---
title: "Stable Takes Episode 1: Open USD, tokenization and AI agents"
description: "Dan Romero and Simon Taylor on Sibos, the Open USD launch, and why agentic bank runs are less likely than banks keeping the checking account and losing the customer."
path: "/blog/stable-takes-episode-1/"
date: "2026-10-02"
label: "Stable Takes"
image: "/cards/blog-stable-takes-episode-1.webp"
---

**_Introducing [Stable Takes](https://x.com/stable_takes), a weekly mini-podcast from Tempo on stablecoins, tokenization, and [agentic payments](/learn/agentic-payments/). Watch the conversation 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), open on [Sibos](https://www.sibos.com/), where tokenization moved from the innovation stage to the opening keynote and the question changed from how to do it to how to get it into production. From there: the [Open USD launch](/blog/ousd-is-live-on-tempo/), why "private issuer" was the term institutions kept reaching for, and what changes when sharing the reserve is the default. They close on [agentic bank runs](https://www.fintechbrainfood.com/p/agentic-bank-run) and the version of the risk Simon thinks is real, which is keeping the checking account and losing the customer.

This conversation is lightly edited for clarity.

---

**Dan Romero:** Good to have you. How was Sibos?

**Simon Taylor:** Sibos was really fun, man. It felt like the
year of tokenization. I've been going to Sibos since 2014, and back then it
was a bit more of a dirty word; it was only in the innovation stages. This
time it was the opening keynote, it was the middle of the show floor, and it
was the conversation you heard over lunch walking past people. So it felt
really different for a long-time Sibos goer.

There's been a real tone shift as well. The [DTCC](https://www.dtcc.com/digital-assets) is about to go live with a good chunk of the top thousand securities in October, and the conversation has shifted from "my God, how do we do this" to "how is this ever going to scale? The stuff we've got is already breaking. How do we get this to production?" That's really powerful. That was the big theme.

I also think the gnarly regulatory issues are still there. We've had a lot come out from the SEC and others, but people are worried about Europe and the UK and cross-border, and there's still a lot of that to get through. But the scale here is enormous. Stablecoins are 400 billion in market cap, JPMorgan's moving three or four trillion a year, Citi's moving two trillion a year, and the DTCC moves quadrillions a year. This is going to be enormous.

What does it look like from afar? Did you see or hear anything? Did anything hit your timeline from Sibos other than me spamming you?

**Dan Romero:** It is a bit of a different world, and stablecoins
and GENIUS are now bringing things together. For all the time I've spent in
crypto, it felt like every three years there was interest from TradFi, and
then prices would go down, or some major scandal would happen, and they'd
disappear for a bit. Post-GENIUS, we both know people are taking this stuff
pretty seriously.

There is a bit of a question of what the use case is for a stablecoin versus a tokenized deposit. You've educated me on this: the gnarly banking regs and capital ratios mean stablecoins are useful for a sliver of things, but for banks, tokenized deposits are really the unlock. Whereas tech companies and global marketplaces don't have to think about that stuff, so for them stablecoins are great. It's literally taking money that would otherwise sit in a fiat bank account and moving it onchain, giving them more flexibility, programmability, and upside from the net interest margin.

But I always try to figure out what the actual catalyst is to get folks from "this is a pilot, or an innovation group," which you know well, to "this is the bread and butter of how we move money every day."

**Simon Taylor:** That's coming now. What was really interesting
is this shift in realization that those tech companies are marquee clients.
You think about a DoorDash, a Deel, these people getting into stablecoins
today, many of them on Tempo. Those are the dream clients if you're a big
bank, and you listen to them when they say they need to do something. They
say, "I want to off-ramp stablecoins." And tokenized deposits are a great way
to off-ramp a stablecoin.

But the other side of it, where they're paying more attention, is the capital markets stuff around 24/7. Nasdaq's going for 24/7; the London Stock Exchange has gone 23/5. There's a push because Hyperliquid is out there, and it's inevitably coming. How these things converge is still being figured out, but that capital markets piece is really where the action is as well.

So it's a dual front. My biggest corporate customers are asking me to do something with stablecoins. I can give the money movers tokenized deposits, and they'll find that really useful. But somebody has to do payouts, and a tokenized deposit doesn't help them, because it's just moving it to a bank account. It's not getting it to the hardest parts to get to.

**Dan Romero:** Maybe this is a good time to transition to the
big news of the week: [Open USD
launched](https://tempo.xyz/blog/ousd-is-live-on-tempo/). You were at Sibos.
Was anyone talking about it?

**Simon Taylor:** Five or six different folks from institutions
told me they really like the idea of it not being a private issuer, and that
it's not necessarily adversarial. The folks at some of the biggest stablecoins
will tell you they're not trying to be adversarial. They'll do a deal with
you, and they want to work with you. But the perception here really matters.
The fact that you can redeem at par, the fact that it's super low fees, and
the fact that you do share in that reserve makes it a bit more egalitarian.

