提示834：DLocal（DLO）：与Daniel Mahncke和Shawn O’Malley探讨长达十年的多倍增长潜力

不过我觉得，这本质上就是本地业务的机会特性。

Although I'd say that it's sort of just the nature of the business opportunity for the local.

我不认为这能带来什么竞争优势，至少不像佩德罗·阿尔特所想的那样。

I don't see any competitive advantage coming from that, not in the same way that perhaps Pedro Arnt does.

反过来讲，市场机会越有吸引力，按照资本主义的逻辑，就会有越多竞争者涌入。

And if anything, I think the more attractive the market opportunity becomes probably, that's how capitalism work, the more competitors will try to get in.

但别忘了，我们之前提到过的DLocal的规模优势还在。

But then you obviously have the scale benefit of DLocal that we talked about.

所以这至少在我眼里，又回到了尼克·斯利普的规模经济共享模型——DLocal可以把因交易量带来的成本折扣，让利给商户。

So this sort of goes back, at least in my mind, to Nick Sleep and his model of scale economies shared, which is that DLocal can pass on the lower cost to serve that comes from those volume discounts that it gives to its merchants.

这和我们投资组合里的Wise的做法很像，都是通过降低费率来提升用户价值。

So that's a similar thing to what Wise, which is another company that we have in our portfolio, is doing as well, sort of lowering the take rate as part of the value proposition to its users.

如果DLocal能更快地降低自己的运营成本，就能进一步压低商户的收费。

And if DLocal can lower its own input costs faster, then it lowers the merchant's pricing.

这样一来，净费率可能保持稳定甚至提升，而商户表面上看到的费率——也就是标价——却会持续下降。

Well, then the net take rate would stabilize or even expand over time, while the gross take rate, so the sticker price that basically the merchant has to pay, would keep falling.

但问题是，服务成本的降低空间是有底线的，就像TPV增长也有上限一样。

The problem with this, obviously, is that there is a floor to, you know, cost to serve savings just as there is a ceiling for TPV growth.

你既没法再内部压降成本，也没法靠增长来抵消费率的持续下滑。

And you can neither save on costs internally nor outgrow the falling take rate.

这归根结底就是看空的理由。

That's in the end the bearish take.

所以最终还是得看你能提供哪些增值服务，或者说整体的价值提升。

So ultimately it does come down to the value added services you can deliver or at least the value add generally.

在这种情况下，最有可能的就是提升转化率。

And in this case, that's most likely a conversion uplift.

如果你能持续证明，用了你的服务后转化率更高，那对那些通过你支付通道处理数十亿美元交易的商家来说，价值就非常明显了。

If you can sustainably demonstrate that conversion rates are higher with your service, then that is obviously going to be valuable to merchants that are routing billions of dollars through your payment rails.

那DLocal到底提供哪些产品来确保这种转化率提升呢？

So what products does DLocal offer to guarantee this conversion uplift?

其中一个例子就是他们所谓的智能支付网关，或者叫智能本地支付方式。

One example is what it's called smart pigs or smart APMs.

这一点对订阅制商业模式的运作至关重要。

And what that does is pretty vital to how the subscription businesses generally work.

西方的支付基础设施主要依赖信用卡，而像巴西的PIX这样的市场则不同，关键区别在于信用卡可以被保存为所谓的‘令牌’。

So one difference between the payment infrastructure in the West, which is obviously based on credit cards and the payment rails in markets like Brazil, PIX, for example, is that credit cards can be saved as so called tokens.

当你订阅Netflix并提供信用卡信息时，Netflix会保存你的卡信息，每月自动扣款。

So when you subscribe to Netflix and give it your credit card details, then Netflix will store your card and charge it every single month.

当然，它不会保存你的完整卡号，而是生成一个令牌，让系统每月自动向这张卡扣款。

And, obviously, it doesn't save your credit card number, but it creates a token that tells the system to charge this exact credit card every single month.

但像巴西的PIX，甚至印度的UPI这类系统，运作方式完全不同。

And systems like PIX in Brazil or even, you know, UPI in India, they don't work that way.

使用这些系统时，每次付款都需要客户主动发起。

So with those systems, you need the customer to initiate the purchase every single time.

理论上，Netflix没法每月直接从你的银行账户扣款。

So in theory, Netflix couldn't just charge your bank account every single month.

每次扣款前，都得先征得你的同意。

It would need to ask you every single time before it does it.

你能想象，这对用户留存来说简直是灾难。

And as you can imagine, that is pretty bad for customer retention.

每次都要问你：嘿，你真的想继续付Netflix的费用吗？

If every single time you're being asked, hey, do you actually wanna pay Netflix?

而你心里想：我两个月都没看过了。

And you're feeling like, I haven't watched anything in two months.

那算了，不如先暂停，这个月就不付了。

So, nope, maybe I'll just pause and and not pay this month.

DLocal的Smart Picks产品就是一个软件层，它建立在PIX之上，让系统能自动且反复地扣款。

So DLocal's Smart Picks product is a software layer that basically sits on top of picks and gives it the ability to charge a consumer automatically and repeatedly.

就像保存了信用卡那样，用户根本不用每次手动确认交易。

So basically the same way that a stored card would and without, you know, the consumer having to approve each transaction manually.

他们对其他替代支付方式也做了同样的设计，这就是为什么他们叫它智能替代支付，而不仅仅是智能PIX。

And they did the same conceptual thing for other alternative payment methods as well, which is, you know, why they call it smart APMs and not only smart picks.

这正是DLocal为Spotify和Netflix这类公司带来的核心价值之一。

And that's sort of one of the major value adds that the DLocal drives, especially for companies like Spotify and Netflix.

做买卖的基本原则就是尽量减少客户付款的障碍。

Good business 101 is to try and remove friction for customer payments as much as possible.

你绝不希望客户付钱的时候觉得麻烦。

You don't want it to be difficult for your customers to pay you.

能自动每月扣款的订阅业务，和每次收费都要客户确认的业务，其盈利质量天差地别。

And so there is a huge difference in the earnings quality of a subscription business that can automatically charge customers each month versus a business that needs to consult with the customer each time a charge arises.

想象一下，如果你办了健身房会员，每个月他们都打电话问你：'你还想继续付20块吗？'

Can you imagine if you had a gym membership and they called you every month of, Hey, would you wanna keep paying $20, for this membership?

你能确认一下付款吗？

Can you confirm the payment?

那简直太糟了，肯定会有很多人取消，很多人一想：算了，我根本不会去锻炼了。

Would be terrible because you get a lot of cancellations, a lot of people realizing, Yeah, I'm not going to actually work out anymore.

我放弃这个梦想了。

I'm giving up on the dream.

都六个月没去过了。

Haven't been in six months.

所以总的来说，看看DLocal的数据，转化率的提升相当明显。

And so anyways, looking at DLocal's numbers, conversion uplift is pretty significant.

即使使用信用卡，本地化处理也能显著提高转化率。

Even if you use credit cards, local processing increases the conversion rate meaningfully.

如果和国际信用卡交易相比，DLocal的本地服务能带来20个百分点的提升，这幅度非常大。

And if you compare that to international card transactions, DLocal's local service shows a 20 percentage point increase, which is very, very substantial.

另一个提升转化率的产品是DLocal所谓的智能路由系统。

Another product that helps with conversion is DLocal's so called smart routing system.

在任何一个市场，通常都有多种途径可以完成支付处理。

So in any given market, there are usually multiple possible path to basically process a payment.

比如多个本地收单机构，或者DLocal对接的几家本地银行，它们的表现并不都一样。

Several different local acquirers or maybe banks the local is connected to, and not all of them perform equally well.

有的收单机构对某种卡的通过率能达到90%，而另一个可能只有80%。

So one acquirer might approve 90% of a certain car type, while another one only approves 80%.

这种差异其实受很多因素影响。

And that basically varies by, you know, many factors.

可能是一天中的某个时间段。

It could be just the time of the day.

也可能是车型的问题。

It could be the car type.

还可能是交易金额，还有其他几十种因素。

It could be transaction sizes, and, you know, dozens of other factors.

所以智能路由的意思是，DLocal 系统会为每一笔交易动态选择最优路径，以最大程度提高通过率。

So smart routing basically means that DLocal system dynamically picks the best path for each individual transaction to maximize the odds of approval.

即使支付第一次失败了，往往也是可以挽回的原因。

And even when a payment fails on the first try, it's often for, you know, a recoverable reason.

比如银行系统临时出错，或者只是时间上的小问题。

So for example, it was a bank glitch or maybe just a timing issue.

DLocal 的系统会自动判断哪些失败值得重新尝试。

And then DLocal system is sort of built in a way to recognize which failures are worth retrying.

然后调整后续的尝试方式，并自动执行。

So how to modify the next attempts, and then they execute on it.

所以像Affirm这样的公司，如果你熟悉的话，它们在这方面的系统也做得不错。

So Affirm, in case that's a company that you know, basically has a good system for this as well.

而这正是它们在支付领域能保持极高利润率的部分原因。

And it's in part where that company has some of the highest margins in payments.

我从来没想过支付路由还存在这么个问题，得选最优路径。

It never occurred to me that there's such an issue with routing payments to the best path.

从最基础的角度来看，我想问你，为什么支付一开始就会有三成到五成失败呢？

From a first principles perspective, I wanna ask you, why do payments even thirty:fifty fail in the first place?

那为什么换条路径就能改善这种情况呢？

And then why is doing something like changing the pathway help with that?

这背后到底发生了什么？

What's actually going on there?

