Visa Inc. ($V)

Guys, why don't we jump in? Thank you again for everyone for joining us. Day 2 of the Wolfe Fintech Forum. Really happy to have Visa with us. And Jack, I think I've had you on stage a couple of times before, but it's really great to have you back. Jack is the Chief Product and Strategy Officer of Visa. Also happy to have Jen from Investor Relations here. And look, if it's -- now it is about as great of a time to have you with us as any, just given the magnitude of change happening in our space. But there always is, but it seems like there's been more now than usual. And so first of all, thanks again for joining.

Thanks for having me. Great to be back.

We're going to get into the different changes in the industry and how it affects Visa or it doesn't. But before we do, maybe a quick update. You obviously see what the business is seeing in terms of trends. And so U.S. payment volume was up 8% in January, credit up 9%, debit up 6%. Overall, I think it was about 1 point better than calendar fourth quarter results and stable cross-border, generally speaking. Just tell me what this means about the current health of the consumer? And do you believe the K-shaped economy and we're hearing from others. Just what are you seeing out there?

Yes, sure. Again, thanks for having me. It's great to be back. Look, we're obviously living through some tumultuous times right now. But I have to say the watch words for the consumer driving out of the data that I'm seeing are stability and resilience. I mean you just rattled off some of the numbers. We're just seeing consistent growth performance quarter in, quarter out, month in, month out. I lead our product team and our strategy team, I sometimes focus on some different metrics than some of my colleagues. One of my favorites is processed transaction volume growth because to me, that's the most pure indication that we've got about how users are engaging on the platform. And that number has been incredibly resilient and consistent on a broad global basis for quite some time. So look, there's a lot going on in the world. You might expect that to be changing the way consumers are behaving with respect to their spending patterns. Overall, we're really not seeing that. You asked about K-shaped. Look, I don't really see K-shaped. It's probably a little bit more letter E-shaped if I've got growth rates on the vertical axis, right? We're just seeing consistent levels. There's definitely a differentiation in the level of growth rate that we see across the spectrum. But it's a positive growth rate at the bottom end and it's a positive growth rate at the top end. And the gap between the 2 has been relatively consistent as far as...

Yes. We've actually heard that. I remember even Chris on the last earnings call was talking about how some of the more affluent are spending a little faster growth rate-wise, but consistent, right, across all the areas and they're all growing. All right. That's great color. Maybe just a quick follow-up would be with all the political turmoil going on in geopolitical in the Gulf region, et cetera. I mean, any impacts from -- on cross-border gas prices? How do you as a company think about that or FX volatility?

A few things I'll say on that. First of all, it's early days. Hopefully, this goes by quickly. who knows what's going to happen with -- like this one. We monitor this stuff very closely. We're always looking at the granular detail of the data with respect to how different corridors are doing with cross-border, how different categories are performing. You mentioned cross-border, you mentioned energy prices. We'll monitor those things. What I would say, though, is it's worth remembering, we're incredibly diversified. We have countless cross-border corridors across the world that we are constantly looking at. So -- we'll see what happens, but I'm kind of looking at it on an aggregate basis and thinking if things are already diversified, [ payment ] gas, by the way, energy prices, we're seeing the volatility there. But overall, gas prices in the U.S. are a fairly small proportion of our overall purchase volume.

Okay. All right. Let's go into the more interesting topics, which I think is really what's everybody's minds. For what it's worth, I mean, as an analyst, we probably don't even field questions over too much concern over the fundamental growth of Visa. That's been consistent and been strong. It's mostly top-down questions, what is the regulatory environment going to look like? And how does that affect? But more recently, it's been agentic, right, or AI in general. And so just let's start there. I mean, you've said that agentic is the next wave of commerce, and you've said it's an opportunity. So just outline what this wave of commerce actually entails. I think a lot of folks are still trying to figure it out.

