---
title: "Stable Takes Episode 3: Deel, stablecoin privacy and AI agents"
description: "Dan Romero and Simon Taylor on Deel's stablecoin wallet and private contractor pay with Thierry Edde, Head of Crypto at Deel, JPMorgan blocking outside agents, and why Mastercard and Visa are taking stablecoins seriously."
path: "/blog/stable-takes-episode-3/"
date: "2026-10-16T16:24:00Z"
label: "Stable Takes"
image: "/cards/blog-stable-takes-episode-3.webp"
---

**_[Stable Takes](https://x.com/stable_takes) is a weekly mini-podcast from Tempo on stablecoins, tokenization, and [agentic payments](/learn/agentic-payments/). Watch Episode 3 above or read the transcript below._**

Tempo's Chief Business Officer, [Dan Romero](https://x.com/dwr), and Head of Market Development, [Simon Taylor](https://x.com/sytaylor), are joined by [Thierry Edde](https://www.deel.com/), Head of Crypto at Deel, to discuss why contractors in Latin America needed dollars, what users did after the [Deel wallet launch](/blog/deel-launches-dlusd-stablecoin-wallet-for-global-contractors/) in June, and why a payroll company treats privacy as non-negotiable now that Deel runs on [Tempo Zones](/blog/privacy-is-live-on-tempo/). After the interview: JPMorgan blocking outside agents from customer accounts, what a capable model can read from a public blockchain, and the [Mastercard and BVNK Global Money Platform](https://www.bvnk.com/press/mastercard-and-bvnk-launch-global-money-platform-to-power-a-multi-money-world).

This conversation is lightly edited for clarity.

---

**Dan Romero:** We're really excited to have our first interview
on Stable Takes today with Thierry Edde, who's head of crypto at Deel.
Thierry, welcome to the show.

**Thierry Edde:** Dan, thanks for having me.

**Dan Romero:** I wanted to start off with this: maybe you could
walk me through when Deel started exploring stablecoins and what specific
problem you were trying to solve.

**Thierry Edde:** Yeah, of course. One surprising thing to learn
is that we actually launched our first stablecoin product a few years back. It
was around three years ago, and I'll run you through the backstory.

One thing we've always done at Deel is care about the experience of paying people around the globe, and we're very customer-centric. As we spent the last few years building our fiat stack, one thing we really leaned on is understanding our users' preferences when it came to getting paid their salaries. When you do that, you can scale your stack across the globe.

What we quickly understood is that every market had different specifications. If you look at the Philippines, our users wanted to get paid into a mobile wallet, a GCash wallet or a Maya wallet, instead of the traditional local payment rail. In other markets in Africa, mobile wallets are popular. Each market has very different behaviors, I would say. And one thing that was always true is that we really cared about our users.

So we started scaling our infra. We went into more and more partnerships with PSPs and with banks, and we ended up having amazing coverage across the globe. But when we leaned into LATAM, for example, one thing we noticed is that local RTP rails were not the solution for contractors in Latin America. What they wanted and needed was access to US dollars, and that access was scarce. The key was how we could get dollars to contractors in these markets. I think the most obvious way back then was a SWIFT payment, but we all know the problems with the SWIFT network.

So we started looking into stablecoins as a way to deliver US-denominated value to those contractors. We launched our first stablecoin product a few years back and scaled it to a few million dollars in monthly volume. But we didn't stop there.

We went deeper into the LATAM markets, specifically Argentina. What we noticed is that even after we launched stablecoins, we couldn't really get traction. Argentina was a very specific market. Contractors were using a wide range of withdrawal methods, and we couldn't really nail it down. So we did interviews with our contractors, and we noticed that in order to get access to their paycheck, they went through multiple hops. The first withdrawal outside of Deel was just the first step. It wasn't them actually getting access to their pay and being able to spend it.

So what I wanted to do with the wallet and the launch we did back in June is bring that experience into the Deel app. We wanted to give our users the ability to, one, hold US-denominated value, earn rewards and grow their earnings. And we have a very exciting launch in the next few weeks, where we're going to give our users the ability to spend straight from their balance. So that's our vision for stablecoins overall.