It's taken some of the defensiveness away, from "I have to do tokenized deposits to protect my deposits" to "wait, I can be a reserve backer for this." It's changed the conversation quite dramatically. I don't think that's got all the way out into Sibos yet; it was only the people who live and breathe digital assets who were ready for it. But I heard from people that the fear around stablecoins, this term that kept coming up, was private issuers. Private, whatever. That's the fear. So disarming that, for the institutions, is going to be really big.

You can see it in who's involved. There are 200 companies now, a lot of banks and financial institutions. People will look at the list of companies around Open USD and see something that looks like them. This is the first time anything's been quite that big. What's your take on it?

**Dan Romero:** The thing you mentioned, that the sharing of the
reserves is the default, is powerful. You could make the argument: we have
this stablecoin, it's successful, large market cap, and we're willing to cut a
deal with you. People don't want to have to do that work. So if the standard
version is "I have a lot of customers, I have a lot of distribution, I've
worked hard to build that business over decades, I should participate in as
much of the upside as possible," that resonates.

We can see it in enterprises dipping their toes in the water. A large global marketplace may have more participants and customers than active stablecoin users today from a consumer standpoint. So the idea that you'd then have to negotiate a one-off deal with an issuer who, at the end of that term, could change the deal is a bit of a non-starter.

It's good competition for the market. I don't think it's guaranteed that Open USD will win by default, but it is setting the new standard: if you own distribution, that's valuable, and you should be rewarded for it. That's a shift in the market relative to the last 10 years.

**Simon Taylor:** It shifted the market, and it shifted the
perception of what I can get from it and what I can do with it. It shifted the
perception of stablecoins themselves, from "I have to do a tokenized deposit,
it's my only answer" to "I'm going to do a tokenized deposit because it solves
for a good chunk of my client base, but this stablecoin goes outside the walls
of my institution really easily."

I keep coming back to Henri's example, from Stripe now: a stablecoin is like getting into orbit. It's expensive to get there, but once you're there you can move around really easily, and it's tough to get back. Although every year it's getting cheaper and easier to get in and out of orbit with Starship and everything, and I think that's happening with tokenized deposits.

The one thing I say to every single bank is, if you want to lean into your client demand, use your tokenized deposits as a one-to-one swap for stablecoins and be the off-ramp they want you to be, because you can handle volume as a bank that exchanges in long-tail markets just can't. That brings up a bunch of regulatory questions that change by individual market. This stuff's hard, but as a business case, it makes sense.

**Dan Romero:** The other thing that's interesting about Open USD
and Open Standard generally is in the initial set of founders. You have
Stripe, but also Visa, Mastercard, Coinbase, and Shopify. Everyone's
approaching it from a different perspective, although arguably Visa and
Mastercard are approaching it from roughly the same one. But the credibility
from the two big card networks, at a place like Sibos, is an important flag,
because even with GENIUS there's still a bit of a reputation issue around
stablecoins and crypto. Is this appropriate for a 200-year-old financial
institution? Visa and Mastercard add a lot of credibility from the TradFi
side.

It's also critically important to have Stripe and Coinbase, who are different layers of the internet financial stack. And then Shopify, which outside of Amazon, there's no one bigger in e-commerce. There was a lot of criticism on the timeline when they announced back in June: this is just a bunch of companies putting their name next to it, this is not serious, they're not going to be motivated. It's worth paying attention to who the core founding companies are, and seeing how much they lean into Open USD over the coming months.

**Simon Taylor:** A short anecdote. One bank came up to me and
said, "Simon, you know about tokenization, you're the podcast guy. A big card
network pitched me on Open USD. What's that all about? Is it true?" And I
said, yes, it is. I then heard that exact same story two or three times.
What's interesting is the signal of how seriously those institutions are
taking it. They're already going to their banks and pushing this out there.

You can't be all things to all people all the time, but you can be something with a specific use case. The card schemes with settlement have something. As a failure mode, that's a pretty great failure mode. And then you see the capital markets side that I live and breathe every day. There's going to be a lot coming in the next couple of years. It's got a bright future.

**Dan Romero:** Let's finish on this. You wrote a [piece on
agentic bank runs](https://www.fintechbrainfood.com/p/agentic-bank-run) this
week. Is this imminent? If I'm running a bank, should I be worried about
agents moving my customers off to the latest and greatest high-yield savings
account, or a checking account?

**Simon Taylor:** Do you know why I wrote that piece? Because it
was [Apollo's chief
economist](https://www.apollo.com/wealth/insights-news/insights/daily-spark/is-an-agentic-bank-run-coming)
who set all of that off with a piece saying, "Won't agents just move deposits
around to a higher yield?" It's tempting to believe, because it intuitively
makes sense. It just doesn't fit with experience. With Plaid, with Atomic,
with all of these fintech apps, I've been able to do that for the past five
years. Chime will give you a single button to switch your direct deposit. It's
incredibly hard, because of the perceived switching cost for consumers. Even
if in reality it's super easy, the perceived switching cost is way too high.

