Banks Are Going Onchain on Weekends

Sy Taylor  0:10  
Welcome to Tokenize, the show focused on stablecoins and the institutional adoption of tokenized real-world assets. My name is Simon Taylor, host of Tokenize for today, and of course, author at FinTech Brain Food and head of market Neva Tempo, and I'm joined by the one and only Kai Sheffield, head of crypto at Visa. How are you doing, Kai? I'm fantastic. We've got a legend, a great friend, mentor. Excited for the show today. Yeah, making a return to tokenize, but a debut on the news show. It's Christine Catalini, who's founder at the MIT Crypto Economics Lab, and you were the co-founder of Lightspot. Christian, always good to have you back. How are you doing, sir?

Christian Catalini  0:46  
I'm doing great. I mean, two of my friends are on the show here, so I'm looking forward to this conversation.

Sy Taylor  0:51  
Let's do this. Before we jump in, I got to remind viewers and listeners that views and opinions of our contributors today are their own and might not reflect those of companies they represent. Please don't take anything we say as tax, legal, or financial advice, and to remind you that this podcast is sponsored by our friends at Modern Treasury, who I had dinner with on Tuesday evening. Shout out to Dmitri and the guys. This episode is sponsored by Modern Treasury. Stablecoins are here, and so are checks and ACH-they're not going away. RTP and Fed now and new types of payments rails are emerging all the time. The challenge now is integrating all of these without slowing down. Modern Treasury offers one single API for fiat and stablecoins, helping teams launch payments products in days, enter new markets, and serve more customers. Trusted by companies like Procore and Navan and Morse, and backed by over $600 billion in payments history, learn how to adapt to changing payments rails with scale and confidence at moderntreasury.com. All right. Thanks, sponsors, and big, big news this week to lead off with. Open USD is live on Tempo, Ethereum, Solana, and Base. This is the Open Standard organization that announced OUSD in June and switched it on yesterday on those chains. And uniquely, businesses can mint and burn stablecoin, the stablecoin OUSD, at one to one, so at par and at low cost through Coinbase, Stripe, Visa, and Mastercard, and of course Bridge, the Stripe company. Bridge issues the token and the reserves stay at BlackRock, Bernie, and LeadBank. So, Christian, you've seen stablecoins come and go. You've seen attempts to do tokenized money come and go numerous times. What are your thoughts on all things stablecoins, and how important do you think the economic model is here?

Christian Catalini  3:12  
So, look, as I wrote when Open USD was announced, I think it's a really important effort to some extent. There's two potential scenario for all this, right? Either one currency, one asset becomes kind of the the interop layer for the digital economy, or we're gonna have massive fragmentation across stablecoins. It's been amazing to see that you know the group has brought together not only for the most prominent names in the space, but also like a longer tail of other participants, I do think that it would fill a big gap in just making interoperability really seamless between wallets, fintechs, you know, even traditional financial institutions. The big question mark is going to be, you know, what does Zach do now that is essentially you know leading this effort and building the team in really ensuring what is a very difficult and delicate? Not only governance. I'm less concerned about you know the governance structure. I think you know the group has been tough on that. But like with all of these efforts historically, it's like how do you get momentum? How do you convince even your own founding partners to actually go? There's always a little bit of a waiting game where everybody's waiting for somebody else to make the assets actually truly meaningful and useful, it looks like partners have committed liquidity, which I think is a great first step. The next step is like you know there's some really interesting use cases that a more global and interoperable stablecoin could serve. How do you bring those to life? I mean the proof is really in the usability and application pudding. I think that's going to be the test. Of course, this stripe DNA brings a lot of practical focus on things that this could do for merchants. And there was a post right by Zach, I think, a few weeks ago, on what are the key things that we want to solve for. I know we have Kai here, so you know, I'll be remiss not to surface that at some point, as we've seen back. The day with Libra, you bring in together members that are gonna have conflicting incentives on how this technology may shape. And so, how does the Open USD team navigate that? How do you get momentum so that people can really see this as a solution? Personally, I think it's strictly superior to you know everybody using stablecoin from a single issuer. It's a more powerful approach to that problem, but yeah, eventually the market votes.

