in technology world.

231

So as AI and new technologies become abilities for further differentiation,

232

as we grow our business and scale becomes an increasingly important differentiator as well,

233

and perhaps most importantly as fragmentation continues to occur across the emerging world because of all the innovation that's occurring through AI,

234

through blockchain, through new real-time payment methods, through local credit card schemes that are being launched,

235

that increased fragmentation generates increased value in what we're offering.

236

And then the final frontier is more and more financial infrastructure products.

237

So credit products, KYC products, physical work products, all these things that we've announced we're working on.

238

When you put all of that in the blender,

239

you begin to see a set of levers for potential take rate flattening or even take rate increase

240

as long as we're successful across those different levers and market structure that changes.

241

So all of that is what guides are thinking.

242

Let's build the merchant relationships.

243

Let's maximize TPV growth for now while we work on addressing all of these levers

244

that certainly will put a bottom to the take rate compression.

245

And ideally, you could even see the curve begin to grow again.

246

And if you now compare your argument with numbers that you're delivering as an investor, without looking at them, I would say, OK, you're telling me as an investor,

247

I should trust that this will be fine long term because there will be plenty of opportunities.

248

And if I just look at what's happening, I see the opportunities, right?

249

The opportunities are arguably too many to even chase all of them because you have to prioritize.

250

There's so many opportunities.

251

And the second thing I would say as an investor is, well, if you're not going to lose deals on price,

252

if you're willing to lower your pricing for some strategic customers, okay, but then I do want to see very,

253

very high growth numbers and you are delivering that.

254

So I'm like, okay, I believe you.

255

If your growth would be not that impressive and margins would be going down, I'll be telling you, hey, I mean, one thing has got to go up, right?

256

But you are showing very, very high growth.

257

Spot on.

258

I mean, we need to sustain very consistent high levels of growth.

259

I mean, law of large numbers, at some point, will kick in a little bit, but we need to sustain TPV growth because if not, then you do run into a conundrum.

260

Well, yeah, everything you're saying is nice and wonderful, but there's no gross profit growth, right?

261

And that's simply, we need to avoid that.

262

But again, when you start seeing the algorithm, the algorithm, if we execute, is TPV grows the most,

263

gross profit grows less because of the margin compression, and EBITDA, net income, grow more than gross profit because the operating leverage begins to kick in.

264

In a business, in a financial model, that's actually cash generative.

265

There's no working capital needs, right?

266

And so when you look at that, it's really, regardless of where the valuation lands out,

267

it's an extremely strong financial model that compounds this cash generation over time.

268

And eventually, it's the famous quote around the voting machine eventually turns into a weighing machine, right?

269

I think that kind of cash generation eventually is recognized.

270

That's also the logic behind, and I know you didn't ask this, but sort of the share buybacks.

271

Essentially what we're saying is this financial model generates so much cash

272

that we can invest in the necessary future growth that we need to sustain.

273

We can keep liquidity buffers just in case,

274

and even then there's enough to continue to buy back shares at an increased clip,

275

which will improve our earnings per share and improve our return on equity.

276

So we're really comfortable with the strategy we have.

277

I think now it's just a matter of the whole sector is kind of clearly not right now

278

being viewed very favorably, right?

279

You look at the multiples on everyone and they're not doing well.

280

So we'll grow through that.

281

Yeah, last comment on that, Ali, so then we can move on with the next question.

282

So also one piece of context that I like here is that this is coming from you as the now CEO, but somebody with a very, very strong CFO background.

283

One thing that usually makes me very skeptical is

284

if the CEO paints this rosy picture and then at the same time the CFO sells like 75% of his shares,

285

which is a real example which I just saw like a couple of days ago.

286

The CEO is like, the future is bright and then the CFO just sold like 75% of shares. And then

287

that makes me feel like you really need to have like very high conviction in the vision to say

288

that but if the numbers have such an obvious relationship between growth

289

and you see the operating leverage going like basically everything you're

290

saying what I want to say to the viewers is the

291

listeners is I feel like people who look at the numbers see

292

that you're just making a lot of sense so yeah over to Ali

293

but I feel like you're the numbers for us to believe you.

294

It's not like you have to believe in a visionary.

295

It's like, don't you see the numbers?

296

Look, always look at the numbers first, then listen to the company.