I don't think there's a big probability we're going to see a bank run anytime soon. But I do think you can keep the checking account and lose the customer entirely. There was another [study](https://www.americanbanker.com/news/why-ai-models-favor-fintechs-over-traditional-banks) that showed up in American Banker, by AIVO. It says AI models dramatically prefer fintech companies. When you ask for an SMB bank account, Mercury comes up 327 times out of 2,000, then Wise 184 out of 2,000, Relay 68 times out of 2,000. The only bank that showed up anywhere above 10 times was Chase, at 60.

You could argue they're just not very good at AEO or GEO type of stuff, but it's the recommendation, and now it's the ability to open the account, that these things are doing. So the opportunity around anything with friction, like cancelling subscriptions or annual insurance renewals, there's a bit more going on there. But this isn't the bankpocalypse anytime soon, and it wasn't the SaaSpocalypse either. The market loves a narrative, but there is a there there, and banks should really think about what to do.

I'm interested in your take on Amazon blocking versus Shopify leaning in. There are two different modalities of how you play with this stuff.

**Dan Romero:** That is the key thing to think about, and I agree
with you. No one who has a Chase account right now, because Muse or Instinct
has had a couple of good weeks, is instantly switching over and weekly
optimizing for wherever their account is. But to break it down a couple of
different ways:

First, this is going to be a slow change, and a bit of a generational change. In the same way that Robinhood has made a ton of progress with younger folks in terms of becoming their first brokerage account, and now offering a lot of services to be a bank, that's an issue for the Schwabs and the Interactive Brokers of the world, and they've had to react to it. That's just market competition.

Second, and this goes to the Amazon and Shopify point, all companies are going to initially react where they want control. They're used to having control. If the modality is changing for consumers, and it'll naturally happen with young consumers first, they want to interact with services that are agent-friendly. The services that lean in, the services that effectively have MCPs.

And Shopify: you can manage your Shopify store with Muse. This is less on the consumer side, but if you know anyone who has a Shopify store, clicking buttons in the Shopify interface is not the highlight of their day. They like using Shopify because of the great functionality, and it lets them run an e-commerce business, but the admin component is something they'd love to get that time back on.

So you're going to see MCP and agent-friendly banks. Mercury is a great example. Mercury's had an API for a while, and now they have an embedded AI agent, called Command, and an MCP server. The services that are agent-native are going to win over the medium to long term. You won't be able to be a business without being agent-native.

Amazon may be a rare exception, where they're just so large. I think Amazon's competitive advantage is the fast shipping. But even there, there are plenty of things I order on Amazon where I could wait another three or four days if I can now find the equivalent product on a Shopify store. In the near term they have the ability to keep control, but in the long term they're going to lose share. And they're pretty data-driven, so the moment it starts to feel like, we really need to give in here is when they may change, especially due to their advertising business.

The second thing, to your point that this isn't going to happen tomorrow, I agree. That said, in the same way as credit card churning and people who optimize points, I think that culture will grow, because people like free money. If all I have to do is tell Muse: I have a check from the gas company from when I moved across the country ten years ago, and it's literally been sitting in an account. I think that will permeate, and it may be a great way for these agent platforms to show immediate value to people: get you money. Whether that's finding a way to reduce your utility bill or your cell phone bill by arguing with the support agent, or finding money that might be sitting in lost and found. And then they're going to keep working down the stack, to the point of, "It looks like you're using a checking account from blank bank, and we think you could change it over. By the way, I know what town you live in, they still have a local branch here. Two taps and we can move it over. And I can move all your bill pay over."

So that will happen at some point. Is it happening tomorrow? No. But people have to take it seriously, in that it's not a bank run as much as the competitive bar has been raised. If you're going to wish this all away, it's not going away.

**Simon Taylor:** That time-to-magic moment, the first magic
moment for an agent, is how you win. There are a lot of reports of user
drop-off on AI applications generally, and if it can save you money, that's a
great example of a magic moment. Something that feels like, "My God," and it
pulls you in. Therefore, as an institution, you have to give the agent a
reason to choose you. You can't rest on your business model of, "Hey, they'll
never claim that refund," because that's actually good for us. You need to
lean in a little bit more and be mindful of that.

I also think there are a bunch of fraud, cybersecurity, and genuine data leakage issues that need to get ironed out. So my advice to institutions is not "don't block it," but it's also not "don't not block it." It's to find a way in which there are some bits you want to protect and make sure agents can't touch, but also find a way you can learn with real customers.

**Dan Romero:** It's the same thing that stablecoins offer for a
lot of these institutions. Is it going to change your business tomorrow? No.
But taking it seriously, educating yourself about it, and deeply understanding
your own business and where this can plug in, that's the logical move. Whether
you're talking about stablecoins or agents, this stuff is not going away, so
you have to take it seriously.