Cuy Sheffield  5:25  
Yeah, huge shout out to the the open standard team, and they've they've just been cooking and moving incredibly quickly. And it's interesting when you think about like the concept of a consortium and and bringing together a number of large entities with distribution and and scale. The first challenge that a lot of people go to is like, well, how are they going to agree on anything, and are they going to be aligned enough, and then are they going to be able to move quickly enough? And we've seen consortium approaches in the stablecoin tokenized deposit space that have been announced years ago and still haven't gone live yet, and so I think that the biggest thing was really just this concerted effort of we've got to ship a product, and having under Zach's leadership and a great team there at Open Standard that that's still kind of being built out, being able to move quickly and get this product off the ground. I think that's the first step in demonstrating that you can really have an operating model where you move quickly and you iterate with a product while still having the backing and the support and the distribution of major entities. And then there's something really cool to me about this balance of cooperation and competition. And I think that that kind of echoes some of the DeHock principles and the early days of Visa, and so I like the idea that you can get OUSD at Visa through VSP. You can get it through BB&K with Mastercard. You can get it through Stripe. You can get it through Coinbase, and so there are many different places that you can go. It's not a single entity, a single kind of place that you have to be, and then you know you can cooperate around the standards and how to issue it. But then we expect to have vigorous competition among these members around the use cases and the value add services and the products around it. And I think that that's a really good market structure. That it's going to take time to play out, but I think it's off to a really good start. Yeah, and Kai,

Christian Catalini  7:20  
really building on on your comment there, right? So I do think the original Visa is exactly the the right way to think about this. That also surfaces an interesting tension, which is to some extent, there's two futures for Open USD, two successful futures, but there are different degrees of intensity. One is it becomes kind of a Linux foundation for open standards around money movement, and the asset is kind of the the forcing function, right? For standards on merchant payments, on the reverses payment, all sort of interesting applications where some partners will have conflicting views, but somehow Open USD manages to find the superset that everybody can agree on, and then you compete in the market, so that's one scenario. The other one is probably the most ambitious one, which is it becomes like a visa-like entity that many years after it's its own very profitable business, its own network. Of course, you can see that I could imagine some founding members would prefer one or the other, depending on where they sit in the stack. And and I think that's the ultimate tension that at some point will materialize. But maybe it doesn't matter, right? So maybe if you bring use cases to market early on, you get adoption, and then the thing solves itself. But those are two very different visions for where you could go. Like, does it become a super profitable network of networks, or does it become kind of a piece of standard infrastructure that everybody really trusts and relies on is super neutral, but it's never meant to make too much money to some extent.

Sy Taylor  8:47  
Yeah, I think that's interesting because it's a question for sort of 20 years' time. We don't know if that's going to be even a consideration until this thing is successful. Job number one is is get the thing off the ground. It was however many years, like 2030, years before Visa even considered that. Just I've been at Cybus in Miami all weekend, dialing in from my phone in a hotel room, and it's been so interesting to me to hear financial institutions talk to me with absolute excitement around Open USD because I think a lot of the oh we don't like stablecoins. Was because of private issuers, and I think private issuers have done really well. You know, Circle and Tether will continue, I think, to play a massive role in the ecosystem, and they've made huge strides for innovation, and they'll continue to do so. But there are still some very large financial institutions that see that as adversarial. The feature of Open USD, I think, is this neutral entity that is doing everything it can to give away as much of the economics as possible, and incentivizing as many participants as possible in the reserve sharing, and really changing those economics. And one thing that I think gets lost is yes. No mint burn fees. People are excited by that, but the most excited feedback I got was redeeming a pot, one to one redemption for stablecoins and USD. Like I don't want to be thinking about having to maintain some hedging against the tiny movements of a stablecoin against regular dollars. $1 is $1. It should work like that, and so to get that feedback has been really interesting. And I've also had other folks go, "Oh, we could extend this idea. What about O Euro and O G B P and O security and so on? So a lot of excitement, but also super super early days. Just checking O USD dot FYI. I think there's 477 million of supply so far. I think there are bigger commitments behind it and more to come. But you know, shout out to the guys that have delivered all of this. I know the team at Tempo is hugely excited by this and what it can unlock for big capital markets players. And certainly, I can say that I've been in conversations with some of the largest capital markets folks again in the last few days, but also the last year or so, and this has been a huge unlock for them. They they see the opportunity for something that, unlike a deposit, can actually go outside the walls of any one individual bank, but unlike a private stablecoin, feels somewhat credibly neutral. So interested in your reflections on that, Paul.