297

I never like investors who say, I don't even bother listening to the companies because you're always sort of saying what you have to say.

298

At the end of the day, you're investing in a management team, whether you like it or not.

299

But I do think that the numbers always have to be there.

300

And I think, yeah, I completely agree with what you just said.

301

Speaker 2 Spoken like some of a CFO background.

302

Pedro, I want to look ahead.

303

I found this quote online actually, a guy called Francisco Alvarez DeMaldi, co-founder of Riverwood Capital, describes you as someone of really tremendous long term vision.

304

So I'm going to go through like crypto and AI in a moment.

305

But is there a structural trend in emerging markets

306

that think most investors are underestimating but would give you a strong conviction about long-term growth?

307

I don't think they're underestimating it to be fair

308

because I'm going to give you the single strongest secular trend that drives D local.

309

I think it's just an issue of timeframes.

310

And again, to Francisco's quote, and I'm flattered by Francisco's quote,

311

the way you make outsized return in tech investing, and I learned this from the guys at Bailey Gifford, is on tenure.

312

The way you really make the huge alpha is you pick your winners and you ride them over the long run.

313

That's why I also like where D-Local is.

314

It's this mid-cap space, more on the smallish side, where so much of the growth and the appreciation of the stock is ahead of us.

315

It's a lot easier to go from $4 billion to $25 billion than it is to go from $25 to $30.

316

There's more risk in the investment, but that's just risk-return adjustment.

317

If you pick the right ones, that's where you make the outsized returns.

318

And the trend here, I don't think the market doesn't get it.

319

I just think the market, sadly, is increasingly more and more short-term, short-term, short-term.

320

Even the so-called long-onlys, you know, when you actually look at their trading patterns, a lot of what they're doing is like three-month flips.

321

But the trend is just look at the demographics of the emerging world.

322

Look at the starting point in terms of continued usage of cash and continued under investment in financial infrastructure.

323

And it's like, this is such an obvious macro trend.

324

Like you have billions of consumers moving into higher levels of disposable income,

325

digitizing their lives and going to move away from cash into payments.

326

And you have companies across the developed world that need to access those consumers.

327

I mean, I always say growth resides in the emerging world over the next 20 years, and that's where the demographics come in.

328

And so when you look at the confluence of those trends,

329

the addressable market that D-Local plays in is the single most important trend behind our investment thesis.

330

Now, the market gets that.

331

The problem is, yeah, that's a five-year view.

332

It's a 10-year view.

333

I wish I could manage my book that way.

334

I really can't.

335

As managers, what I learned in my 26 years, 25 years at Mali is build for the long run.

336

The other stuff, it falls into place.

337

Yeah.

338

As an investor, I hate it so much when I feel like management or especially the CEO are

339

talking about what's currently hot.

340

For example, one of the things I hate the most, and you also come from the startup world,

341

One of the biggest red flags to me is companies that always pivot to the current hot thing.

342

A couple of years ago, they were in the creator economy, then they were crypto startups, now they are of course all AI startups.

343

Everybody and their grandmothers now an AI startup.

344

It's like, come on, man, where's your credibility?

345

And then we know they're going to do something else in two years.

346

And obviously, if you are actually not just an investor, if you have been operational or have been working in a serious,

347

ambitious company before, you know that if somebody changes their long-term vision every five months or six months,

348

that means the goals they set internally are optimized.

349

And it's like if you imagine everybody in the company doing work, it's like a couple of vectors all in a different direction.

350

And yeah, so I think even though it annoys you that maybe the market is a bit short-term thinking,

351

I feel like looking at the stock price of the local since you're optimizing for the long term,

352

this would just mean that when your thesis playing out is visible in the numbers and the margin aspect,

353

for example, is visible and people see that you might be profiting from stable coins and the agentic web,

354

when that becomes visible, I feel like the local stock price and then the share buybacks hit and everything,

355

I feel like they could just show up as like a very aggressive stock buy.

356

Obviously, you can comment on that, but I feel like it could just lead to an even more aggressive re-rating of the stock.

357

Anyway, so let's talk about some of those sexy growth aspects.

358

So for example, latest number I have is like you've been processing 41 billion in payments across emerging markets.

359

And I don't know how to call this.