Yes. Darrin, I'd say I've been doing this payment technology thing for a really long time, 20-plus years. And I will honestly tell you, I have not stared into a bigger growth opportunity than what we have ahead of us in the development of the agentic web and -- agentic web broadly that then will turn into agentic commerce, turn into agentic payments. But I haven't seen anything like this since the dawn of e-commerce itself in the late '90s or early 2000s. We see it as a generative growth opportunity. And I want to make sure we try to get out of the zero-sum mindset, right? There's just too much of the, hey, is agentic a way that the existing pie is going to get sliced up differently? I don't think so. This is about new business models, new transaction flows and increasing the velocity of money. And we're seeing it in very practical ways. I'll give you 4. Think about how agents can take friction out of payments. There's still friction in digital payments, right? We see fall-off rates on transactions. We see decline rates in transactions. Agents are going to be able to take the friction out of payments and transaction success rates are going to climb. When we see transaction success rates climb, we see payment volumes increase. Two, we're going to see what I call increased payment density. What do I mean by that? Think about a fixed amount of purchase volume or purchasing power split into a number of transactions. We've actually been seeing increasing payment density for a long time. Our average transaction size from 2015 to now has dropped 20% from around $55 to around $45. You might say, "Oh my gosh, that's great. Because at the same time, our number of transactions has tripled to $300 billion. And so you're seeing the increased payment density, and it's coming from subscription purchases, streaming payments, gaming payments, creator, all these new use cases that have lower tickets and higher velocity. And we think agentic commerce is going to completely unshackle that, right? You think about an agent -- we can be kind of lazy, right? I'm buying a bundle of goods. I'm kind of at the same merchant. I'm buying all the stuff at the same time because I'm there. Your agent doesn't need to do that. Your agent is going to split that purchase into as many little purchases as it needs to. Streaming payments completely...

Find the best price, find the best product in 3 different locations.

Do you want to buy an hour worth of something, do you want to buy a day worth of something instead of a month or a year. So I really think we're going to see a significant increase in transaction volumes coming from that. Then we're going to see faster digitization of payments. There's still a lot of undigitized payment flows out there. My favorite here is B2B payments. There's an enormous amount of B2B payments volume left to be digitized and a lot of friction left. And when you think about onboarding suppliers, raising purchase recs, invoicing, paying, reconciling all that stuff, it's still painful. Agents will rip that friction out and we'll see a massive digitization of B2B payments. We're doing some really cool work with Ramp, by the way. There's some stuff up on YouTube, if you wanted to take a look at it. But those guys are integrating AI agents for their business buyers with virtual Visa cards in hand. And it's basically an invoice to pay instantly with full-on reconciliation from the get-go. It's a really cool if you're into B2B payment.

They were just -- Ramp was here this morning actually talking about that to some degree.

And then I'll give you one more.

Yes please.

We're not takers on the agentic dystopia argument, right? Like every other wave of technology that has hit the payments ecosystem has generated growth. We've seen it with e-com. We've seen it with mobile. We've seen it with mobile cloud. We've seen -- and we see this the same way. The economic forecast that we're looking at consensus is anywhere from 80 basis points to 150 basis points of GDP growth coming out of the efficiencies at a macro level that we can gain from agentic. And we know that when we see GDP growth, that flows to PCE growth, and that just flows straight into our ecosystem.

All right. That's very encouraging. Look, on that note, I mean, we still -- despite everything you just said, and you answered some of it, but we still get questions about whether this means -- this is a good thing or a bad thing. Is Visa a winner or a loser in agentic -- maybe just articulate if there's anything else you can add on to specifically why Visa could win in this versus others that might also try to offer these offerings.

I think there's some lessons in history there. I'm not going to go dwell on it. But again, you think about the complexity of the transaction that came into existence at the advent of e-commerce, right? It was riskier. It was less identifiable. There was no means of having a physical identifier in place. We adapted our technology. We created new authentication mechanisms. We changed. It was a moment that was made for us. Same thing with mobile technology. We created tokenization. This is the same. These transactions are going to be riskier. They're way more complicated. You've got an agent in the middle, the agent needs an identity. You need to secure that identity. You need to validate it. You need to collect more data in order to be able to ensure the security of trans -- all that stuff. And we're working on that. So in a way, this is made for us. And maybe just for fun, I could bring it to life with an example. So let's say you were so compelled by my growth argument, Darrin, that when we're done here, you go hit Claude Code or Cursor or Codex, whatever your favorite coding platform is and start spinning up commerce agents because you want to realize the opportunity. At some point, like very quickly in your development process, you're going to have to equip your agent with the ability to buy things. And you're going to have to decide what you want it to use to buy things, right? I would argue that you will want your agent to be able to easily connect to the sources of funds that your human has, Like, say, the -- 14,500 financial institutions that we are already connected to with Visa credentials that store most of the money that is used to buy things in a person-to-merchant context. And you're going to want to be able to use the credential everywhere where things are sold. maybe like at 175 million merchants that already accept Visa cards, right? So -- by the way, you also want -- you want to keep your human happy. And they trust their cards. They trust that brand. They trust that security. They like their rewards. So I think you're probably going to default to card credentials as opposed to something experimental that has no acceptance footprint that doesn't have a lot of other stuff. You're also going to worry about security, right? Agency is great to automate things on behalf of a buyer and a seller. It's also a tool that fraudsters are increasingly using and things are getting more dangerous out there. And we process 300 billion transactions a year, almost 1 billion transactions a day. We use the data from those transactions to evaluate every single transaction in subsecond time with hundreds of parameters to make sure they're safe and they're secure. You're going to want that, right? You're going to want our token technology that is able to bind the credential to your agent and identify your agent and carry the extra data payload you need. None of that exists in these alternate payment mechanisms. And then lastly, you're going to want everybody to understand that this is a trustable payment method. And we spent decades building trust in our brand. And when we present the Visa brand in a payment transaction, the merchant trust it, the buyer trusts it, and the agents are going to trust it, too. So look, we're excited about this. Like this is opportunity of the -- like I said, of the order of e-commerce itself 20, 25 years ago.