Christian Catalini  11:24  
No, I think I think that's exactly the the way to think about it. And look, the ultimate test is like I don't doubt that even given the initial set of partners, Open USD can do really well in the forward-looking fintechs that are you know building wallets, all sort of new infrastructure and products and services, and are kind of more crypto forward. I would say the ultimate test is like: is this a more comfortable solution for a bank that is being courted maybe by Circle with Ark? Come onto our network, you know, trust our asset, or banks that are getting together with probably confused objectives on like oh we should issue our own stablecoin. I think there's at least one effort right with 21 or more institutions that's trying to do that. Can Open USD essentially convince them that like why bother? This is already the right structure. Is it neutral enough? You know, or do you worry about oh the influence of you know a Stripe or Coinbase or in Mastercard, in Visa into that mix, right? That that's the ultimate test, and and I think to be honest, your intuition of like it's all about shipping product is the right one, which is like if this consortium can get together effectively and ship wonderful product experiences, people will gravitate towards it, and the economic design is the right one. It's very similar to what we had in Libra. Look, a few months into the Libra process, there were already discussions between different members that have conflicting incentives of some features of the standard. So I think you know Zach and team should expect some of that tension. As soon things get real, you get excitement, but you also get people starting to pull in different directions, so that's one test. But to me, the ultimate test is like: Can you convince the GSE that this is a better solution for interrupt than whatever they may be discussing? You know, across the spectrum, from like clearing out and tokenized deposits to maybe using some form of tokenized money market fund or whatnot to the things like finality, where you're trying to get as close to the central bank metal as as possible, that would be ultimate victory for this standard if you can convince banks to peel off and start using

Cuy Sheffield  13:32  
it. I I think one of the biggest opportunities are large emerging market banks, and from the the initial announcement in the summer, one of the most interesting kind of set of participants is you know you had FMB, you had APSA, you had Ital, you you had some of these very large banks that are seeing stablecoins in their backyard, but they haven't really invested and built and integrated stablecoins into their products yet, and I think we see that as a big opportunity for Visa. We've got close relationships with many of these banks. They need support, services, infrastructure to help build those products out. And I think it's also really important when they go to regulators. You know, we've gotten this feedback from banks in a number of markets. Like it means something to say that we're working with Visa and Open Standard, this entity that has some of the largest companies in the world, kind of coming together around this stablecoin, and that the reputations of those entities go a long way towards getting some of the first kind of regulatory approvals, you know, for these large banks to be able to to get to the space. And so that's something that I think we'll we'll continue to see, and then I think capital markets is just wide open. As there'll be many use cases in both trading, on-chain lending, that there's just there's room for more competition. It's it's early enough that I think OUSD has a really important role to play.

Sy Taylor  14:55  
Yeah, definitely, Kai. I mean, on that, if you look at the Open US. List of companies. Anyone could look at that and go, "Oh, I see something that looks a bit like me. I don't know if that was true of Libra. I don't know if that's true of the 21 Bank Consortia. I don't know if that's true of many other things. But in this one, it really is has made a concerted effort to be for everybody. But it's this interesting dance of sort of there are some short-term use cases that have real traction, and the failure mode of this is it wins at those use cases and nothing else. And if it does that, it's an enormous success. And if it can do more in capital markets and beyond that, then it gets really, really interesting. And I think one of those use cases links us nicely to the next story, actually. And I found it so interesting that one of the UK's largest banks, Lloyd's, used USDC to settle $750,000 with Visa. This was a seven-day live pilot. Lloyd's bought the USDC through UK company Artax. Shout out to those guys who are a UK-regulated digital asset exchange, and the obligations were booked through their corporate markets branch in Jersey, and then sent to Visa in the U.S. Really nice regulatory work here. The head of digital assets at Lloyd says stable currency could be particularly valuable for cross-border payments when moving money between markets, currencies, and infrastructures, adds times and complexity. Kai, we've been talking about card issuers sort of using stablecoins for scheme settlement for for a while here. And interesting that USDC here is still involved. Interesting that a UK bank is involved. Lots to pick on, but I know this is a big topic for

Cuy Sheffield  16:36  
you. Yeah, huge shout out to the Lloyd's team, and you know this is a a major major financial institution that's going on chain settling on a Saturday and Sunday. We purposely wanted to do the the seven day pilot so that we could prove that the weekend out. And I think it's one. It's exciting to just see how far stablecoin settlement has come, and starting with crypto companies, and then stablecoin native fintechs, and then regular fintechs, and then fintech-focused banks, and then you know now major banks, and so it's just this progression of adoption has has been really cool to see. And I think the other thing that that resonates a lot is we've talked a lot for years. If you just you have to get off zero in the sense of if you've never really made on-chain transactions as a large financial institution, the earlier you can get to the point where you can test end to end, you know, send a million dollars one place, get money sent back, and I think we we did this first time in I think it was 2021 was like the the first on chain transaction that Visa made, and it was like such an intimidating moment of like, whoa, like we're we're gonna send money over a blockchain, and how are we going to manage gas? And what is this going to look like? And and we did it, and we're like, okay, like that nothing broke, nothing blew up. Like it's it's possible, you know, to be able to to move money on chain. And I think we're seeing large FIs all over the world looking for what are the opportunities of very clear scoped pilots that have you know low risk that give us an opportunity to be able to start to move value on chain and and to be able to prove some of the most interesting properties like the ability to move value over the weekends, and so I think it's it's great that Lloyd's took this on to say let's demonstrate how we can use USDC use public blockchains you know moving value over the weekend and we are excited to partner with them on that, and I think you'll see many other banks look at settlement as this first use case to get some experience that then can evolve into many other use cases in in the future. But Christian, your your thoughts on the the settlement space and how how large banks should be looking at kind of dipping their toes in getting their first on chain transactions?