360

I don't like calling it the crypto community because crypto, to me, I'm very skeptical of a lot of crypto things

361

but let's say stablecoin enthusiasts or bulls say stablecoins would be very very dominant

362

and we see that trend happening i'm not an expert in this topic to be honest

363

but looking at the numbers they're getting a lot of traction

364

right they're already doing like massive numbers i think today you're

365

doing around a little bit less than 10 of your volume is on stablecoin

366

or like kind of place in

367

that world how how do you see this playing out for for d local

368

and just transparently i'm no expert on this

369

but maybe the audience knows more about this i don't think

370

i'm a super expert either i mean we've been spending a lot of time on it

371

and we now have clients and we offer a fairly complete suite of products um

372

so what i can share with you is what we're seeing in the enterprise world.

373

And I think that's a very important distinction because when you look at where cross-border payments were truly broken,

374

I think it's more in the consumer and the SMB side.

375

And they were broken because it was slow and it was extremely expensive.

376

So if you're a consumer and you're doing relatively smallish volumes of trade

377

that you were trying in the past to move money across borders through the traditional banking system,

378

stablecoins are an absolute game changer from a cost and speed perspective.

379

When you move into the enterprise world, it's not exactly that transformative.

380

We're already moving money for our merchants across the globe through fiat very quickly.

381

That's exactly the value proposition of the large enterprise cross-border players.

382

There's friction, but we're solving the friction.

383

So the merchant isn't necessarily seeing it.

384

They already pay wholesale FX prices.

385

These are very sophisticated treasuries.

386

So actually, when you look at it today at larger volumes, FX is not cheaper on stable than it is on fiat.

387

It's actually usually the opposite, right?

388

And so although it's a product that we're trying to push

389

and that we're very excited about because we see the potential of these rails, it's not a segment yet,

390

enterprise stablecoin treasury movements that we're seeing massive adoption in yet.

391

I think it will happen.

392

One of the reasons it will happen is just such a better technology to build on

393

that I think some of the better products and the better stacks that will emerge will be stable native.

394

But I think they'll be successful almost more because they're just better technology than because of anything inherent in, oh, it's stable versus fiat.

395

I also think this whole stablecoin push is making fiat rails finally more efficient.

396

Like this is what they needed to get prodded on more than anything else.

397

At the consumer end, we offer our merchants the ability to be paid in crypto,

398

in stablecoins, or to settle in stablecoins to their end users.

399

I'm still seeing very little interest on that front from global merchants.

400

You see it in very specific corridors or markets where because of capital controls or some sort of geopolitical issue,

401

you have a combination of it's very difficult to move fiat and consumers have become very versed in paying with stable coins.

402

And so that combination maybe makes a market attractive to offer stable coin at checkout.

403

But I think that one's still not here yet.

404

And where we do begin to see the most interest

405

and where most of our business is are the actual stable

406

or crypto or digital asset native business models that sell somehow those services

407

that need the on ramps and the off ramps to and from fiat.

408

No. So, hey, D-Local, I sell USDC,

409

but I need someone to help me be able to collect fiat in Indonesia to give the person the USDC

410

or the USDT.

411

Can you do that ramp for me?

412

And that's really the translation between fiat and the stable world for consumer and SMB use cases.

413

That's where I'm seeing the most short-term potential and the most growth.

414

The one that excites me the most long-term is the first case I mentioned.

415

It's actually business-to-business and treasury movements through the blockchain.

416

I think that one's going to take a little bit longer before it takes off.

417

Okay, very interesting.

418

And for people like me who have only been following this on a very, very surface level, to be honest, quite surprising.

419

So basically, the B2B experience in fiat is kind of solved.

420

So it's just not that relevant right now.

421

Yeah, but in the long term, it will be.

422

Solved is a strong word because, as you know, there's always things that can be improved.

423

And again, the blockchain is such a better underlying infrastructure.

424

But the pain points at the enterprise level are nowhere near as clear

425

and as obvious as they are at the SMB and consumer side.

426

That I would say for sure.

427

Yeah.

428

Yeah.

429

Just in general, just random side note.

430

I'm from Switzerland.

431

People call it the land of banking and finance.

432

And if I, from my bank, I make a payment to, I don't know, a friend in America, costs me 20 francs and takes a couple of days.

433

I don't, why?

434

I don't understand.