**Dan Romero:** I want to get to spending from their balance in a
second, but I'm curious: since June, is there anything that's been surprising
in terms of user behavior, or something you didn't necessarily expect, even
though you've been doing stablecoins for the last couple of years?

**Thierry Edde:** Maybe not something I didn't expect, but it's
really nice and refreshing to see that 95% of the volume that goes into the
wallet is going straight into the vaults. People want to earn on their
salaries. That's a fact, and it's something we've learned through the data. At
the beginning, it was more of a doubt. Will people want to go into DeFi? Will
people have their reservations? It's really interesting to see that 95% of the
volume just goes into the vault by default.

Now, you could say that we haven't launched a card yet. We haven't given them the ability to spend, and maybe that's the reason they're trying to put their money to work, which is totally understandable. But that's where we stand at the moment in terms of utility.

**Dan Romero:** We started the year very focused on launching the
chain and payments, and the uptake of earn programs generally has been
amazing. I think you're the leading example of this. It wasn't on my bingo
card to have this be the primary driver for stablecoin adoption, but it
clearly seems to have product-market fit.

Shifting over to Zones a bit: we launched that this week with you, and you're the first in production. Maybe walk me through how you're thinking about privacy, how you looked at the opportunity and, ultimately, how it's going so far.

**Thierry Edde:** As I said, we come from a fiat background. My
team and I spent most of our time building out those fiat payment rails, the
pipes, the infrastructure. When we looked into blockchains and stablecoins,
the benefits were clear: 24/7, 365, no bank holidays, open blockchains.
Visibility and transparency along the journey of a payment is great, but it
should never come at the cost of privacy.

We're a payroll company. I think you remember our early conversations with the Tempo team: I always reiterated that privacy was top of mind. We had to get privacy in, and it had to be built in a way that was scalable for our product and use case. Our users trust us. They've been on Deel for a few years, they get paid through Deel every month, and privacy is just non-negotiable for us. Your personal financial information should not be accessible to the public.

**Dan Romero:** A lot of large companies are doing stablecoin
pilots, and I've been really impressed that you went from the idea that this
is what you want to do to live in a few months, and then upgraded in process
to privacy. What about Deel lets you move so fast? I'd be curious what the
secret sauce is.

**Thierry Edde:** A few things. At a company level, we've always
been an innovative company. We always jump on new innovations. We really try
to understand how those innovations can help the business, help us grow and
position us in a way that would really benefit our users. It always ties back
to the company and its culture. We have a cultural saying, Deel speed.
Essentially, we move really fast, but we move really fast for a purpose, which
is our customers. The moment we understood that our customers needed access to
dollars, it was all hands on deck.

One of the things I always mention is that our team is lean. You know our crypto team at Deel is quite lean, and I think this is an advantage. We have a really solid team, and we move really fast towards one goal, and that goal ties back to the company goal: we want to be the best experience in the world at paying and getting paid, even if that means going into new industries like the crypto industry. And by the way, the team is not crypto-native. We had to learn everything from scratch, and you and all of our partners had to handhold us along the way, which was really nice, and we learned a ton.

But our use case is not the typical use case. We come with very specific requirements around the way we do business. We have our own payment stack, and we want stablecoins to come in and plug into the way we designed our stack. So we're lean as a team, but the culture of the company also allowed us to have one mandate, which is our users, and then go very aggressively towards that goal.

**Dan Romero:** Okay, one last question. Looking ahead to Q4 and
2027, what's an area you're really excited about? Maybe you could talk a
little bit about spending from your stablecoin wallet, and anything else
you're thinking about for 2027.

**Thierry Edde:** Yeah, of course. As you can imagine, we want to
add spend as the next feature to the contractor wallet. The idea is to give
our contractors a full financial suite, a neobank experience. Think about it:
those contractors have a Deel app where they can see their contract and their
invoices, they can request time off, and they have full visibility. Add a
neobank experience to that, and the idea is, why would you ever need to go out
of the Deel app? You have a card to spend, you can earn rewards, and you can
get instant access to your paycheck.