Christian Catalini  18:40  
I mean, look honestly. Also, beyond the the value of learning, so I think any financial institution that wants to stay relevant needs to experiment with this above and beyond the immediate benefits. And so, yes, settlement is really important. Eventually, programmability, collateral, all sort of new solutions that that will be possible once assets are probably correctly represented on chain will become available to everyone. The key here is also like the cost of liquidity, the cost of credit, the cost of essentially any sort of financial intermediation is coming down, and yeah, as an institution, you need to start experimenting with this. I feel like these pilots are really glimpses of what we can do in the future, and I imagine over time, you know, they'll scale up dramatically across a number of things. What's interesting to me is that, in a sense, it's almost like we're still in this phase where you know you're running the the car on the road that's designed for horses, and so every one of these pilots has to create all these kind of convoluted structures and and flows, but eventually, as more of this is all on chain, suddenly having that composability will really unlock all sort of interesting applications. So it's almost like a slow ramp up, but I would imagine. As soon as large institutions really come on chain and start putting really interesting assets on beyond cash and cash equivalents, there's going to be a whole you know new economy of how these assets get used more efficiently. So yeah, very optimistic about what what's possible here.

Sy Taylor  20:16  
All right, let's just take a quick pause here while we hear from our sponsors. Stablecoin operations usually mean a wallet from one vendor and on-ramp from another, and then controls stitched together across all of them. Visa's stablecoin platform fixes this fundamentally. You can mint, move, and manage stablecoins across OpenUSD, and you remain your own custodian, all in one single environment. Then stablecoin link cards let you spend balances anywhere Visa is accepted. That's Visa, the global leader in payments, and of course, sponsor of this show. You can find out more at Visa.com forward slash crypto. This episode is sponsored by Privy, a Stripe company. Stablecoins can move money anywhere, as we well know, but only with powerful wallets at their core. Trusted by more than 160 million accounts across 180 countries, Privy powers secure, customizable wallets that enable you to go global from day one, from business banking and payments to trading platforms and consumer apps, Privy powers the wallets behind modern financial products. Start building with Privy. Learn more at Privy.io. All right, the next story is that Citibank is going to let its corporate clients accept stablecoins. This is Spring by Citi, the bank's merchant acceptance platform, which once surpassed life. We helped work on another company I used to work for. So, Coinbase virtual accounts run on Citi's virtual account and wallet, part of its banking as a service stack. And business customers can accept, hold, and pay fiat, and incoming dollars convert into stablecoins on arrival. CityBooks Spring in 2019 with global payments on cards and PPR and local payment methods. Coinbase now fits alongside that as its stablecoin solution. Guys, I'm interested in your view here. Stablecoin acceptance for merchants seems to be the one area that hasn't really taken off. I think Shopify has had this capability for for quite some time. I don't know that consumers are banging the door down of large merchants to pay with stablecoins. But do you guys have any evidence otherwise? Is that bias of mine? Am I am I wrong on that? Do you think that's going to change soon? Christian, Carl, your thoughts?

Christian Catalini  22:42  
I do think it's a much more compelling solution when you go cross-border, right? So, if it is a global merchant trying to accept payment, maybe that that's a more viable payment method. But at least domestically, I mean, all of these banks, PSPs have kind of struggled historically, as you pointed out, right? City had an effort before in this arena, there's a lot of reasons why you know the point of sale is is still dominated by you know other methods. I think where it's interesting is that if you zoom out, right? So here you have two entities that clearly are going to be competing on different fronts, each embracing a piece of the other in a way that's potentially like a slight expansion. I don't think that's the end state, right? I would imagine banks will will want to control much more tightly some of their interfaces, especially if it is their merchant network. You know how that happens. Is it going to be with a partner? Are you going to do it in house? On the other side, I do think we need to be more prepared for a world where all of these different companies and all of these different assets and instruments will blend, and to some extent, it won't matter for consumers or merchants. They will just use whatever it's available to their wallet, to their particular situation in time. But I do think you raise a really important point, which is like, okay, what will trigger switching, right, from from maybe a more costly payment method for the merchant to something like this that can be more bare bones, more final stablecoin settled.