你可能会以为，主要问题就是跨境交易带来的。

You might guess that the main problem is just a cross border problem.

但实际上，对于普通的境内银行卡支付，失败率只有1%到5%左右。

So for ordinary domestic card payments, only about one to 5% of payments fail.

所以这很合理。

So it's reasonable.

但跨境支付的失败率很容易上升到百分之十五、二十，甚至百分之二十五。

But cross border payment failure rates can easily go up to fifteen, twenty, even 25%.

所以这个问题的主要原因是银行对这些交易的了解非常有限。

So the main cause for the problem is that banks only have limited details on those transactions.

他们只能看到一些基本信息，比如卡片信息、余额、金额、地点这些，根本没法判断商户或者收单方到底靠不靠谱。

So they only see, you know, basic details like card details, balance, amount, location, all those sort of things that don't really tell you if, you know, the merchant, the acquirer, whatever is actually trustworthy.

然后风控系统马上就会介入，毕竟，宁可多拦住一笔，也不能漏掉一笔，这本来是好事。

And then, you know, the fraud detection quickly jumps in because, obviously, it's generally a good thing if you stop one payment too much instead of one too few.

但另一方面，对商户来说，这显然是个大问题。

And but on the other hand, it's obviously a big problem if you're a merchant.

大概有百分之三十五的持卡人在遇到支付被拒后，很可能会放弃这家商户。

So about thirty five percent of cardholders are likely to abandon a merchant after experiencing a decline generally.

不过，像Spotify和Netflix这样的公司，这个数据可能并不完全准确，因为作为消费者，我知道它们是正规公司。

Probably the data is not fully true for, you know, companies like Spotify and Netflix, because I as a consumer know it's a legitimate company.

但这说明一旦支付失败，对商家来说就是个大问题。

But it still shows you that once a payment fails, it's a huge upset for the merchant.

而DLocal的做法是找出问题的原因，然后主动解决。

And what DLocal does is figure out the cause for that problem and then counter it.

假设问题是超时了，

So let's assume the problem was a time out.

DLocal的系统会识别出来，然后一小时后自动重试，完全不用人工干预。

Then DLocal's system recognized that and then tries again in an hour without anyone having to do anything.

如果用来订阅的卡片过期了，DLocal会自动换用你其他地方正在用的新卡，比如你在Spotify上用的那张。

And if maybe the card used for the subscription was old and is no longer active, then DLocal automatically uses the new one, which, for example, you use on your Spotify subscription.

这样系统就知道有张新卡了，接下来就会用这张新卡尝试支付Netflix的费用。

So then it knows there's a new card, and now it will try that one for Netflix.

如果问题是出在海外收单行，比如Spotify用来发送支付请求的银行，DLocal就会改用它合作的本地收单行来处理。

And if a foreign acquirer has the problem, so, you know, for example, Spotify's bank that sends the payment request, then DLocal routes it through a local acquirer instead, which it partners with.

最后一个产品又是另一个先买后付的BNPL产品，对吧？

And the last product is yet another buy now, pay later BNPL product, right?

我觉得到了2026年，任何支付公司都不可能不提供先买后付服务。

I feel like there's no payment company in 2026 that does not offer a BNPL solution.

不过DLocal的模式跟通常的不太一样，因为它自己并不放贷。

Although it seems that DLO calls is very different from the usual ones because they're not a lender in itself.

信贷风险还是由专门的先买后付合作伙伴承担。

The credit risk still sits with specialized BNPL lending partners.

这部分风险不会体现在他们的资产负债表上。

It's not on their balance sheet.

DLocal只负责搭建技术连接，把那些先买后付平台接入商户的支付流程。

DLocal is only responsible for the connected technology that plugs those BNPL lenders into its merchants' checkouts.

没错。

Right.

你可以把它看作一个聚合平台，让商户通过DLocal一个集成接口，就能接入多个新兴市场的先买后付服务商。

So you can basically think of it as an aggregator that gives a merchant access to many BNPL providers across emerging markets, basically via DLocal single existing integration.

为了提供这套技术支撑，DLocal会从这些先买后付合作伙伴的收入中分得一部分。

And to provide that technological overlay, DLocal then takes the share of the revenue that the BNL partners earn.

所以他们就是靠这种中间交易来赚钱的。

So that's sort of how they make money on that intermediate transaction.

这并不是靠利息收入，DLocal 也不用承担任何潜在的信用损失。

So it's not interest income, and DLOCAL does not need to absorb any potential credit losses either.

如果你看看像 Malley 或 New Bank 这样的公司，这些支付平台根本不存在信用风险。

If you think about companies like Malley or also like New Bank, there's no credit risk in this payments company.

那么当你考虑所有这些产品时，有没有哪个是别人没法复制的？

And so when you think about all these products, do any of them stand out as being uncopyable or unclonable?

我的意思是，模仿他们的做法有多难？

I mean, how hard is it to mimic what they do?

我猜答案是否定的。

I assume the answer is no.

说实话，我觉得他们的护城河并不在产品本身或者差异化上。

And to be completely honest, I don't think the moat is in the products or the differentiation, you know.

Smart APM 确实不错，对商家来说非常有价值。

Smart APM is nice, and it's very valuable to merchants.

但在巴西，比如，政府已经在推出一种叫PIX自动支付的功能，这种功能以前没有DLocal是做不到的，现在可以实现定期订阅付款了。

But in Brazil, for example, the government is already rolling out what's called PIX Automatico, which enables recurring subscription payments that weren't possible without DLocal before.

所以，随着时间推移，肯定会有一些创新出现，能解决看空者的担忧，让新兴市场的支付变得更简单。

So, you know, over time, there will be innovation that probably, and that solves the bear case, makes it easier to have payments in emerging markets.

而且，Smart PIX和成功率数据是捆绑在一起的，多少还是有点数据优势，但我觉得这算不上真正的护城河。

And there's still some data advantage to the extent that, you know, Smart Pix is bundled with success rate intelligence, But it's not really a moat, I would say.

它确实能稍微好用一点，但关键是你得想清楚，十年后这种优势到底还能剩多少，又能因此多收多少钱。

Know, it's part of it works slightly better, but you just got to figure out ten years from now how much better is it actually and how much can you charge for that.

在我看来，真正的优势在于规模、监管和市场碎片化。

So the key advantages that I see are scale, regulation, and fragmentation.

DLocal现在覆盖六十多个市场，接入了六百多种本地支付方式，拿到了三十八张监管牌照，还有更多正在申请中。

So DLocal operates in more than 60 markets, has over 600 local payment integrations, and 38 regulatory licenses, and a bit more than it does and still in process.

就像我一开始说的，拿到这些牌照往往要花上好几年时间。

And as I said in the beginning, it can easily take years until you get those licenses.

如果你已经拥有三十八张牌照，还有十二到十五张在申请中，那任何竞争对手想复制这套东西，都得花上很长很长时间。

So if you have 38 of them and, you know, 12 to 15 outstanding, it will take a lot of time until any competitor can copy that.

当你想到智能路由这个功能时，比如，它必须依赖海量数据和大量交易对手才能发挥作用。

So when you think about, you know, the smart routing feature, for example, that only works with tons of data and many counterparties to switch to.

DLocal每年处理的交易量高达35亿笔，覆盖40多个、甚至50多个市场。

So DLocal processes, you know, 3,500,000,000 pay in transactions every single year across, you know, 40 plus, 50 plus markets.

这种优势会随着时间推移不断积累。

And it's an advantage that just compounds over time.

路由越精准，就能赢得更多商户，商户多了，交易量自然就上去了。

You have better routing, and better routing leads to winning more merchants, which leads to more volume.

最终，这会让DLocal掌握更优质的数据，而数据好了，路由又会变得更智能——这个循环你懂的。

Then in the end, that gives DLocal better data, which again, know the drill, improves routing.

你觉得那些资金雄厚的大商户会不会干脆自己开发支付系统，把所有功能都内部化，省下原本付给DLocal的费用？

How do you think about the risk that these big merchants with very deep pockets will just build payment solutions themselves, bring everything in house to save this money they would otherwise be paying to DLocal?

我刚接触DLocal那会儿，确实更担心这个问题。

I think I was more worried about that when I first looked at DLocal some years ago.

但现在我不这么看了，这完全是因为DLocal做得够好。

And I don't look at it fully today because of anything that DLocal did.

从那以后，我为这个节目考察了将近90家公司，大部分都是和你一起看的。

It's mostly that since then, I've looked at, again, close to 90 businesses for this show, most of them with you.

其中有不少公司实际上都是DLocal的客户。

And and a considerable number of them are actually DLocal customers.

这让我明白了一点：机会成本无处不在。

If that taught me anything, it's that opportunity costs are everywhere.

我根本看不到亚马逊或奈飞会去投入资源。

And I just don't see Amazon or Netflix spending resources.

再说，那得花多少钱和时间去搞定全球最分散的支付方式。

And again, that's money and time on figuring out payment methods in the most fragmented part of the world.

佩德罗·阿尔恩特在最近的财报电话会上也说过，随着业务增长，商家往往会和DLocal合作得更紧密。

So, you know, Pedro Arnt actually said in the latest earnings call that merchants tend to work even closer together with DLocal when the business grows.

当然，你也可以从另一个角度看这个问题。

You know, you could look at that differently.

你可能会想，企业越大，支付环节就越关键，自己做反而更合理。

You could say, well, the bigger the business gets, the more important it gets, the more sense it makes for them to bring those things in house.

但实际情况并不是这样。

But that's not what you see.