435

Why can't I just say 20 francs?

436

Perfect.

437

Perfect.

438

That was exactly my point when I said, I think the other thing that's happening with the emergence of stables, the reason is because they never had to do it any better.

439

Like the technology has clearly been there.

440

If you look at many of my peers and ourselves, right, we move money instantly across the globe.

441

You know, you talk to Errol Wollocks, you talk to, then they're doing the same thing.

442

So the reason is just that they could get away with that.

443

I mean, there's a classic example, I think, of just market dynamics not generating sufficient competition for that to improve.

444

That's going to change very rapidly, whether through stable coin or just,

445

you know, financial institutions getting their act together on whatever systems they choose to move their money around.

446

Because if not, everyone's going to take their business to Revolut, to Wise, to Nubank, to Monzo,

447

to the neobanks that are building much more efficient global money movement corridors using the underlying infrastructure of DeLocal.

448

That's amazing.

449

Yeah.

450

Another topic that I'm personally incredibly excited about.

451

Agents.

452

Let's talk about agents.

453

Do you personally see a world where AI agents are soon doing a majority of the buying and or selling for consumers?

454

Of course, you are thinking about it, but what are you thinking about it and how are you positioning the local?

455

Yeah, we can go on for like six hours just on this topic.

456

Yeah.

457

But a couple of quick reactions.

458

I think it's even more, I was talking to someone about this the other day.

459

I think it's even more specific than 2026, at least for us at the local.

460

It was like you could see it happening.

461

It was towards the end of last year.

462

It was November.

463

It was Opus, what is it, 4.6?

464

That's when you said, okay, now something has just tipped.

465

The end-to-end capacity in the workflows that I'm seeing people build in the company, the jump up.

466

So we have something we call DePlanner and Decoder.

467

So essentially we've built agents that do the product speccing,

468

dPlanner, and Decoder then the product speccing builds the tickets that Decoder then reads.

469

Decoder determines whether they can program that end-to-end and goes ahead

470

and produces the code till PR for a human to do the final review or says,

471

I think I can't handle this one and throws it back into the developer pool.

472

And then the developers will vibe code it, right?

473

But so we're already at roughly 60% of all the code that's getting sent to PRs

474

that are fully end-to-end done by decoder, which I actually think is pretty good because you hear Anthropic saying they're close to 90-ish, 100.

475

If they're there and we're at 60, I think we're doing a really good job.

476

But we saw that just as the models have gotten progressively better, you could just see the whole thing tipping.

477

So I'm a huge believer like you are.

478

Okay.

479

Second question.

480

Agents will replace so many of the tasks that we carry out today.

481

Like that's just a foregone conclusion.

482

If you haven't realized that, then you're in trouble.

483

But I take solace from two things.

484

First of all, we're an infrastructure company.

485

Like we don't have consumers.

486

We don't touch consumer flows.

487

There's no disintermediation of our core business.

488

Right.

489

So the question becomes, does the agentic payment mandate or whatever,

490

does it do away with the need for a payfac or for payments infrastructure?

491

And the more I've spent time on this, the more I'm comfortable that the answer is no. Payfax and payment infrastructure,

492

especially Payfax, are potentially actually a slight winner of all this.

493

Because agents will be better at optimizing which payment method to use.

494

Humans are still sometimes driven by just, I'll just pull out my card, it's what I've always done.

495

The agent will figure out what makes sense, loyalty points, cost, insurance.

496

Hey, this one's better with a local payment method.

497

This one's better with crypto.

498

This one's better.

499

So it's going to push fragmentation.

500

Fragmentation increases the need for delocal.

501

And at the end of the day, even as we work with all these agentic payment mandates that are coming out, whether it's Stripe with OpenAI,

502

whether it's Google, the agent is still bringing the payment mandate based on its understanding of who built the agent.

503

So you as a shopper or as a payer.

504

But then it's still looking for the pipes and the infrastructure.

505

You still need to process the payment.

506

Okay, there is one world where I would say, ah, if this is the direction that agentic takes us,

507

then potentially the middlewares and the middle layers become less relevant.

508

And that is if all consumers migrate to basically instructing their agents, always pay with crypto.

509

Because then all the agent has to do is go and carry out a transaction on the blockchain.