Cuy Sheffield  24:05  
Yeah, we we still have not seen any meaningful demand for direct consumer to merchant stablecoin payments, and and it's not that there aren't any merchants that accept. I mean, like you said, Shopify accepts, Stripe is accepting. There's just not material demand of of consumers who would rather pay, and I think it's also the success of of stablecoin linked cards has really taken people who have a stablecoin balance and want to spend it. It is much more convenient for them, and and now with many of the rewards programs and the competition in the stablecoin linked card space, the argument of if I have a stablecoin balance, why wouldn't I spend it on a card? Why would I pay it, you know, directly to the merchant? Have a worse experience? It's just we have not seen that that play out. I think the area where it makes the most sense is is on the B 2b side. So I read this story more as kind of preparing for some of the B 2b use cases. We've heard. Some of the the title transfer companies are now accepting stablecoins. That that people are doing kind of large value. What used to be a wire is now a you know $100,000 stablecoin payment. I could see many of those use cases that those kind of consumer or kind of small business wire use cases move over. I think that the experience that Citi is enabling here is something that I expect to be really common across most banks. If you fast forward you five years, you should be able to have someone send you a stablecoin, and it automatically just shows up as a deposit, or you should be able to withdraw from your deposits and have it sent out as a stablecoin. And I think we're starting to see more examples of that seamless ability to treat a stablecoin like digital cash and automatically clear it, have it show up as a deposit. There's a lot of work on the back end to make that smooth, but there's no real incentive for banks, at least in in the U.S. to hold a stablecoin on behalf of a consumer, and so you need that ability to let someone accept it automatically, have it show up in their bank account as a deposit, the same way that they would if they walked into a bank with you know $1,000 of cash that they were depositing.

Christian Catalini  26:14  
Yeah, I think you raise a really interesting point.

Sy Taylor  26:16  
I think that B 2b thing is super interesting, as as Christian was was about to allude to as well. The wires have been just such a painful thing for anybody trying to do B 2b payments. ACH isn't a lot better; it taps out above a certain threshold, and and international wires are even worse. And so, kind of having this instant low cost solution is something that there are a portion of businesses that deal internationally that could find that extremely valuable. I find it so interesting as a theme that you have a very large bank leaning into this, and having a partner like Coinbase kind of allows you to do that. Citi's been quite open to partnering with fintech companies. They've done stuff with Navan. They've done stuff with Gusto. I believe they've done a little bit with Brex as well historically, although I might be wrong. And so, kind of interesting how in their corporate and investment bank they have done quite a few partnerships that you go under the radar and don't necessarily get the most attention. But it's a good way for an incumbent to innovate if they can't necessarily engineer their way into having competitive solutions. Is like do some partnerships with fintechs. These are these are not necessarily your enemy.

Christian Catalini  27:23  
Yeah, and and look, if you if you want to elevate, I I think there's a underlying question that every large bank will need to answer, which is like it's clear the world is going towards the world with greater interop. A lot of the friction, you know, will disappear. And how do you play your cards, right? Do you embrace a stablecoin? Do you try to go clearinghouses plus the deposit tokens or tokenized deposits? Do you use a tokenized treasury? Like, what what is that interrupt solution, and what interrupt solution you're going to use for different use cases? So, I think what what makes this interesting is that you're seeing glimpses of banks dipping their toes on the stablecoin side and trying to understand what do we gain, what do we lose. I mean, there's the looming debate around deposit flight and the yield and clarity. That's of course top of mind for many of these banks. But it's interesting to see that different parts of the banks will have very conflicting incentives on this.

Cuy Sheffield  28:16  
Yeah, I think speaking of deposit flight, yeah, maybe we we should move to the the next story. It's funny that we we've covered deposit flight so many times on the show, but the context of of how you know it's come up mostly has been stablecoin rewards and yield on stablecoins, and just this this battle that played out in DC between the banks and the the crypto lobby, which ultimately concluded in really nothing happening in terms of clarity not going through, and the status quo has remained on on genius. But for the the last story, we wanted to talk about the Apollo chief economist says agents like Amuse will cause deposit flight, so the consumers don't just want a better deal. An agent can move deposits that earn you know fraction of a percent to above 3% This went absolutely viral. I feel like we got takes all across the the spectrum, and so Simon, have to give a shout out to you for your your brain food last week, which I thought was was really really good. Christian, this covers a bunch of topics that you and I have have discussed in the past. So, Simon, you want to start with just like what your high level take was on this?