以前商家来找我们，通常只是想解决一个具体问题，比如帮我在巴西处理收款，怎么把钱从那个国家弄出来；现在呢，商家们已经把新兴市场的支付当成全球南方整体战略的核心了。

So it used to be a merchant coming with, you know, a narrow problem, like, help me fix pigs in Brazil and how I can get money out of that country, to now where merchants are basically treating emerging market payments as a core strategic priority across, you know, the entire global south.

这不只是巴西的问题。

And that's not just Brazil.

也不只是墨西哥。

That's not just Mexico.

还包括DLocal运营的另外五十多个市场。

It's also the other 50 plus markets that, you know, DLocal operates in.

你可能会问，如果十年后只有巴西、墨西哥和阿根廷重要了，那怎么办？

So you could argue that, well, what happens if only Brazil, Mexico, and Argentina matter in ten years' time?

但换个角度看，你怎么能忽视未来二三十年里，会有十个非洲国家变得比现在重要得多呢？

The other side of it is, how do you not know that there are 10 African countries that will be significantly bigger and more important in twenty, thirty years' time than they are today?

我就想到新加坡这样的例子。

I mean, I just think of these stories like like Singapore.

而且很明显，那是完全不同的地区，起点也完全不同。

And obviously, it's a totally different part of the world and totally different starting position.

但他们在短短几十年里取得的成就，如果你在DLocal运营的全球市场中，看到其中两三个地方有类似的发展，长期来看潜力就非常大。

But what they achieved in just a couple of decades, if you see anything like that in just two or three of the markets in the entire world that DLocal operates in, you would have huge potential in the long run.

DLocal合作的这些公司，其实并不希望DLocal变成垄断企业。

So these companies that DLocal works with, they sort of have an incentive for DLocal to not become a monopoly.

这让我想起我们之前做的Copart那一期。

It sort of reminds me of our Copart episode a while back.

当时有一种情况，保险公司即使发现Copart运营得更好，还是会把业务分给它和它的主要竞争对手。

There was this dynamic where insurance companies would split their volumes between Copart and Copart's main competitor, even if Copart was the better operator.

因为它们不想让Copart掌握太大的定价权，变成一家独大。

Because they didn't want Copart to become a monopoly with just too much pricing power.

这个比喻其实特别贴切。

That's a really good analogy, actually.

我觉得主要区别在于，保险公司的客户群体，也就是Copart的服务对象，本身就是一个集中度很高的市场。

I think the major difference that I could think of is that the insurance market, so Copart's customer base, was a pretty consolidated space or is a pretty consolidated space.

所以，不需要太多公司就达成了这种共识。

And so it didn't take many companies to agree on that approach.

客户群越广，这种共识就越难达成。

And the wider the customer base, the more difficult that would be.

所以，未来确实存在这种风险，但我认为和Copart的情况并不完全一样。

So, you know, it is a potential risk in the future, but I don't think it's the same as with Copart, at least not a 100%.

问题不在于客户群体整体，而在于每个客户个人，觉得把业务分散出去能让自己在谈判时更有优势。

So not so much from a perspective of, you know, the customer as a group, but more so that a customer individually feels like he's in a better negotiating position when he can diversify volume.

这正是DLocal面临的难题。

That's sort of the problem for D Local.

而且，DLocal的客户群体高度集中。

And again, D Local's customer base is highly concentrated.

另一个风险是，随着时间推移，我们讨论碎片化这个论点时，它的说服力会越来越弱。

So one other risk that I see is that the fragmentation argument just gets weaker over time when we discuss this.

当只有四个关键市场，而这些市场对Netflix、Spotify和亚马逊都至关重要时，DLocal要想继续用碎片化来解释自己的优势，就会越来越难。

So, know, when there are only four markets, and they are all that matter to Netflix, to Spotify, to Amazon, obviously, it will get harder for DLocal to justify, you know, the the fragmentation argument that they currently have.

所以本质上是在告诉亚马逊：你得接受我们的价格，因为我们提供的不只是巴西、阿根廷和墨西哥，还有其他市场。

And basically telling Amazon, well, you have to pay the prices because we do not only offer Brazil, Argentina, and Mexico, but also for the other markets.

如果亚马逊不看重这一点，他们就不会多付钱。

If Amazon doesn't care about that, they won't pay up.

这种现象我们其实从优步身上也见过，就拿我们投资组合里的公司来说吧。

So we basically know this phenomenon from Uber, you know, to to stay with some analogies on the companies that we own in in our portfolio.

他们运营的前十座城市，贡献了业务的很大一部分。

So, you know, the top 10 cities it operates in are basically responsible for a very significant chunk of the overall business.

如果同样的情况发生在南美最重要的几个市场，那也是我担心的问题。

So, you know, if that happens, again, to the most important markets in South America, that's also a problem that I see for DLocal.

在继续之前，另一个区别是，DLocal 到目前为止采取的策略正好相反。

Maybe before we move on, another difference is that DLocal has taken the opposite approach up until now.

他们没有试图仗着优势去争取最高利润的合同。

So they didn't try to flex their muscle and sort of get the highest margin deals.

他们更看重交易量，优先拉拢大商户，所以之前我才提到，他们的抽成率出现了明显下滑。

They prioritized volume and onboarding these large merchants, which is why I mentioned that before their take rate had this sharp decline.

所以这种规模经济的共享模式，在我看来，也降低了客户因为担心价格被宰而把业务转到别处或自己做的可能性。

So this scale economy's shared model also lowers, in my opinion, the likelihood of customers being afraid of price gorging at any point than either, you know, ship volume somewhere else or do it in house.

那你怎么看那些西方大玩家，比如 Stripe、Adyen，还有 PayPal 之类的？

And so what would be your take on the big Western players, Stripe, Adyen, and also PayPal and so on?

他们现在的竞争力怎么样？

How do they shape up?

目前的情况是，商家同时和 Stripe、Adyen 以及 DLocal 合作。

So the way it currently works is that merchants work with Stripe and Ation as well as DLocal.

西方公司负责西方市场，而 DLocal 则处理所有新兴市场的事务。

So the Western players take the Western market and then DLocal handles everything related to the emerging markets.

归根结底，我还是得说，这种碎片化的问题，我每天都要提十遍。

And in the end, it only comes down to once again, I have to say that, like, 10 times a day fragmentation.

如果这些市场继续保持今天这种高度碎片化的状态，Stripe 或 Adyen 在这些地方投入大量资金就没什么意义了，毕竟他们的本土市场机会多得多。

So if the markets stay as fragmented as they are today, makes very little sense for Stripe or Ation to invest a lot of money there compared to projects in their home markets.

再说，竞争本来就挺激烈的。

I mean, there's a lot of competition.

所以你看，有了DLocal这样的规模型竞争对手，进入新市场就更难了。

So, you know, you have a competitor at scale with DLocal, which never makes it easier to get into a new market.

你还得向投资者解释，为什么因为这个投资周期，利润率在很多年内都会持续偏低。

And you also have to explain to investors why margins will be structurally lower for many, many years because of this investment cycle.

当你是一家利润率远高于普通企业的公司时，投资者就不喜欢看到这种变化。

When you're a company that makes significantly higher margins than, you know, the average business, investors don't like when that changes.

不过DLocal这么做情况就不一样了，毕竟过去五年他们的股价一直被严重打压，我们待会儿可能会聊到为什么会这样。

And, you know, it's sort of different when DLocal does it, because, you know, that stock has been absolutely hammered over the last five years, and we'll probably get into why that happened.

但目前它的市盈率是15倍，而Adyen的市盈率还在25倍左右。

But it's currently trading at 15 times earnings, and Adyen is still a company that's trading at 25 earnings.

所以我的观点是，他们在市场情绪上的损失空间更大。

So my argument here is just that they have a bit more to lose in terms of the market sentiment.

某种程度上，这其实还是机会成本的问题。

So to an extent, I mean, comes back to opportunity costs.

确实是机会成本的问题。

It's opportunity costs, yeah.

而且我觉得，这些公司的整体架构也有关系。

And I think it's also the general setup of these companies.

像Adyen或Stripe这样的公司，完全是自己从头搭建了整个技术体系，特别注重垂直整合，部分原因是因为很多竞争对手经过并购之后，变得越来越老派，垂直整合程度也降低了。

So what companies like Adyen or Stripe have done is they basically built their entire tech stack on their own and focus very heavily on vertical integration, which in part, because many of their competitors have become older and less vertically integrated over time through M and A.

说白了，PayPal就是被他们打败的，因为要打的仗太多，根本不知道该集中精力打哪一场，结果全丢了。

I mean, you know, PayPal lost to them because they had so many battles to fight that they just sort of, you know, lost them all because they didn't know where to focus on.

但Stripe和Adyen能这么做，是因为他们本质上首先是商户收单服务商。

But the reason Stripe and Adyen could do that is that they are, you know, first and foremost merchant acquirers.

所以，支付基础设施早就搭好了。

So, you know, the payment infrastructure is already there.

整个体系都是围绕信用卡网络构建的，但在DLocal的市场里，情况完全不一样。

Everything is, you know, dominated by credit card rails, and that's obviously different in DLocal's markets.

正因如此，DLocal的架构也截然不同。

And because of that, DLocal is also built differently.

它更像一个横向平台——比如Netflix过来接入，重点不是靠自己重新造一套技术，而是通过优化现有资源，把那些地方性的复杂问题给简化掉。

So it's much more of a horizontal player where, you know, Netflix comes around and plugs in, and then it's about getting rid of all of the, you know, complexity, not necessarily through your own tech stack, but by just optimizing, you know, how to use what you have at hand in those markets.