510

And it won't take too long before the Amazons of the world basically say, get rid of all the intermediaries.

511

Just give the agent our address on the blockchain and they can do the transfer.

512

We're not anywhere close to that world.

513

There's still so much value in what non-blockchain payment methods offer, where it be credit cards,

514

real-time payment networks, bank transfers.

515

So you still need the infrastructure for all that.

516

And we just need to make sure that our stack is primed to offer the best kind of conversion

517

when we receive an agentic mandate,

518

and that we're working with all of these agentic protocol companies to onboard as many of our payment methods as possible.

519

Because if the agents can only deal in credit cards, then that's dumb.

520

Like the emerging world doesn't use credit cards.

521

Our role is how to make sure that all of these digital wallets, all of these local payment methods are part of these protocols.

522

So that's on what we're doing, right?

523

We're doing a lot on the cost side, on the building an entirely different company from how we code to how we reconcile.

524

I think that for us, Agentic will be able to drive that operating leverage I was talking about if we execute well.

525

I don't think it disintermediates as well.

526

And then final, sorry, I've given you a long answer, but like I said, we could go on for six hours here.

527

You asked about the shopping experience and payments because that's, remember, e-commerce for us is only a third of the business, right?

528

Specifically e-commerce, then tourism, advertising, we have others.

529

So we've also looked at agentic of saying, okay, where do we think our merchants are at the most risk of being disintermediated also?

530

Because those are the first ones that we should go and pitch to them our agentic capabilities to help them say,

531

hey, if everything, if travel is a good example.

532

So travel, you can make a really strong case that an agent

533

that optimizes what you're looking for is significantly value adding to you.

534

Like I need to get from here to here.

535

These are my loyalty points.

536

This is the price I'm willing to pay.

537

This is the flexibility I have on time.

538

Go and figure it out for me.

539

Like what we take endless hours doing online.

540

Which you won't even have to tell your agent because your agent will have contacts.

541

It will have your name.

542

It will have your wallet or will have access to your local. For sure.

543

It has your calendar, right?

544

So yeah, yeah.

545

Get me there and back to the November conference I have.

546

So those you could say,

547

wow, okay, that one, there's a very utilitarian angle to how we probably shop for travel,

548

not holiday travel, hold on a second, but for travel in general, right?

549

On the pure shopping piece, there's two things there.

550

First of all, e-commerce is more than about just the discovery.

551

It's about the returns, the logistics, right?

552

And that an agent can't do.

553

An agent can't yet figure out an entire logistics network.

554

So I think the e-commerce companies still have a lot of value they can protect from being disintermediated by agentic commerce.

555

And maybe that's the reason also OpenAI switched from saying, I'm just going to try to have everything happen on my

556

shopping agent to the shopping agent can be instrumental in the discovery,

557

but then I'll spit out the traffic to the app or to the e-commerce company because I can monetize that short term,

558

but maybe also because I realized that there's a lot

559

that the consumer still wants to have the relationship with the e-commerce company.

560

And the other thing, this is the last one on this, is when you start talking not to us,

561

because we tend to be very rational, very utility driven, but you start talking to shoppers at large,

562

you begin to realize that shopping isn't always about optimizing.

563

It's actually about the joy of the discovery, of the journey, of browsing through different websites.

564

That angle of the discovery of shopping and the experience of shopping, I wonder what's going to happen.

565

I mean, it's the same defense bricks and mortar retailers make of the physical experience.

566

Now it's funny because now it's the digital guys saying, no, wait a minute.

567

There's a digital experience that the consumer wants.

568

He doesn't want the agent to solve everything for him.

569

So this will happen.

570

Agents will revolutionize everything.

571

But like always in tech land, which verticals, which industries sometimes doesn't always play out as the initial euphoria, I think indicates.

572

Now from a cost perspective, it's just mind blowing how much more efficient our companies are going to become because of what's happening.

573

Pedro, I love that take.

574

I think it's so good, especially the take that in the future payments will become more fragmented, like you would expect consolidation.

575

It's so contrary and so good.

576

I think I'm going to steal it and repeat it to look smarter.

577

Okay, one final question to close.

578

I saw that you said once, one of your greatest fears that keeps you up at night is somewhere in Sao Paulo, there are three kids working in a garage,

579

inventing something that will make the things you're working on obsolete.