Sy Taylor  29:29  
Yeah. So, I don't buy that agents will cause deposit flight any more than fintechs cause deposit flight. If you look at sort of 2013 to today, bank deposits went somewhere from around 9 trillion to about 19 trillion. Banks are just fine. Deposits haven't gone anywhere. What they have done is miss out on all of the incremental revenue opportunities that fintech created, and I think that's that's been meaningful. Fintech now represents something like 4% of global financial. Services revenue, if it keeps compounding at this rate, it'll take about 20 years before that half of all financial services revenue. Banks are not growing anywhere near the same compound annual growth rate. It's something like 21% versus 6% So banks are slowly eroding in market share, and this will amuse will and agents like it will massively compound that issue. So yes, okay, agents could move your deposit, but think about the consumer kind of perceived friction of moving your deposits, where your direct deposit goes. Oh, what's going to happen to all of my bills, and what's going to happen to like my kind of everything that comes out of it, my mortgage, and yada yada yada. I've got all of these other things that are linked to it. The fact is, there are fintech solutions that, with a single click, will take care of all of that for you. But that doesn't get rid of the perceived difficulty from consumers. There's a psychology issue here that I think is much much bigger. But there was a really really interesting study by AIVO, they ran more than like 2000 conversations with ChatGPT, Gemini Perplexity, proposing as a small business owner choosing a bank, and in this, the banks came up almost never. So Mercury won with 327 endorsements, followed by Wise with 184 and Relay with 68, the best rated bank Chase got to 60. I think this is going to be the really big thing. If I go to my agent and it's helping me manage my money, which it can already do via Plaid, by the way, at least on a read-only basis, and it can sort of help me find new financial products. It's not necessarily going to recommend a bank, and that I think over time compounds to be something really meaningful and something you should pay attention to. So, in the in the full piece, the Agentic Bank run, which you'll find in my social account or fintechbrainfood.com or brainfood dot xyz soon, then you will sort of see that I break down. What should banks do about it? What will fintechs do about it? And the lessons you can learn from it. But it was a really fun piece to write because I I don't buy the deposit flight story at all. Christian, do you disagree with me, or what are your thoughts?

Christian Catalini  32:14  
You know, and I had very strong opinions on on X when this came out. I I think it was a clever headline by Apollo, it drew a lot of clicks and a lot of engagement. But a few thoughts. So first of all, I think recently there was a report by Clarivis on on this old deposit flight story, and they painted as deposit flight. If you look at the data, it's just people moving you know assets to fintechs and crypto exchanges to trade. So first of all, point one: There is no evidence of the positive flight to date. But second, as you start thinking about what the agentic economy will do, this is what I've been spending, you know, most most of this year on friction and any sort of business that relies on what I call for for those that are old enough to remember the blockbuster late fees. That model, where whether you're an insurance company and your consumer doesn't shop around, or you're holding and earning NII or net interest income because people are lazy, that stuff is going right. So long term, are we going towards an economy that will favor business models that serve consumers and businesses more accurately and better? Yes, and like you pointed out already, Simon, this is nothing new. FinTech has been doing it progressively for for more than a decade. This is going to be an accelerant towards that. Now, many reasons why deposits are sticky. Well, first of all, you know, think about FDIC insurance and the guarantees that come with that. Now we know that genius stablecoins are probably as safe as an FDIC insured deposit, right? Especially if it's you know HQLAs. But more broadly, the idea that everybody will shift in mass because Muse is recommending or Grokbot or whatever Google comes up with. I mean, there's going to be multiple of these, and that's going to cause a run. That just makes for a good deadline. It's not how it's going to unfold. That said, the reality I think is going to be that over time, consumers and businesses will be a lot more tougher with their cash and cash equivalents. Today, the only interrupt player in the United States for your money is unfortunately a checking account. Whether you connect to your fintechs, eventually your salary, everything, your mortgage, like you were pointing out, everything has to route through this clunky piece of infrastructure, and that's what has made this extremely sticky. But you know, we have number portability for mobile. Eventually, you know, you can imagine a scenario where we're gonna have really good interoperability between all these different services. That's when I do think the agents will make the right calls and will drive business. So if you're a bank, yes, I think you should have a plan for a future where the comfy part deposits are going to go away through competition, through fintech, through stablecoins, through everything. But yeah, is this something that is almost like the agentic swarm that? We've seen with OpenAI, you know, attacking arguing face because that's going to be the same. Where all the funds are going to move away. That that that just makes for a good headline.