比如巴西的PIX，印度的UPI，还有很多其他可能的支付方式。

And that could be PIX in Brazil, could be UPI in India, many, many other potential methods as well.

所以我想说的是，这完全是另一种价值主张。

So I guess the point being, it's a very different value proposition.

如果让我打个比方，这就像是亚马逊要想在巴西或者主要市场跟Mercado Libre竞争，得砸进去几百亿美元。

And if I can just bring up another analogy here, it's similar to how Amazon would need to invest tens of billions of dollars to compete with Mercado Libre in Brazil or their main markets.

因为电子商务在各地根本不是一回事。

Because, you know, e commerce is not the same everywhere.

消费者的需求也不一样，这可不是亚马逊习惯玩的游戏。

And the customer needs are different, and that's not the game that Amazon is used to playing.

而DLocal和Adyen的故事，其实跟这个情况非常相似。

And it looks like the DLOCAL ADIAN story is actually quite similar to that dynamic.

我之前在录音前查过，Adyen在巴西已经运营了大约十年，比DLocal还要久。

I looked it up before our recording here and it looks like ADIAN has been operating in Brazil for about a decade, which is actually longer than DLocal.

所以可以说，他们只是没在那个市场投足够的钱，也许根本就没把它当重点。

So we could argue that they just haven't invested enough money in the market and maybe it wasn't a priority.

这跟亚马逊在巴西投资不如Mercado Libre的情况很像，就是没下足够本钱去真正争夺市场。

That's sort of the same argument that we've seen with Amazon investing in Brazil relative to Mercado Libre, that they just haven't put in enough money to really prioritize winning.

所以最终，Meli在巴西彻底赢了亚马逊，DLocal也赢了Adyen。

So ultimately, Meli has pretty much won Brazil over Amazon and DLocal won it over Adyen.

这可能只是因为他们更懂当地市场，产品也更贴合当地需求。

And so it might just be their understanding of the market and their product offering being better suited for that market.

这种事我们也不是第一次见了。

It's certainly not the first time we've seen that.

很多出海的公司都有类似的经历，这其实是个普遍现象。

It is a common theme from many of the companies that we've studied that have expanded globally.

不过我们也不该把这个说法太当真，毕竟最终我们还是得靠那些西方商家在DLocal运营的市场里拿份额、扩规模。

We shouldn't double down too much on the narrative though, because in the end we still need those Western merchants to win market share and expand into the markets that DLO is actually operating in.

不只是支付领域的对手，我们还得指望亚马逊继续留在巴西，别彻底被挤出局。

Not just the competitors in the payment space, we also need Amazon to still be in Brazil and not totally get lost there.

总体来说，我对Spotify、Netflix或亚马逊在未来几十年在发展中国家抢走大量市场份额并不担心。

I'm not concerned overall that companies like Spotify or Netflix or Amazon will take a lot of share in developing markets in the next few decades.

但当我想到简化支付这个观点——这某种程度上是DLocal的看空理由时，我也会想到稳定币。

But when I think about simplifying payments, which is to some extent the DLocal bear thesis, then I also think of stablecoins.

那么，你觉得稳定币会对DLocal的业务构成多大威胁？

So to what extent do you feel concerned that stablecoins could be bad for business for DLocal?

事实上，DLocal在今年四月就推出了一款叫‘稳定币全服务’的产品。

Well, DLocal launched a product they call Stablecoins Full in April of this year.

和大多数支付公司一样，现在每家都在做稳定币业务，从入金、出金到结算，样样都涵盖。

So like most other payments companies as well, everybody is a stablecoin product by now, and they offer anything from, you know, on ramp to off ramp to settlement and so on.

对DLocal来说特别有意思的是，全球三分之二的稳定币都集中在新兴市场，人们主要用它来对冲本币贬值的风险。

And what makes this especially interesting for DLocal is that two thirds of all stablecoins are held in emerging markets, basically as a hedge against the local currency falling apart.

单是阿根廷，一年内的稳定币交易额就达到了三千四百亿美元，其中大部分是为了绕过资本管制进行跨境交易。

And Argentina alone did something like $34,000,000,000 of stablecoin transactions in a single year, most of it cross border to get around capital controls.

真正让人担心的是，企业可能会直接用稳定币来结算交易，或者DLocal自己也在用，但这样一来，原本靠高额汇差赚的利润就几乎没了。

And the threat from DLocal is that either companies will use stablecoins themselves to settle transactions, or maybe that DLocal is doing so, but at significantly lower margins with the current high forex spreads basically not existing anymore.

你可能还记得我们在《Remeddling》那一期提到的‘最后一公里’观点：在新兴市场，人们并不会像西方人那样把加密货币当投资品来囤积，尤其是手头有闲钱的时候。

And you might remember my last mile argument from the Remeddling episode we did a while back where, you know, in emerging markets, people don't save or invest money in crypto the same way that people in the West do, where people, you know, invest there, especially if they have spare money that they don't need right now.

在新兴市场，情况很不一样，人们需要这笔钱随时用来支付各种开销。

And in emerging markets, it's pretty different where, you know, people need that money to constantly pay for things.

所以阿根廷的人并不想让USDC余额一直躺在钱包里。

So someone in Argentina doesn't want a USDC balance sitting in a wallet.

他们要的是银行账户里的比索，税已经缴清，合规手续全办妥，再把稳定币换成当地法定货币，最终打到本地支付系统里。

They need pesos in their bank account with the, you know, tax settled, with all of the compliance done, and then converting that stablecoin into local fiat and actually pushing it into the local way.

所以真正需要的，就是这套流程。

So it'll be that.

巴西的PICCs，布宜诺斯艾利斯的银行账户，这些才是DLocal真正派上用场的地方。

PICCs in Brazil, a bank account in Buenos Aires, all of that is what DLOCAL is needed for.

稳定币确实能让结算成本更低，但顶层服务还是少不了。

So stablecoins basically make the settlement cheaper, but you still need the top layer.

把稳定币兑换成客户当地的货币，这一整套操作，本质上只是降低了DLocal处理支付的结算成本。

So everything that actually turns the stablecoin into local currency for the customer, which in turn actually means it just reduces DLocal's cost of settling a payment.

对。

Right.

不过我得说，这只是我的一个推测，我觉得这样解释更合理，但也许十年后回头看，实际情况会不一样。

Although I should say that, this is my working theory and that makes more sense to me, but perhaps we look at it ten years from now and it worked out differently.

我还以为你会站到对立面去打这个赌呢。

I thought maybe you would take the other side of that bet.

好的。

Okay.

在我们聊财务数据、激励机制这些之前，DLocal 还有一个地方跟我们最近分析过的一家公司很像。

Well, before we get to the financials, the incentives and all that kind of stuff, there is one other parallel that DLocal has with one of the companies that we covered just recently.

就像 CASB 一样，DLocal 两年前也遭过做空报告的攻击。

And so just like CASB, DLocal was the target of a short report a couple of years ago.

关于这个，有什么我们需要了解的吗？

So anything we should know about that?

是的，我对做空报告的印象一直不太好，总觉得它们往往抓住一些细枝末节，硬生生编出个故事来赚快钱。

Yeah, I think I might just have the wrong experiences with the short seller reports, because I feel like most of them just try to find the smallest things and then, you know, sort of create this narrative in order to make a quick profit.

不过这次倒是挺意外的，报告一出，股价当天就跌了百分之五十。

This one, I should say, worked out pretty well because the stock dropped 50% in a single day after the report dropped.

不过说实话，我觉得做空在市场运作得当的情况下，确实是个重要的功能。

So to be fair, I think that, you know, generally, shorting is an important function of the market when done right.

这次的做空报告来自一个挺有信誉的机构，就是Muddy Waters。

And in this case, the short report came from a pretty reputable name, which is Muddy Waters.

而且这也不是2022年的事。

And it wasn't 2022.

我们都清楚2022年发生了什么，市场直接崩了。

We all know what happened in 2022 where the market tanked.

那时候有很多公司估值严重虚高。

There were a lot of way overvalued companies.

如果你是做空者，通常都会提前抛出做空理由，所以那会儿市场上到处都是这类报告。

And, you know, if you are a short seller, you mostly pull out a short thesis before, so there were a lot of those flooding the market.

但说实话，到现在回头看，那份报告其实没什么实质内容，甚至可以说根本站不住脚。

And yet, I think it's fair to say that by now, there just wasn't much to it, if anything, actually.

他们当时声称TPV被夸大了，抽成比例高得不切实际。

I mean, the claims were that, you know, TPV was overstated, that the take rate was too high to be realistic.

那时候还是挺高的，不是百分之零点九。

Back then, it was still high, not 0.9%.

而且创始人把公司账户和商户的钱混在一起了。

And the founders, you know, mixed up business account with the merchants' money.

还有，最后一点， insiders 在2021年IPO后锁定期一结束，就抛售了大约十亿美元的股票。

And also, and that's sort of the last point, that insiders sold about $1,000,000,000 in stock right after the lockup period that followed the IPO back in 2021.

所以你总能看出来，有些指控比其他的严重得多。

So you can always see that, you know, some claims are more severe than others.

但我觉得2021年的时候，大多数股票的估值都高得离谱。

But I think back in 2021, again, most stocks, they traded at absurd valuations.

所以，我觉得，当DLocal的股价涨到350倍市盈率的时候，锁定期一结束就卖一些股票，完全说得通。

So, you know, selling some stock after the lockup period ended when DLocal was trading at a multiple of 350 makes a lot of sense if you ask me.