580

Last question to close is, have you seen anything recently of that nature that's either scary or exciting?

581

It's the AI native companies, right?

582

I mean, it's the ultimate example of the three kids in the garage.

583

First of all, because lots of them really are, you know, just grew up.

584

For you to have grown up with agents means you're like 18 or 19 if you're starting a company right now.

585

Because if not, simply they weren't around.

586

But just, it's a lot harder to do what we're doing, which is to transform a non-AI native company into an AI first company,

587

because you'll never become AI native.

588

I'm having to deal, as we speak, with all the change management challenges of people realizing,

589

holy crap, like so many of the jobs that exist today at the company will not exist anymore.

590

and what does that mean for the company and that makes for really challenging change management

591

whereas if you're starting it from scratch by definition you're going to move faster

592

so I think for us what's been most humbling

593

and I think what's been most challenging of all this is

594

I'm acutely aware now of how much more difficult it'll be to go from where we are to AI first

595

than everyone else who's simply building it from scratch at zero, right?

596

So I've seen some payment companies, fortunately not direct competitors, but that have like seven employees.

597

And when you look at the TPV per employee or the gross profit per employee, it's just like absolutely off the charts.

598

Pedro, thank you so much for doing this.

599

Thanks, guys.

600

Great questions.

601

This has been a lot of fun.

602

Really like the work you guys do.

603

So, like I always tell my team, keep pushing.

604

We will, we will.

Download App

Everything you need to speak fluently

AI PronunciationScore every sentence

IPA PracticeMaster every sound

VocabularyBuild your word bank

Vocab GameLearn while playing

Scan QR

Download on theApp Store Get it onGoogle Play

FOR YOU

Recommend Video

264 sentences

Daniel Ek: Life After Spotify, Broken Healthcare Incentives, Catching Disease Early & AI's Potential

Daily ConversationWorking Adults

142 sentences

How Americans see EU Tech

Daily ConversationWorking Adults

57 sentences

William Blair’s Louie DiPalma: We estimate SpaceX will spend about $50 billion on AI capex this year

Daily ConversationWorking Adults

304 sentences

The Economics of Owning a Restaurant

Daily ConversationWorking Adults

Vocabulary and speaking notes for this lesson

This video has 604 sentences and 8055 words to shadow. The speech runs for 50:14. The speaker talks at a natural 160 words per minute, close to everyday conversation. 83% of the words are among the 3,000 most common in English; the rest is worth studying before you start.

Key vocabulary in this video

15 less common words from the video, with pronunciation and meaning:

merchant /ˈmɝt͡ʃənt/ (noun) — A person who traffics in commodities for profit.
emerge /ɪˈmɝd͡ʒ/ (verb) — To come into view.
infrastructure /ˈɪnfɹəˌstɹʌkt͡ʃə/ (noun) — An underlying base or foundation for a building, organization, or system.
solve /sɒlv/ (verb) — To find an answer or solution to a problem or question; to work out.
investor /ɪnˈvɛs.tə(ɹ)/ (noun) — A person who invests money in order to make a profit.
stable /ˈsteɪ.bəl/ (noun) — A building, wing or dependency set apart and adapted for lodging and feeding (and training) ungulates, especially horses.
fiat /ˈfaɪæt/ (noun) — An arbitrary or authoritative command or order to do something; an effectual decree.
crypto /ˈkɹɪptoʊ/ (noun) — A secret supporter or follower.
commerce /ˈkɑ.mɚs/ (noun) — The exchange or buying and selling of commodities; especially the exchange of merchandise, on a large scale, between different places or communities; extended…
margin /ˈmɑɹ.d͡ʒɪn/ (noun) — The edge of the paper, typically left blank when printing but sometimes used for annotations etc.
stack /stæk/ (noun) — A large pile of hay, grain, straw, or the like, larger at the bottom than the top, sometimes covered with thatch.
blockchain (noun) — A shared record of past transactions in a distributed network, typically used for cryptocurrency.
stripe /stɹaɪp/ (noun) — A long region of a single colour in a repeating pattern of similar regions.
expand /ɪkˈspænd/ (verb) — To change (something) from a smaller form or size to a larger one; to spread out or lay open.
gross /ɡɹoʊs/ (adjective) — remarkably great, big, vast in an often unpleasant way; (of behaviour) Highly or conspicuously offensive.