Sy Taylor  35:06  
I don't know if you saw the Meta stock was up like 15% and the S and P bank stocks was down 5% It's like it's like Citrini research all over again. But I don't exactly think this is the bank

Speaker 1  35:17  
pocalypse. So yeah,

Christian Catalini  35:19  
and and look, I think I think this is where I mean the market is is is doing a lot of vibes lately, but people have been thinking about the AI bet as essentially it's all going to be entropic and open AI. I've written extensively about this, but I strongly disagree. I think the winners in the AI market will be the holders of complementary assets. So what you're seeing with news is essentially looking at a massive platform with massive distribution that can create wonderful product experiences. Saying, we don't need the the frontier model, right? We can make it actually useful to consumers. Meta is not going to be the only one. I think you're going to see Google eventually, even Apple. Everyone is going to play in that. Some leading fintech will do the same. Yeah, that's just going to happen.

Cuy Sheffield  36:03  
And so then, how should banks respond? I think even if you know you don't believe, it sounds like we're pretty much in consensus that this deposit flight boogeyman-that you know-it's just kind of this fear that is thrown out there for any new technology. But there is this steady progressing over a longer period of time, reduction in friction, and any business model that is expecting a consumer to not have perfect information or shop around is is going to have more difficulty. Christian, how do you think about put yourself in the shoes of you're a a bank CEO, you're looking at kind of this future of agentic banking, agentic finance. Like, what should you do about both continuing to maintain the customer relationship, and then also I'd love your thoughts on how do you think about like the sovereign AI stack in for banks to actually progress and use AI. Like what what do you need to do as as a bank CEO to remain relevant the next decade? Yeah,

Christian Catalini  37:00  
look, I don't believe in the pacing the fintech frontier. I think if I was leading a global bank, I would just embrace it and rather disrupt myself than waiting to be disrupted. Build something customers want. Right now, customers will want more choice. They will want better products and services. You can be ahead of that game, and now that can be very painful for companies that have a meaningful share of revenues in their payment business coming from park deposits, but look, it's happening, right? So you can slow it down. You can try to make it uncomfortable. You can create all sort of friction. If you look at the music labels, right? So the ones that embraced streaming and understood that the the economics of distribution of music were going to zero, and they had to win in a different game. They tried, so I don't think this is a extinction event for the incumbents that embrace the technology and understand that yeah, the blockbuster late fees in the financial services industry are going away, and we should play by new rules. There's plenty of ways to add value. There's plenty of way to make that deposit still sticky and useful, but there's going to be new ways. And to answer your second question, I I think it's going to be almost ridiculous a year from now or two years from now for any fintech or bank executive to send their traces back to a closed frontier model. They're going to realize that part of what makes them a viable business is they're essentially a finance verification factory. They can get information in, better measurement in, and they're gonna manufacture better tokens out of that process. And if you don't own those traces, if you don't own your weights, if you're not running on a model that maybe it's like an open weights model plus your proprietary context, you're essentially slowly sending your secret sauce away to someone else that could be a disruptor. Now, of course, you know the labs don't have licenses, and so there's always going to be friction. But imagine a world where you're essentially just a dumb pipe, the holder of licenses, and all the value is created at the ChatGPT level or at the cloud level. I think that's not the future banks want. And Simon was already going there by saying, you know, if your agent tells you which bank account to open, then well, they're the massive aggregator, and as we know, aggregators are really valuable when things get unbundled.

Cuy Sheffield  39:09  
So, Christian, should banks build their own agents? Should they build CLIs? Should they build MCPs? Like, how do you think about just staying relevant from a distribution standpoint? Like, do you think our MCP is going to be like, you know, the same way every bank has a mobile app today. They're going to need to have an MCP in the future, and they're going to need to connect into these agents. Like, what what would you do on the the front end side of it?

Christian Catalini  39:31  
Absolutely, I I think they need to embrace the technology. They need to dip their toes and look. Maybe in the end, they're going to use a third party for some of this, but nothing beats the exercise of going through the motions of building your own MCP server, trying to make an agent available to your customers. It's not that different than back in stablecoin early days, right? We were all there when the forward-looking institutions were playing with permissionless blockchains, and maybe there were pilots, maybe there were early experiments. I think all the all of. Those institutions are better off today because of that learning.