如果我是CEO，我可能也会卖掉一部分股份。

Know, if I was the CEO, I might also sell at least some shares.

但其他那些我们见过的更严重的指控，看起来完全是假的。

But, you know, the other claims that we have seen that are more severe, they seem to be outright false.

我的意思是，DLocal 请了独立的第三方机构对这些关于客户资金的指控进行了审查。

I mean, D Locust bought ran an independent review with outside investigators and on these specific client funds allegations.

审查结果表明，商户的资金和公司资金确实是分开管理的，而且和银行对账单完全吻合。

And the review basically verified that, you know, merchant cash and corporate cash set in separate accounts and matched the bank statements.

还有一件事我得提一下，这简直让人惊讶——当时没有一个商户选择离开。

One thing that I should also add, and that's pretty astonishing, not a single merchant left back then.

股价跌了百分之五十。

The stock dropped 50%.

做空报告都出来了，可还是没有一个商户离开公司。

The short report was just there, not a single merchant left the company.

这表现真不赖。

Wow, that's not bad.

再看费率的变化趋势，我觉得这方面也没啥好担心的。

So looking at the take rate and how that's developed afterward, I guess I wouldn't be too concerned that there's anything wrong with that either.

不过说正经的，你之前提到，Muddy Waters 主要是拿 DLocal 的费率和 Stripe 比，说它高得不合理。

But jokes aside, I think you mentioned that Muddy Waters mainly compared DLocal's take rate to Stripe and therefore argued DLocal's take rate was unreasonably high.

DLocal在阿根廷、尼日利亚和埃及这些市场运营，涉及外汇兑换和分期付款，这种业务模式本身就意味着更高的抽成比例。

With DLocal operating in markets like Argentina, Nigeria, and Egypt with FX conversion and installments layered in, that is just a structurally higher take rate business.

如今三年过去了。

And so after all, the three years have passed.

总交易额从一百亿涨到了四百五十亿。

TPV went from 10,000,000,000 to 45,000,000,000.

他们确实产生了实实在在的现金流。

They do generate very real cash flow.

他们派发股息，还在回购股票。

They pay dividends and they're buying back stock.

所以事情根本没那么可疑，远没有空头报告渲染得那么糟糕。

So things are definitely not really sketchy at all and certainly not as sketchy as the short report would have made it seem.

我得承认，创始人确实犯了一些新手错误，某种程度上招来了这些关注。

I got to admit though that it seemed like the founders did make some beginner's mistakes that sort of, I would say invited this sort of attention.

虽然空头报告内容空洞，但确实还有一起联邦诉讼，指控DLocal没有充分披露阿根廷外汇管制的风险。

While the short report definitely lacked substance, there was also a federal lawsuit regarding DLocal not adequately disclosing its Argentina forex control risk.

看起来，阿根廷业务从一开始就不怎么顺利，尤其是在沟通方式上。

And it appears that the Argentina operation has been somewhat of a mess since the beginning, especially in terms of how they communicated it.

不过，这件事的积极一面是，现在换了新的CEO。

And I would say the positive effect of that is that nowadays, you have a different CEO.

我觉得他们让财务总监佩德罗·阿尔恩接任CEO，这绝不是偶然。

I don't think it's a coincidence that they took a CFO in Pedro Arnt to become the new CEO.

但我真觉得，让佩德罗当CEO，对公司来说是最好的结果。

But I would actually, I like that having Pedro as CEO is I think the best that could have happened to this company.

威廉，听上去咱们这儿有个关键人物啊。

William It sounds like we got a real key man here.

你说佩德罗·阿尔恩是这个投资逻辑的核心？

Pedro Arnt is a vital part of the thesis, you say?

确实是，但我想强调，这不光是因为他以前在Mercado Libre待过。

He is, but I want to mention it's not just because he was at Mercado Libre.

我觉得这确实是好事，对吧？

I do think it is a positive, right?

他看起来就像已经在公司待了二十四年了。

I mean, he looks like he's been there for twenty four years.

他亲眼见证了公司从一家小创业公司成长为拉美最大的企业之一。

So he basically saw it go from a small startup to one of the biggest companies in Latin America.

作为CFO，他在这一过程中发挥了重要作用。

And as the CFO, he played a significant role in that.

他还在牛津读过书，后来在贝恩咨询工作过。

And he also studied at Oxford and and worked for Bain Consulting.

把这些经历都加在一起，他确实像是那种你投资一家公司时希望由他来当CEO的人。

So when you take all of that together, he does seem like the type of guy that you wanna have as a CEO for a company that you're gonna invest in.

尤其是因为他接手这份工作，纯粹是因为认同公司的愿景。

Especially because it was so clear that he only signed up for this because he believed in the vision of the company.

他本可以去拉美地区最顶尖的职位之一，那才是他更轻松的选择。

He could have easily said what was one of the best positions where you can work in Latin America.

他在DLocal公司还持有大约0.8%的股票。

He also owns about 0.8% of the company in stock at DLocal.

刚开始听起来可能不多，但他并不是创始团队成员，三年半前才加入公司。

This might not sound like much in the beginning, but again, he's not part of the founding team and he only joined three and a half years ago.

所以这家公司给员工的激励股权比例非常高。

So generally, the incentive ownership in this company is massive.

创始人和管理团队加起来持有公司大约三分之一的股份。

It's about 33% of the company is owned by the founder and the management team.

也就是说，他们真的把自己的利益和公司绑在了一起。

So there's a lot of skin in the game.

那这个激励机制具体是怎么运作的呢？

And how does that incentive system work?

我们对这方面了解得不多。

Well, we don't know a lot about

至于激励机制，因为DLocal虽然运营总部在乌拉圭，但法律上注册在开曼群岛，这就意味着它不用像美国公司那样公开披露财务信息。

the incentive system because while DLocal is operationally headquartered in Uruguay, it's officially incorporated in The Cayman Islands and that basically means it's legally exempt from, you know, the paid disclosures of a US company or, you know, that a US company usually has to publish.

所以我没法告诉你佩德罗·阿尔特具体拿多少报酬，也不清楚他的激励计划细节。

So I can't really tell you what Pedro Arnt is making, nor exactly what his incentive program looks like.

但我们有的是整个管理团队的综合数字。

But what we have is sort of a blended number for the entire management team.

2024年的时候，这个数字大约是两千万美元，而两年前才五百万左右。

So that was about $20,000,000 back in 2024, which was up from about $5,000,000 just two years earlier.

在股票和期权方面，DLocal 发放三种类型的股权。

And in terms of stock and options, DLocal hands out three kinds of stocks.

也就是期权、限制性股票，还有绩效单位。

So it's options, restricted stock, and then performance units.

至于比例，说实话，看起来并不太吸引人。

And the mix, I gotta say, doesn't look too inspiring.

大部分奖金都是以限制性股票的形式发放，管理团队只要留下来就行，什么都不用做。

I mean, the majority of the bonus is paid in restricted stock, for which the management team doesn't need to do anything beyond just sticking around.

这种做法我们通常都会觉得不太理想。

And it's one of those things that you often point out as something that you don't like to see.

至于绩效挂钩的股票期权，只占整体报酬的一小部分。

And then, know, the performance based stock options, they're just a very small part of the overall payout.

因为我们不清楚每个人的合同细节，所以可能只能假设所有的绩效期权都归佩德罗·阿尔恩特所有。

So since we don't know anyone's individual contract, we still might assume that all of the performance based options sit with Pedro Arnt.

但这显然只是猜测，我觉得可能性极低。

But obviously, there would be speculation, and I think it's highly unlikely.

我想赶紧再回过头聊聊阿根廷的情况。

I do want to quickly get back to Argentina.

你说那边的情况很混乱，具体是指什么？

When you say that things have been messy there, what exactly does that mean?

阿根廷的整体经济环境对DLocal的业务有什么影响？

And how does the macro there impact DLocal's business?

毕竟阿根廷的收入占拉美地区总收入的20%左右。

Because in the end, Argentina is about 20% of LATAM revenue.

这可不是个小数目。

So it is not a small portion.

是的，阿根廷市场确实很复杂。

Yeah, Argentina is a complicated market.

很长一段时间里，阿根廷一直是DLocal最赚钱的市场之一，某种程度上正是由于这种复杂性。

So for a long time, it was one of DLocal's most profitable markets, which to some extent was because of the complexity.

多年来，阿根廷一直有个叫SIPO的政策，说白了就是一种限制或钳制。

And then for many years, Argentina had something called SIPO, which means trap or clamp, if you will translate it.

这本质上是一种关于外汇管制和汇率等方面的监管措施。

And it's basically a regulation about currency controls or exchange rates and all of that sort of stuff.

其中一个难题就是把钱转出国外。

So one difficult thing was basically getting money out of the country.

如果你是一家支付公司，在阿根廷收的是比索，但你得给海外的商家付美元，那你就必须把钱转出去。

And if you're a payments company collecting pesos inside Argentina, but you owe a global merchant dollars outside of Argentina, well, that means you have to get that money out.

正因为难度大，DLocal就能向商家收取高额费用，作为帮他们实现跨境交易的服务费。

And because it was difficult, DLocal could charge high fees to merchants for its service of making it possible to be able to transact internationally.

没错，但这类监管变化太快了，今年可能是优势，明年就变成劣势了。

Right, but due to just the fast pace of change in regulations like this, it could have been an advantage one year and then it's a disadvantage in the other year.