Yes. No, it sounds like it. I've seen data points that Visa and if you look at Mastercard, there's a combined 450,000 attempts to hack data centers by fraudsters per day, believe it or not, on the networks that you guys have to pull back, which you do a good job doing, obviously. I can imagine how much that can increase with models and agents helping fraudsters, right? I mean -- so there's no doubt that security is going to be paramount here. Can I ask you, I mean, how do the announcements from Stripe and OpenAI, it seems like they were trying to almost acknowledge that they have to work with the ecosystem. Stripe talked about using Visa tokens. OpenAI, I think acknowledge they're not doing as much as they thought in-house, right? And so how does that really change or impact your thinking on this at all?

Yes. I mean there's...

You probably knew this already, but how should we think about it?

Yes. Look, it's -- I mean, it's early days in agentic commerce, right? We've kind of started to turn the corner on consumers changing behavior from old search to agentic search and that terminating in a payment transaction or a purchase transaction at the merchant website. We still have a ways to go before you've got a more fully automated agentic purchase transaction. That's going to require standards. That's going to require a new form of interaction on the web. And so you're seeing this alphabet soup of standards start to propagate, right? And it's -- in my mind, it's because we're moving from a web construct that had browsers driven by humans interacting by servers that were coded to be human readable to a place where what we really need is browsers driven by machines on behalf of humans with policy constraints on them. And then server side that is machine readable. And these standards need to exist to make the 2 machine readable sides work. So we're seeing it at the web level with standards like model context protocol. We use that for our intelligent commerce platform to enable agents to connect to it. You're seeing it at the shopping level, right? Google's UCP and OpenAI's ACP, like those were standards to say, here's how you should share inventory data. Here's how a shopping mission should unfold. And then there's the payment standards. And that's where we come in. And this is where some of the stuff you were talking about a moment ago is really relevant. Like we've been -- these things take time to bet in, but we published our -- these intelligent commerce specs almost a year ago, our trusted agent protocol that's about agent identity over the course of this summer. It takes time for these things to get adopted, but you saw Stripe announced that they are embedding Visa Intelligent Commerce into their processes. And that's our play. Like you go back to what we were just talking about, the compelling features of our platform as a payment scaler for agentic commerce are there. But if you want to use it, you're going to have to adhere to the standards. And those standards are going to embed and be interoperable with the web standards and with the broader shopping.

That makes sense.

And that's how you should think about it.

That's helpful, Jack. You mentioned micro transactions. You brought it up before in terms of more transactions also just as an opportunity. But there have been other players saying they think they can capture this. So just help us again with what is Visa's role here? And how do you think about pricing on this?

Yes. This is a fun one. I would say -- by the way, there's probably no platform in the world that has processed more low-ticket, high-velocity payment transactions than Visa. And we have adopted and adapted over time. Like have you ever tapped your Visa card to get on transit here in New York?

Sure.

It works pretty well. That didn't used to work. We had to modify our technology, different transaction payloads in there. We had to modify our pricing and commercial framework to make sure that worked. We've done that in transit. We've done that in vending. We've done that in parking. We've done that over and over and over again. And we just see this as the next one. It's not a question of if we will do this. It's just a question of how we will do this. We've got a lot of experience in batching transactions. We could probably settle them straight through. We need to figure out like does a merchant recipient really want to receive $0.01 every couple of nanoseconds or would they actually rather have that batched up and receive a few dollars every few minutes. So those are kinds of things that we're working through from a technical standpoint. We've got a lot of experience doing this over a long, long time.