Phrasal verbs you will hear

look at (verb) — To observe or watch.
talk about (verb) — Used to draw attention to the speaker's characterization of someone or something.
work on (verb) — To apply mental or physical effort to (something) in order to shape, form, improve, investigate, solve it, etc.
deal with (verb) — To handle verbally or in some form of artistic expression; to address or discuss as a subject.
come from (verb) — To have as one's origin, birthplace or nationality.
carry out /ˈkæ.ɹi aʊt/ (verb) — To hold while moving it out.
come in /ˌkʌm ˈɪn/ (verb) — To enter.
figure out (verb) — To come to understand; to discover or find a solution; to deduce.

Pronunciation to watch

The speaker uses 194 contractions and reduced forms, such as you're, we're, I'm. Say them the short way, as you hear them.

The “th” sounds: thesis /ˈθisɪs/, algorithm /ˈælɡəɹɪðm̩/, bother /ˈbɑðɚ/
The “sh” and “zh” sounds: efficient /ɪˈfɪʃənt/, essentially /ɪˈsɛnʃəli/, translation /tɹænzˈleɪʃən/, innovation /ˌɪn.əˈveɪ.ʃən/, operational /ˌɑ.pəˈɹeɪ.ʃə.nəl/
Long words — get the stress right: infrastructure /ˈɪnfɹəˌstɹʌkt͡ʃə/, importantly /ɪmˈpɔɹ.tənt.li/, essentially /ɪˈsɛnʃəli/, innovation /ˌɪn.əˈveɪ.ʃən/, operational /ˌɑ.pəˈɹeɪ.ʃə.nəl/

How to practise with this video

Listen to the whole video once without speaking and note the words you do not know.
Start at 0.75× speed, shadow it sentence by sentence, then go back to normal speed once it feels easy.
Record yourself and compare with the original, paying attention to words like merchant, emerge, infrastructure.

Grammar in this video

The structures the speaker uses most, with the exact words from the video:

Structure	In the video
Present perfect continuous have/has been + -ing — an action that started earlier and is still going on	I've been saying · we've been seeing · I've been using
Present perfect have/has + past participle — a past action that still matters now	haven't heard · have never lived · you've built
Passive voice be + past participle — the focus is on what happens, not who does it	being launched · is recognized · being viewed
Relative clauses who / which + clause — extra information about a person or thing	issue, which is · people who look · doing, which is

What is the Shadowing Technique?

Shadowing is a science-backed language learning technique originally developed for professional interpreter training and popularized by polyglot Dr. Alexander Arguelles. The method is simple but powerful: you listen to native English audio and immediately repeat it out loud — like a shadow following the speaker with just a 1–2 second delay. Unlike passive listening or grammar drills, shadowing forces your brain and mouth muscles to simultaneously process and reproduce real speech patterns. Research shows it significantly improves pronunciation accuracy, intonation, rhythm, connected speech, listening comprehension, and speaking fluency — making it one of the most effective methods for IELTS Speaking preparation and real-world English communication.

Shadowing technique: read the full step-by-step guide →

Shadowing Practice: I Left a $100B Giant to Build dLocal | CEO Pedro Arnt - Learn English Speaking with Video

Choose a practice mode

Download App

Inside a YC Startup Disrupting a $900B Industry That Runs the US

The Digital Transformation of Banking and Financial Services in a New Era

Why React Native apps make all the money

The Plan for a Unified Nordic Stock Exchange Explained

View From The Top with Roelof Botha, Managing Partner of Sequoia Capital

Billion-Dollar Beauty: A Conversation With Hailey Bieber and Lauren Ratner

I MADE $15B WITH CHAT GPT

I Bet Nobody Has Ever Explained the Backend Ecosystem Like This Before

Daniel Ek: Life After Spotify, Broken Healthcare Incentives, Catching Disease Early & AI's Potential

How Americans see EU Tech

William Blair’s Louie DiPalma: We estimate SpaceX will spend about $50 billion on AI capex this year

The Economics of Owning a Restaurant

Vocabulary and speaking notes for this lesson

Key vocabulary in this video

Phrasal verbs you will hear

Pronunciation to watch

How to practise with this video

Grammar in this video

What is the Shadowing Technique?