情况变得非常快。

So things change quickly.

如果你看看最近几年，甚至就看每个季度，就会发现阿根廷业务的利润率波动有多剧烈。

And if you look at the last couple of years and even just quarters, you will see just how volatile the margin of the Argentina business has actually been.

你知道，由于MLA逐步推行的政策调整，情况至少在一定程度上稳定下来了。

You know, things have settled, at least to some extent, because of the policy changes that MLA pushed through over time.

他彻底取消了所有外汇管制，这当然不错，但对DLocal来说也有不利的一面，因为以前他们就是靠这个赚钱的。

So he basically got rid of all of the currency controls, which, you know, of course, is good, but it's also to some extent bad for DLocal, because that's where they made their money.

一方面，现在的利润率比以前低多了。

And on one hand, the margin is now lower than it is before.

但另一方面，你也可以说，波动性降低了，政治风险也减少了。

On the other, you could also argue, you know, there's less volatility and also less political risk.

所以我觉得，这既是利润率的压力，也是盈利质量的提升。

So I see it as, you know, a margin headwind, but also quality of earnings upgrade.

所以如果让我总结到现在为止的观点，你买DLocal主要是为了抓住两大顺风因素。

So when I have to summarize the case up until now, you would basically buy DLocal to take advantage of two major tailwinds.

一个是新兴市场，尤其是拉美地区的增长和数字化，另一个是全球科技巨头进入这些市场，扩大份额，更加依赖DLocal。

The growth and digitalization of emerging markets, particularly LatAm, and then just global tech giants expanding into those markets and increasing their market share and and further reliance on DLocal.

DLocal 是参与这一趋势的最佳方式，因为它在这些市场中规模最大、关系最深、技术最强，而且不像西方竞争对手那样高度垂直整合——而后者通常被视为劣势。

And so DLocal is the best way to play that trend because it has the most scale in those markets, the best relationships, the tech, and it's not as vertically integrated as Western competitors, is usually seen as a negative.

但在这个地区，这反而成了优势，因为它让公司更灵活，更能应对各地复杂的监管挑战。

But in this geography, it actually can be a benefit because it helps with a more dynamic company and helps them to juggle the regulatory challenges in those geographies.

没错，你说得非常到位。

Yeah, think that's pretty on point.

我觉得在一段时间内，DLocal 的股价还是会持续波动。

And I think DLocal will remain a highly volatile stock for a while.

归根结底，我最关心的是，有没有什么因素可能真正动摇 DLocal 作为这些大趋势最大受益者的地位。

And in the end, the only thing that matters to me is sort of figuring it out whether anything could reasonably disrupt DLocal status as sort of the main beneficiary of these mega trends.

我确信，随着时间推移，竞争一定会出现，这很正常。

And I'm quite certain there will be competition over time, as there always is.

一些客户可能会把部分业务转移到新竞争者那里，以此来分散风险。

And some customers will probably shift part of their volume to that competition to sort of diversify, you know, the the volume.

虽然这在短期内可能带来波动，但我相信，DLocal 凭借数据优势，以及在这些市场积累的监管经验，依然会占据主导地位。

And while that can create volatility in the short term, I also believe that, you know, the data advantages as well as also the regulatory advantage of, you know, having experience in all of those markets.

我不敢说这是个护城河，但我觉得现在想彻底颠覆他们已经很难了。

I don't wanna say that's a moat generally, but I do think it's too much to completely disrupt them at this point.

我还得提一句，每当市场波动导致股价下跌时，DLocal都会进行战略性股票回购，这一点我相信你和我都懂。

And I should say that whenever volatility causes the stock to decline, DLocal conducts strategic share repurchases, which is something that I know you and I can appreciate.

战略性的股票回购确实总能让我加分，但看起来上一次大规模回购还是在2023和2024年。

Strategic buybacks are are something that always are gonna win some points with me, but it does look like the last meaningful share repurchases happened in in 2023 and 2024.

已经好几年没动静了。

So it has been a few years.

现在有新的回购计划要推出吗？

Is there a new buyback program looming?

在回答之前，我想补充一点：如果我的判断最终被证明是错的，我猜原因更可能是变现能力不足，而不是客户流失。

Well, before I answer this, I want to quickly add that if the thesis eventually turns out to be wrong, my suspicion is that an inability to monetize would be the reason, not necessarily lost volume.

回到资本配置的话题，正如你所说，上一次大规模回购是在2023和2024年，每年大约一亿美元。

But getting back to to capital allocation, as you said, you know, the last meaningful buybacks were in 2023 and 2024, which was about $100,000,000 in each year.

今年早些时候，董事会又批准了新的3亿美元回购计划。

And earlier this year, the board authorized another $300,000,000 buyback program.

我记得三月份的时候，股价还接近30亿美元，现在涨到了40亿美元左右，但即便如此，未来两三年内公司还是能回购7%到8%的股份。

So I think it was in March when the stock still traded closer to $3,000,000,000 And now it's about $4,000,000,000 but that's still, you know, 7% to 8% of the company repurchased in the next two to three years.

考虑到DLocal股价的波动性，如果他们耐心等待合适的时机再出手买入，实际回购比例可能会更高。

And with the volatility of DLocal, it might be more than that if they just wait for the right moments and then buy the stock.

这是其中一方面。

So that's one part.

另外，除了股票回购，你们还能通过分红获得回报，分红占自由现金流的大约30%，股息率在3%到4%之间。

And then also beyond the buybacks, you are being paid through a dividend, which is about 30% of free cash flow, resulting in a yield of about three to 4%.

公平地说，DLocal本质上是个轻资产的业务模式。

To be fair, I mean, DLocal is a is a relatively asset light business.

这意味着他们没什么太多需要再投资的地方。

And what that means is there aren't that many reinvestment opportunities.

他们不需要大量资产就能支撑起现有的运营。

They don't need that many assets to support what they do.

所以，发分红是合情合理的。

So it can make sense to pay a dividend.

不过，如果股价真像看起来这么低，我更希望这部分钱也能用在战略性的股票回购上。

Although I would rather see that being used for strategic buybacks as well if the stock is as cheap as it seems.

我也更倾向回购，但我觉得派发股息的部分原因，可能是为了让创始人获得回报。

I would also favor buybacks, but I guess part of the reason for the dividends is that the founders are getting paid that way.

他们的基本工资其实不高，而且我觉得他们也不打算抛售股票，这其实是个好事。

Mean, they have sort of a low salary and I don't think they plan to sell any stock, which is also a positive.

所以他们也想向投资者传递一个信号：我们对自己产生现金流的能力很有信心，这才选择派发股息。

So they also wanna signal to investors, Hey, we are confident in the ability to generate cash flow, and that's why we're paying a dividend.

而且，我们也想拿到应得的回报。

Plus, we wanna get paid.

那并购方面呢？

What about the M and A front?

他们以前有收购过其他公司吗？

Do they have any history of making acquisitions?

几乎没有。

Not really.

DLocal最近想收购一家叫AZO Finance的公司，也就是AZA，这是一家总部在肯尼亚、专注于非洲市场的跨境支付公司，2024年融资时估值大约1.5亿美元。

Mean, DLocal recently want to buy AZO Finance, AZA, which is a Kenya based and Africa focused cross border payments provider that had been valued at roughly $150,000,000 in a 2024 funding round.

但这笔交易拖得比他们预期久，最后E Local只买下了对方的一项技术资产，花了大约2300万美元。除此之外，这家公司基本没有并购的历史，也看不出有什么并购的打算。

But the deal sort of took longer than they initially planned, and then E Local eventually only acquired one asset or sort of technology from that company for about $23,000,000 So beyond that case, I think there's pretty much no M and A history or even ambition right now.

所以如果要总结的话，这是一个轻资产、高利润的生意，还通过股票回购和分红回报股东。

So if I had to summarize, you have a capital light business, highly profitable, paying back shareholders through buybacks and dividends.

我觉得这已经算是很不错的情况了。

And I think there are worse things than that.

对吧？

Right?

并购通常都被认为会破坏价值，而且这种看法确实有道理。

M and A does, of course, tend to have the reputation and rightfully so that it destroys value.

大多数收购对股东来说都没什么好处。

Most acquisitions have not been good for shareholders.

像Constellation Software和伯克希尔这样的频繁收购者，那是另当别论的。

And serial acquirers like Constellation Software and Berkshire are, of course in a league of their own.

但凡普通科技公司或支付公司搞并购，结果多半都不太好。

But whenever your ordinary tech or payment company is going for M and A, it's probably not gonna work out well.

所以我很高兴DLocal并不想玩这个游戏。

So yeah, I am glad that it's not a game that D Local is keen on playing.

那你怎么看股票薪酬这方面的问题？

But how do you think about things on the stock based compensation front?

他们的股票回购真能显著减少总股本吗？还是说DLocal发的新股太多，把回购的效果全抵消了？

Are the buybacks actually reducing the share count materially or is DLocal issuing so many shares that it basically offsets the buybacks that they're doing?

我可以告诉你，股票薪酬只占营收的0.2%，听起来挺不错，但这么看不对。

So I could tell you that a stock based comp is only 0.2% of revenue, which sounds very good at first, but that's not how you should look at it.

因为在这个行业里，正如我之前说的，毛利比营收重要得多。

Because in this business, as I mentioned before, gross profit is much more important than revenue.

用毛利来衡量股票薪酬，比例其实也不算离谱，大概在5%左右。

And SBC measured against gross profit isn't outrageously high either, but it's about 5%.

所以这绝对不是个小数目。

So it's certainly not immaterial.