Sy Taylor  40:03  
I spoke to three different leadership teams at banks last week who'd all been playing with Muse, and then when this news broke, I spoke to another one as well. They all asked me two questions: one, should we block this? They saw that Amazon had blocked Muse, and they were like, "Is that a thing we should do? There was an insurance company called Insurify that did that, and their stock price spiked. I'm getting sort of flashbacks of open finance and open banking, where some of the big banks would like intentionally sort of break screen scraping, and and I think that's a possibility. But also, I don't think the evidence is there for doing that just yet. Like, I think it's a bit early, but I do think you need to understand how that secure interaction is going to work. To the point that Christian was going to make, if if an agent is coming to your data, how are you making sure that that's 100% secure? Because there's so many things that could go wrong for consumers and businesses if that data leaks in some way. I've spoken to many banks as well who are intentionally, knowingly a little bit behind on what they could do with AI because they don't just want to go do an enterprise deal with one of the big labs because those big labs will tell you till they're blue in the face. We won't train on your data. We'll never see your data. But frankly, I think experience with big tech has taught us people say that, and it's true until it isn't. The second question I get is, should we build our agent? And I have a very simple answer to this: only if it's actually good. Nobody wants another shit chatbot. They really, really, really, really, really don't want a terrible chatbot. So don't do it. What you can do is learn. I think, from Public.com and Robinhood, who I think have done this really, really well. So they opened brokerage accounts for people's own personal agents, like Claude Code and Codex, and that sort of thing. They ended up with about 150,000 people who had these separate agentic accounts, tightly constrained, you know, really limited in what they could do. But these power users allowed them to observe what will people do if they had access to the account. And then just a couple of days ago, they announced Robinhood agents, which is like the consumer packaged version of that. So if you think about the difference between Claude Code and using Muse or Instinct, there's like a consumer version and a and a nerd version. That's a thing you could do as a financial institution. You probably have some customers who are massive nerds, and you can give them a really sandbox version of that account to learn from. Coinbase has done this. Mercury's ships an MCP server and an account for agents. Meow, Ramp. Go look at what the fintech companies are doing. There's your roadmap. But then I think the big thing you got to do is give the agent a reason to choose you. You got to make your bank worth choosing. Like keeping customers' money safe is one thing, but like you're in a competitive market now, and the reason you're showing up last, I think on a lot of these AI benchmarks is because objectively your product offer doesn't look as good and certainly isn't described as well on the internet. So I think there's there's three things you can do. You've got to move faster. You just got to move faster. I think this is going to be much bigger than mobile because with mobile, somebody launches a new neobank mobile app, and you do one that's like half as good, and you keep most of your customers. This is like actually the consumer has its own agent. How is a consumer going to delegate authority to that agent for it to be allowed to interact with the bank? I think Visa is doing some work on this with Tap Protocol, and I've seen some other good efforts. But like, you really need to be thinking about that, and I know a lot of consumer finance teams are. So long may that continue. So I'm conscious of the time, Christian Kai. Any any final thoughts before we move on?

Christian Catalini  43:53  
I think my my parting thought is really like, look, we're going towards an economy where if you're not doing something that customers and business really value and want, you're going to be in trouble. And so, to go back to your gradient of how fast you accelerate, the more you rely on friction that's about to disappear, the more you should really hurry up and and both embrace you know stablecoins, tokenization, but also the agentic flows. If you're already delivering value, then maybe you're less at risk, right? The agents will find you no matter what.

Sy Taylor  44:23  
I like it. The agents will find you. Sounds vaguely like something from The Matrix, but there's a bunch of stories we didn't have time to cover this week. The DTCC took a stake in iCapita. Cloudflare opened its closed beta of its monetization gateway, which lets websites charge AI agents per request in stablecoins. The Financial Conduct Authority has started accepting crypto authorization applications ahead of its 2027 regime. Hallelujah! And Robinhood is going to launch perps and weekend stock trading for its U.S. users. Always more. News. I wish we could do a longer two, three hours show, but Kai's got to go be a dad, and Christian's got to go save the world. So thank you everybody for watching and listening wherever you are, Kai. If people want to find out more about you, where do they go to do that?

Cuy Sheffield  45:12  
On X at Kai Sheffield and visa.com/crypto. Christian,

Christian Catalini  45:16  
see Catalini on X.

Sy Taylor  45:18  
You'll find me at sy Taylor on all of the socials@fintechbrainfood.com, and of course at tempo dot xyz, where we're doing lots of interesting things with large financial institutions. Please, please, please remember to subscribe to this podcast wherever you get your podcasts, and leave us a review. Hit the like button if you're watching on YouTube. It really, truly helps us. Thank you, and we'll catch you next time.