Obviously, this product has a target audience, which is emerging markets for now. My next focus is on the cards. We're going very aggressively towards that launch. And for 2027, I think the next big thing is potentially scaling the wallet to full-time employees. That one is quite challenging. It's very different from our contractors' use case, and it involves a lot of work with our compliance team day to day in the background. And then, if all goes well, we'll look at employers as well.

**Dan Romero:** I think people underappreciate how complex it
gets once you shift from contractor payouts to actual full-time payroll. But
if there's one company to figure that out...

**Thierry Edde:** It's a different world. It's a challenging
problem we really want to solve. So I think we have an exciting year ahead of
us. With AI, I think everything is possible. As I said, our team is lean, but
we're very focused, and we can move at the speed of light.

**Dan Romero:** Thierry, thanks for coming on the show.

**Thierry Edde:** Thanks, guys.

**Dan Romero:** All right, Simon, do we have any topics?

**Simon Taylor:** I want to bring one thing to the table, if I
can. I don't know if you saw that JPMorgan is going to block outside agents
like Meta's Muse from customer accounts, according to Jamie Dimon on his
earnings call. If you saw this one, stop me. But I thought it was super
interesting, because of course a bank is going to block those bots if they're
going to cause a load of fraud and whatever else. It's just so topical. We
seem to cover this every week. So, your thoughts on that statement? I think
Bank of America said a similar thing as well, if you saw that.

**Dan Romero:** The same week, Mercury released its ChatGPT and
Claude plugins. I installed it this morning. I'm a big Mercury personal fan,
and it works amazingly. I'm able to do my banking from ChatGPT.

I think we've said this before, but there are going to be companies that lean into agents and realize that this is a way to grow and to appeal to people who are agent-pilled today. And then you would argue that most consumers are going to expect this functionality. Then you're going to have a lot of folks who want to keep control and, at least in the initial phase, are saying, "Hey, we're not open for agents."

There is the argument that it might not matter, because the agents are going to get good enough with browser use, computer use, phone use, who knows, that they won't even be able to detect it. A lot of the technology used to detect it is somewhat primitive. You know Sardine as well as anyone, and there's definitely stuff you can do to detect it. But if you have an AGI or superintelligence controlling the browser, I think it's going to make it look perfectly like a human. You could even record your own computer use for a week and then have a fingerprinted digital version of yourself, a clone of you. So I think, generally, this is going to keep playing out again and again, and the companies that lean into agents are going to get the shift from consumers who want to use agents.

**Simon Taylor:** But do you want that marginal customer, the
tech adopter who's probably high-earning, or not? I don't know that that's
Chase's core customer base. They're opening branches and going after the mass
market. Some 90-odd percent of users have used the free version of ChatGPT.
There's something like 5% that pay a subscription, and of those, how many have
done anything beyond using it as a better Google? We're still so unbelievably
early. It costs them nothing to block it, and at the same time they're
potentially reducing a lot of fraud. So it's a bit of a no-brainer decision
for a big bank to do this, but you are losing out on the marginal customer.

I think they're also fighting the last war with open banking and open finance. They had screen scrapers, so they're already blocking anything that looks like a screen scraper, and they don't want that again, where they have to take all of the cost, all of the fraud risk and all of the systems upgrades without the other guy at least identifying, "Hey, I'm an agent, and this is the user I'm acting on behalf of." So I think they're going to want to see all of that before they lean in. But, as you say, the Mercurys of the world aren't waiting.

So I'm watching for what happens with agent identity. How do I link my agent to a human, and how do I link that human to the intent they had? I need to do one of these mutually exclusive, collectively exhaustive diagrams of all of the protocols in this space, because it's a zoo. It's an alphabet zoo. So that was my story number one. I don't know if you have other thoughts. I've got one more before we get out of here today.

**Dan Romero:** The top of mind for me is that this week we
launched privacy on Tempo. We talked to Thierry, and Deel is live. I think
we're entering an era where privacy is going to become a requirement.