过去五年里，股本数量基本保持稳定，但有了新的股票回购计划，这种情况显然会改变。

And in the last five years, the share count has been more or less flat, but that obviously should change now with the new repurchase program.

好的。

All right.

那我们不妨深入看看财务细节，再讨论一下估值吧。

Well, how about we do some digging in the weeds here and looking at the financials and then talking valuation?

投资者应该关注哪些关键指标和数据？你平时会看哪些数字来判断DLocal是否走在正确的轨道上？

What are the metrics and the numbers that matter for DLO that investors should be aware of and that you would look at to keep track of whether DLocal is on the right path?

因为这明显是一个需要长期跟踪的投资逻辑。

Because this does seem to be a long term thesis that we'll wanna be monitoring.

是的，确实如此。

Yeah, mean, it certainly is.

我觉得首先该看的是TPV，也就是总支付交易额。

And I think the first thing to look at is obviously TPV, total payment volume.

去年，TPV大约是400亿美元，同比增长了60%。

Last year, TPV was about $40,000,000,000, so up 60% year over year.

自2019年以来，年复合增长率实际上接近80%。

And since 2019, the CAGR has actually been almost 80%.

就在最近几个季度，增长再次提速，尽管2024年曾有个疲软的季度，但整体趋势是向好的。

And just in the last few quarters, growth has accelerated, again, after a weak quarter in 2024, so you see the right trend going forward.

正如我之前说的，TPV的增长速度总会快于毛利润，主要是因为抽成率的变化——抽成率在下降，所以毛利润的增长慢于TPV的增速。

And as I said earlier, you know, TPV will always grow faster than, for example, gross profit, simply due to the take rate dynamic, where, you know, the take rate declines, so the gross profit grows more slowly than the underlying TPV.

总的来说，DLocal的财务表现，你可以把它想象成一个字母‘V’的形状。

Generally, DLocal's financials, I think you can sort of explain them as as looking like the letter of a v.

它先是上升，然后下降。

So, you know, it goes upwards, and then it goes downwards.

而‘V’形下降的那部分，正是我刚才提到的动态：TPV增长最快，毛利润则慢一些。

And the part of the v that drops is sort of the dynamic that I described where, you know, TPV is growing the fastest, then gross profit is going slower.

但接下来就到了运营杠杆的部分。

But then you get to the operating leverage side of things.

虽然毛利润的增长慢于TPV，但底层的净利润增长却比毛利润更快。

So while gross profit goes slower than TPV, the underlying profits grow faster than gross profits.

这其实是另一侧向上趋势的表现。

That's sort of the the upward trend of the other side of the view.

我不确定这个比喻有没有帮上忙，但至少有些人应该能感同身受。

I don't know if that helped, but maybe, you know, was an analogy that at least some of you guys resonate with.

说到指标，衡量运营杠杆的一个好方法是看EBIT，也就是税前利息前利润，与毛利润的比率。

And I would say, going to the metrics, a good one to track the operating leverage is the ratio of EBIT, so, you know, the earnings before interest and taxes to gross profit.

这个指标反映的是，DLocal每赚一美元毛利润，最终有多少能撑过运营成本，真正落到净利润里。

Because what that measures is basically of every dollar of gross profit that DLocal keeps, how much survives operating expenses and drops actually down toward the bottom.

等当前的投资周期结束，这个指标应该会呈现明显的上升趋势，而且已经开始出现了。

When the current investment cycle ends, this metric should show a clear upward trend, which has already started.

那关于客户群体的指标呢？

What about metrics regarding the customer base?

说实话，一想到DLocal这么依赖少数几个大客户和几个市场，我还是有点担心。

I mean, it would it would still make me somewhat nervous to know that D Local is so dependent on a handful of just its biggest customers and a few different markets.

那在客户多元化方面，有没有什么趋势是你希望关注的？

So are there any trends on that end that you would want to follow with the hope of seeing more diversification?

DLocal公布了两个关于前50大商户的指标，值得重点关注。

DLocal discloses two metrics for its top 50 merchants that are interesting to check-in that regard.

一个是每个商户服务的国家平均数量，另一个是每个商户支持的支付方式平均数量。

And that's the average number of countries served per merchant, and then also the average number of payment methods served per merchant.

过去一年，每个商户服务的国家数量增长了40%以上，支付方式数量则增长了约50%。

So countries per merchant grew by, you know, 40 plus percent year over year, and the payment methods per merchant grew about 50%.

这还谈不上从大商户身上分散风险，但说明这些主要商户随着时间推移，越来越深地融入了我们的生态系统。

So that's not yet diversification away from the top merchants, but it does show that, you know, the top existing merchants integrate deeper into the ecosystem over time.

这也正是为什么我们的净收入留存率能一直保持在这么高的水平。

And, you know, that's also why net revenue retention stays as high as it currently is.

从地域分布来看，你已经能看到一种缓慢但明显的多元化趋势。

And in terms of geography, you already see a slow trend toward more diversification.

比如在2023年，也就是几年前，前三大市场贡献了55%的收入。

I mean, in 2023, for example, a couple years back, the top three markets made up 55% revenue.

现在这个比例已经降到50%了。

Now it's 50%.

在这方面，我觉得还是有轻微改善的。

You know, it's slight improvement, I would say, on that front.

这当然不是因为其他市场在萎缩，而是因为其他市场的增长速度超过了前三名。

And obviously that's not because the other markets are shrinking, but because the other markets are outgrowing the top three.

明白了。

Okay.

现在该说说这家公司的估值了，你打算怎么评估它的价值？它是否应该纳入我们的内在价值投资组合？

I think it's that time where you tell us what the valuation is for this business and how to think about valuing it and whether it should be in addition to our Intrinsic Value Portfolio?

现在我们不再侧重预测收入，而是关注TPV和整体模型中的净费率。

Well, today, instead of forecasting revenue, we care about TPV and the net take rate for the top of our model.

我预计TPV到2028年之前会以38%的速度增长，接近40%，之后的两三年会放缓到20%。

So I have TPV growing at 38%, so close to 40% through 2028 and then decelerating to 20% after that for the next two to three years.

DLocal自己对2026年的预测仍然是60%的TPV增长。

And then DLocal's own guidance for 2026 is still 60% TPV growth.

考虑到过去的增长势头，以及目前整体渗透率还很低，我觉得我的预测已经相当保守了，实际增长可能会更高。

And considering past growth rates and the still low penetration overall, I do think a haircutted here quite a bit, and we could see more growth than that.

不过取费率这一块会持续下降一段时间，虽然我也不觉得奇怪，未来几年甚至几个季度可能会稳住。

And I have the take rate decrease going on for a while, although I wouldn't be surprised if we see a sort of stabilization in the next few years and perhaps even quarters.

我知道你可能会对这一点持更怀疑的态度。

Although I know that you would probably look at that a bit more skeptically.

但我听了一些佩德罗·阿尔特的访谈，感觉他觉得底部快到了。

But some of the calls, interviews that I listened to from Pedro Arnt made it seem like he thinks the bottom should be coming soon.

不过我本人也挺怀疑的，所以还是预计未来五年内取费率会降到略低于0.7的水平。

But I'm a skeptic too, so I still have the take rate decreasing to slightly below 0.7 within the next five years.

这些假设其实延续了我们刚才讨论的V型走势。

And these assumptions are basically a continuation of that V dynamic that we just discussed.

TPV增长最快，但取费率下滑，所以毛利润增长就慢一些。

TPV grows fastest, but the take rate declines and thus gross profit grows slower.

所以我觉得DLO给2026年的毛利润增长指引大概是30%。

And so I see that DLO is guiding for about 30% gross profit growth for 2026.

而你则预计未来毛利润的年复合增长率在19%到20%之间。

And then you're assuming a gross profit CAGR compound annual growth rate of 19 to 20% looking forward.

没错。

Right.

这也就开启了V形曲线的另一端——投资周期差不多结束了，或者至少已经见顶。

And that basically starts the other side of the V to stay with that example, where, you know, the investment cycle is more or less done, or at least it has peaked.

所以我们应该能看到运营杠杆推动EBITDA和EBIT利润率上升，自然也会带来更高的利润，使得净利润的增长速度比毛利润快至少一到两个百分点。

So we should see operating leverage push the EBITDA and also the EBIT margins up, resulting obviously in more profits so that net income is compounding at least a percentage point or two faster than gross profit.

说实话，回报中相当一部分可能还会来自股息和股票回购。

Honestly, a surprisingly part of the return will likely also come from, you know, dividends and the buybacks.

这正是D Local的特别之处。

So that's the sort of special thing about D Local.

它不仅增长迅速，还是个印钞机。

It's not only growing fast, but it's also a cash printing machine.

如果它能维持计划中的30%派息率，并继续以最近宣布的节奏回购股票，未来五年内仅回购金额就可能轻松达到七八亿美元。

And if it keeps the 30% payout ratio, which is planned, I should say, and keeps buying back shares at a similar pace to what has been announced recently, you could easily have, you know, 700 to $800,000,000 of buybacks over just the next five years.

考虑到我预期股价会有较大波动，而D Local的管理层在资本配置方面一直表现出色，过去也证明过这一点，我认为他们会以相当不错的价位完成这些回购。

And given that I expect quite a volatile stock, and DLocal's management is good at allocating capital and has proven that in the past, I think they will execute those buybacks at pretty good prices.

所以我假设每年股价下跌约2.5%，这样算下来每股大约在17到18美元之间。

So I assumed a share decline rate of about 2.5% per year, which would, you know, somewhat be $17 to $18 per share.