You definitely do. Yes.

And then the pricing thing, yes, I keep reading things that talk about our pricing as it relates to these transactions, but it kind of misses the point that we can actually configure our pricing, and we can create commercial frameworks that really work to make these transactions.

Even for small micro transactions, you can make it work?

Yes.

Okay. There's also a lot of talk about different protocols and really nobody better asked than you. I mean just what does each protocol do? And maybe help us understand your role in it, just to explain it. We get -- there's a lot of jargon out there and a lot of folks aren't familiar with it, so.

Yes. And I was talking a little bit about this a minute ago. It's the Alphabet suit. We need these standards. We really need to bed them in if the agentic web is going to work, if agentic shopping is going to work and if agentic payments are going to work. I -- to be honest with you, I don't know what standards are going to win at the agentic web level. I don't even know what standards are going to win at the agentic shopping level. But there definitely needs to be a common set. Think about how the Internet today works, right? We have HTTP as a set of standards. We've got DNS. We've got all these standards. They're open and they're interoperable, and it makes the web work and grow the way it works today. This is just another evolution. So we need a set of standards that govern how agents interact and machines interact with each other. Our paramount concern is making sure that this functions from a payments perspective. And so we're very focused on that. We've been building and publishing standards for a very, very long time. This is kind of what we do. We like the standards that we created almost a year ago and the specs we created with Visa Intelligent Commerce and with our trusted agent protocol, but things are moving fast, and we're continuing to evolve and adapt those standards.

Okay. One of the areas that I know we and others are excited about is the value-added services that you should be able to provide around this opportunity, given all the data you have, tokenization capabilities. Maybe just help frame that opportunity that you see, at least for Visa to really play a big part in the VaaS around agentic.

Yes. You probably noticed our value-added services growth rates are...

Strong.

Good. They're strong. I'd tell you, part of the driver behind that is the attach rate of our value-added services to digital transactions is high, right? And it's because digital transactions are more complex, they are more risky and they need more full featured capability in order to work well. So back to agentic transactions, we're going to see a move toward an even bigger opportunity with respect to value-added services that we can bring to transactions. I mean we bring fraud scoring. We bring authentication. We bring stand-in processing. We bring new data payloads. We bring controls and policy that agentic platforms are going to need to be able to govern these credentials when they're tokenized. We bring the tokenization service itself. We -- the decryption, like all of this stuff is all sort of under that rubric. So we see it as a big opportunity. And in a couple of ways. One, just the sheer volume of opportunity that I was just rattling off, we're going to be able to build more and more of these value-added services and build them into these transactions. We're also excited about the use of AI itself. I mean we're early adopters in AI. We have the ability to actually screen hundreds of variables in sub-second time for fraud in our authorization stream, something we've doing for a long time, but that's because of AI, the precursor generations of AI to the ones we're seeing right now. So we're excited about not only the opportunity presented by agentic commerce, but the use of AI in our products themselves that will cause us to be able to ship at higher velocity with more sophisticated products.

So the fraud dynamic and the cybersecurity dynamic associated with e-comm, which is really even a higher level on agentic, is clearly an opportunity for you guys, right?

Absolutely.

And then when it comes back to the AI use in your own other just broad VaaS, again, you started touching on it in terms of the data itself being utilized, right, or the AI being utilized for better use of your product. But can you just expand on that again? I mean, where are you using it throughout VaaS to expedite the opportunities?

In fact one last thing, I'll come to that. I just want to mention back on the AI, VaaS thing, because I keep getting asked this question. It's the old -- is AI eating software? And a lot of our VaaS services are software. But I would pause it that software businesses that are rooted in network effects, proprietary data and data scale actually benefit from AI as opposed to have exposure to AI. And so again, 2/3 of our value-added services are attached to transactions in our core processing system, which means they're attached to that network effect, which mean they're attached to that proprietary scaled network data that we have. So there's -- we're super confident in the resilience of those software-based value-added services that we have. Now you're asking what are some use cases and examples of where we are applying AI. I'm almost struggling to think of use cases where we're not applying AI. Yes, I mean, in the protect side of the business, we've been doing it for a really long time. But we're doing a lot more of it, but we're also expanding it into new use cases. Rather than give you a list, I'll give you an example. We have a service we call Stand-In. And so issuers make their own authorization decisions in our network, but they often go down or they have outages, there are reasons why they cannot make a decision. And for a long time, we've had rules in place that issuers can configure. We allow them to configure those rules and say, if I'm down, authorize this, don't authorize that. We've used AI to create an intelligent adaptive version of that system that can effectively replicate what an issuer could do even if they were up. And that just continues to learn. So that's a cool example of the kind of thing that our data scientists and engineers are working on.