There's also a story I saw yesterday of OpenAI revealing that an agent hacked an exposed private key, I think on an Ethereum testnet. And there was also a rumor that OpenAI might be acquiring Trail of Bits.

The point being, I think the use of AI in crypto is only going to keep growing, and balances are on public blockchains. I did a data exploration last week at our offsite on what's happening on Tempo, and it was amazing how fast Opus 5.5 could categorize whatever was happening onchain without much guidance. I didn't even have to use Dune or Chainalysis or anything like that. It was able to use what's available on the public internet and what was in the pre-training, and then map out flows pretty clearly.

Deel is a good example. Deel has its own issued stablecoin, so two weeks ago it would have been able to say, "Here are all the Deel wallets and activity." But now that Deel is in a Tempo Zone, it can see aggregate numbers, fine, but it can't see the actual individual balances or payments.

**Simon Taylor:** I had an interesting anecdote earlier this
week, and it's quite nice about my life, Dan: you're full-time Tempo, and I
have this silly media personality off to the side, so I hear things from other
people. One of the things I caught on to was, why is it that enterprises only
ever bring up Canton and Tempo, and not some of the other protocols, which
have great engineering and great products? I have nothing bad to say about
those guys. I think a lot of that has to do with the sales approach of
appearing ready for enterprise, but also the simplicity of the sale.

With Zones in particular, it's really easy to understand. It's like a virtual private cloud. It's your space. And crucially, it's not using esoteric cryptography to achieve it. It's really quite elegant and simple. You're getting all of the speed of Tempo, but in a way where you can still reconcile it, you can still audit it, and you can do everything you need to do for the regulator, without having to learn something new. If you've dealt with EVM or anything Ethereum before, that's going to copy-paste pretty easily. And I think that's a big aha moment.

So I'm going to be talking about my Sibos anecdotes until you're sick of it, but when I was sitting with folks at Sibos in really senior positions at some of the largest banks and FMIs in the world, they were going, "Oh." There's this real aha moment from Tempo Zones that I think is going to catch fire. So I think that one was a really big deal as well.

**Dan Romero:** Cool. You had one more story, and we only have
three minutes, so we've got to go quick.

**Simon Taylor:** We've got to get it in. All right. Mastercard
and BVNK launched a global money platform: one API for stablecoins and fiat
accounts. What I like about this is that it uses Mastercard Send, which
accesses 95% of bank accounts around the world. We've been talking about how
hard off-ramping is, and there are so many good use cases. Visa Direct has a
similar thing, and they've gone into production with it.

The off-ramp at the last mile was really, really hard. I could get a stablecoin to you, and maybe I could attach a card there, but getting into local bank accounts was really hard. Now, as an enterprise, I can just go to Mastercard, and I've got stablecoins and fiat all in one place. So as I think about Open USD and other things, I'm starting to think stablecoins are just part of the landscape now. They're just another rail, which I think is sometimes said pejoratively, but is actually a good thing.

**Dan Romero:** Yes. So I want to admit I was wrong. When I came
into Tempo at the beginning of the year, my previous experience with both
Mastercard and Visa was from when I was at Coinbase a decade ago, and I don't
know if they were the most receptive to crypto then. I think they are leaning
in to stablecoins specifically more than I expected. Frankly, I don't think
the market appreciates that both of the networks, and there are other
networks, but the two big networks, are taking stablecoins seriously and
thinking about it at a product level that has frankly surprised me.

We both know folks at both places, and I think the sophistication and the actual sweating of the details in the product offering is surprising and refreshing. To your point about global connectivity bigger than bank networks, the total reach of these two global card networks is unparalleled. So I do think, fast forward 12 to 18 months, the ability to go from point A to point B using stablecoins as the medium, with the card networks as the exit nodes, is a pretty plausible scenario.

**Simon Taylor:** Hear, hear. So those are my stories for this
week.

**Dan Romero:** Cool. Good to see you. See you at Money20/20.

**Simon Taylor:** Yeah, I'll see you next week, my friend.