如果我们把这些都综合起来，再保守一点，假设退出倍数在个位数出头，这跟你模型里现在的设定差不多，那最终结果会是多少？

So if we take all that together and we try to be conservative by expecting a low teens exit multiple, and that's what I'm seeing in your model right now, where does that leave us?

在这些假设条件下，加上大约20%的安全边际，预期回报率大约为22%，从当前水平算起。

The expected return under these assumptions and with a margin of safety of about 20% is about 22% from today's levels.

这是在基准情景下的情况，我得说明一下。

That's the case in the base case, I should say.

我同时也建了悲观和乐观两种情景，跟往常一样。

So I also model a bear and a bull case as we always do.

我就不在这儿细说了，因为今天讲太多数字可能会有点过载。

I won't go through them here because I think that would just be a bit too many numbers for today.

但我可以告诉你，在悲观情景下，虽然增长依然还算合理，但利润率开始下滑而不是上升，股价可能很快就会比现在跌掉一半。

But I can tell you that in the bear case, when which growth is still quite reasonable, but the margins start declining instead of growing, the stock can quickly halve from today's prices.

那样的话，你就彻底失去V型反弹的势头了。

And so what would happen is you'd basically lose the V shape.

EBITDA和利润率不会像毛利那样增长，反而会像毛利一样面临压力。

Instead of EBITDA and profit margins outgrowing gross profit, they are gonna be pressured just like gross profit is.

所以这种情况意味着，运营杠杆根本不会发挥作用。

And so it's a scenario where basically you don't have any operating leverage kick in.

但说实话，这跟我们现在看到的情况恰恰相反——运营杠杆其实已经在起作用了。

Which to be fair is sort of antithetical because we already see the operating leverage kick in right now.

如果按这种情景来看，那对DLocal的整体估值意味着什么？

And with that scenario, again, does that leave us overall with DLocal?

我前面说过，我自己个人的投资组合里持有DLocal。

As I said earlier, I own DLOCAL in my personal portfolio.

我的平均成本大概在10美元左右，现在股价在14到15美元之间，但我还是觉得它挺便宜的。

And I believe my average cost base there is close to $10 And right now it's trading for about $14 to $15 But I still consider it quite cheap.

因为这只股票波动大，而且我持有一段时间了，我知道市场一旦出现负面消息或恐慌，很可能还会出现不错的买入机会。

So just due to its volatility and from owning it for a while, I know there's a good chance that there will be plenty of opportunities to probably also buy it after some form of bad news or fear in the market.

所以我的打算就是，今天先买一小部分，毕竟手头现金不多，之后再等股价下跌时加仓。

So I guess my idea would be to buy a small position today, also because we lack cash, and then buy more on weakness.

我对手游业务一直有点顾虑，这应该不是什么秘密，但你刚才提到的这些数据确实很有吸引力。

It's probably no secret that I have hesitations about payment businesses, but the metrics you walk through sound very promising.

从新兴市场增长和国际企业抢占新兴市场份额的角度来看，这个故事整体上也很有说服力，这很可能让DLocal的业务需求进一步增加。

I can't see why the overall story is compelling too, terms of emerging markets growth and international companies expanding their market share in emerging markets, which would likely lead to more reliance on DLocal.

既然你在个人投资组合里持有它，丹尼尔，那我猜你对这家公司一定做了大量研究，才会推荐它。

So if you have it in your personal portfolio, Daniel, that I know you've done a ton of homework on the company to be recommending it.

我们之前说过，希望投资组合里持有15到20家公司。

So we've talked about having a goal of owning 15 to 20 companies in the portfolio.

目前我们正好是15家。

At the moment, we have exactly 15.

所以，我并不觉得非得为了分散风险就非要再加更多公司。

So I don't necessarily feel like we have to add more businesses just for the sake of diversification.

但另一方面，我们目前管理的公司数量也没多到连你、我，还有同事凯尔都应付不过来。

But I also don't think that we're at such a large number of companies to keep track of that we can't do so responsibly between me, you, and our colleague, Kyle.

如果我们把DLocal和Wise都设为2%的仓位，我想问你的是，你觉得这样的配置，对那些估值有吸引力的支付类公司来说，够不够？当然，这不考虑我可能有的任何偏见。

And if we make it a 2% position with wise also being a 2% position, my question for you is whether you think that's enough exposure to attractively valued payment companies, ignoring whatever biases I may bring to the table.

我知道你很喜欢这类公司，Kyle也一样。

Because I know you like these businesses a lot and I know Kyle does too.

整个投资组合占4%并不算多。

4% overall portfolio exposure is not a massive amount.

我知道你和Kyle都很看好Wise。

I know you and Kyle, for example, both like wise.

就像我说的，我敢打赌Kyle也会支持DLocal。

And like I said, I'm pretty sure Kyle is gonna be on board with DLocal.

假如没有我的顾虑拖你后腿，你会不会觉得应该再加大仓位？

I guess the question is, if you didn't have my hesitations to hold you back, would you actually argue for making it an even bigger position?

从长远来看，我觉得我们可能会把仓位提高到超过2%。

In the long term, I might see us establishing a larger position than 2%, yes.

但现阶段，我对这个仓位大小很满意。

But I think right now, I feel very good about that sizing.

再说一遍，我自身也带着一些偏见，尤其是因为我的成本较低，加上自己的持仓，可能有锚定效应，这我一直挺困扰的。

Again, I sort of have my own biases going into this, especially because of my lower cost base and my own portfolio, I probably have an anchoring bias, which I sort of struggle with quite often.

所以我觉得现在2%就挺好的。

So yeah, I feel like 2% now.

如果股价在未来某个时候大幅下跌，但公司基本面没变，我觉得我们可以再加仓2%。

And if we actually do see the stock going down significantly at any point without changes in the business, I think we can buy another 2% in that position.

不过我们也要留意对新兴市场的整体敞口，这个比例其实不算小。

But we also wanna be mindful of our exposure to emerging markets, which is not too small.

我们得考虑Mercado Libre、New Bank，还有DLocal这些公司。

We think about Mercado Libre, we think about New Bank and then also DLocal.

所以我觉得这样安排挺合适的。

So yeah, I think I'm pretty fine with that.

在我用约翰·坦普尔顿爵士的一句话为这个周末收尾之前，你还有什么想补充的吗？

Anything you wanna add before I send us into the weekend with a quote by Sir John Templeton?

没有，我觉得不用了。

No, I don't think so.

我们就把它当成一个小仓位吧。

We'll make it a small position.

我会继续做研究，争取彻底搞明白，也许以后再考虑加大仓位。

I'll keep doing homework to try to wrap my head around it and maybe we'll come back and make it a bigger position.

不过，我们就先说到这里吧。

But yeah, we can leave it there.

好的。

All right.

接下来是约翰·坦普尔顿爵士的名言，他说：我从不猜市场是涨还是跌，因为我根本不知道。

Then it's Sir John Templeton's time and he said, I never ask if the market is going to go up or down because I don't know.

而且，其实也不重要。

And besides, it doesn't matter.

我一直在全球各国寻找股票，问自己：哪一只的价格，相对于我认定的价值来说，是最低的？

I search nation after nation for stocks asking, where's the one that is lowest priced in relation to what I believe it's worth?

四十年的经验告诉我，你根本不需要知道市场下一步往哪儿走，照样能赚钱。

So for forty years of experience, have taught me you can make money without ever knowing which way the market is going.

说实话，我觉得挺欣慰的，因为现在我们这个节目，也一直在全球范围内寻找最好的股票，说不定下一个标的，就是个海外公司。

And I actually gotta say I feel pretty good about the fact that by now, we also search nation after nation here on the show to find the best stocks, and perhaps also the next one will be an international one.

我们还不知道，但很快你就会清楚了。

We don't know yet, but you will find out pretty soon.

那到时候见。

So see you then.

感谢收听《投资播客》。

Thanks for listening to TIP.

在你常用的播客平台关注我们，并访问 investorspodcast.com 获取节目笔记和学习资料。

Follow the investors podcast on your favorite podcast app, and visit the investorspodcast.com for show notes and educational resources.

本播客仅用于信息和娱乐目的，不提供财务、投资、税务或法律建议。

This podcast is for informational and entertainment purposes only and does not provide financial, investment, tax, or legal advice.

内容为通用性质，未考虑您的个人目标、财务状况或需求。

The content is impersonal and does not consider your objectives, financial situation, or needs.

投资有风险，包括本金损失的可能性，过往表现不预示未来结果。

Investing involves risk, including possible loss of principal, and past performance is not a guarantee of future results.

听众应在做出任何财务决策前自行研究，并咨询合格的专业人士。

Listeners should do their own research and consult a qualified professional before making any financial decisions.

本节目中的任何内容均不构成购买或出售任何证券或其他金融产品的建议或要约。

Nothing on this show is a recommendation or solicitation to buy or sell any security or other financial product.

主持人、嘉宾以及投资者播客网络可能持有节目中讨论的证券，并可能在任何时候未经通知而调整持仓。

Hosts, guests, and The Investor's Podcast Network may hold positions in securities discussed and may change those positions at any time without notice.

提及任何第三方产品、服务或广告商，并不表示认可，投资者播客网络不对它们的声明负责。

References to any third party products, services, or advertisers do not constitute endorsements, and The Investor's Podcast Network is not responsible for any claims made by them.

版权归属于投资者播客网络。

Copyright by The Investors Podcast Network.

保留所有权利。

All rights reserved.