One of the things that we get asked about is whether or not agents are going to play a role determining what payment method is actually used. And it's actually perceived as an opportunity, but also a risk. I mean, what are your thoughts on that?

I say, I don't think agents are going to change the laws of supply and demand, right? So if you think about digital commerce, it's on the supply side, on the seller side, it's an incredibly competitive business. And the switching costs are super low. It's a very efficient market. And because the supply side has a lot of alternatives and switching costs are very low, the buying side of digital commerce has a lot of latitude in negotiating power. And so as it sits today, the buying side makes the determination about payment vehicle, right? That's just how it works. Now enter agency. Agency will come first on the buying side to create all those efficiencies and take out the work on the behalf of the consumer and the buyer. It will also come on the supplier side. But imagine your agent goes out and it's trying to buy something for you and the agent comes back and says, Darrin, I got the thing. But now I need to go like buy some gold bullion or something. And by the way, you're not getting your rewards. I think you're going to get a new agent really quickly when that happens. And I think your agent is going to learn what -- within the parameters and policy that you set for it, what you want it to do, you're probably going to instruct it to use your card, maybe you'll do a little bit of optimization around that. But I actually think agency will be a multiplier effect on the existing model, which is buyer decides buyer has leverage.

Right. The consumer should be the one that's choosing what's done. Okay. And then another hot topic, obviously, that sort of connects to this is just what you're seeing in terms of demand for stablecoin. Just what are the greatest revenue opportunities for Visa around it? It does dovetail with questions we get about agentic sometimes because it goes back to the last question, whether folks are going to route it to stablecoins. But just generally, what are your thoughts there for a minute in terms of demand and opportunities for Visa?

Yes. we're very optimistic about stablecoins, too, but it's very different from agentic. Agentic, when you think about the way agents are evolving and the way they're going to integrate into shopping, they're working at the application layer. The change is very obvious as you as a user start interacting with an agent as to what's happening. Stablecoins are infrastructure. This is not a solution in and of itself. This is infrastructure. It's really powerful infrastructure, by the way. And a lot of people don't sort of realize what the underlying bone structure of Visa is, I mean we operate a front-end messaging infrastructure that creates promises to pay in real time across our network. The actual movement of the money happens offline from that, and we use the same settlement infrastructure that everyone else in the world uses. And I can tell you from firsthand experience, whilst we have done an amazing job reinventing the front end of money for the digital era to make it all feel so seamless and convenient that you can just forget about it and trust it. The back end of money, the settlement infrastructure feels like it did 20 years ago. It's brittle. It's not real time. It consumes a lot of liquidity, and it could use a lot of improvement. And we are huge believers that stablecoins can improve that underlying settlement infrastructure. That's the way we see it. Now we then see opportunity for ourselves because we see ourselves as a bridge from that new settlement infrastructure that's going to start to permeate and propagate around the world to connecting it to that front end of money that's already functioning really, really well. I'll give you a couple of examples of the things that we're working on. One, on the settlement side itself, we're starting to settle transactions in stablecoins on our own network. We created that capability about 6 or 7 years ago, by the way, at that point, no one cared. It was an experiment. But as we've gotten regulatory clarity in different markets, we started to open things up. So we opened it up first. We're in Singapore and Hong Kong. And now we've opened it up just in the last 6 months or so in the United States. And not surprisingly, we offer our clients the ability to settle with us 24/7, over weekends and relieve them of the collateral and liquidity constraints that, that old model creates, they're doing it. You go back a year, I said we did this 6 or 7 years ago, a year ago, we had almost no volume on this. We hit $4.5-plus billion in annualized run rate in the last quarter in stablecoin settlement volume. So it's a real deal. that's the kind of boring back-end side. I get a little more excited when you start like connecting it to the front end. And there, we are starting to see, especially in emerging markets, a lot of use cases where consumers and businesses are coming into possession of stablecoins. Some of it's about dollarization of savings and deposits for citizens in countries with a lot of volatility. Some of it is just efficient payouts, right cross-border payouts. So I'm paying my gig economy workers in another country in 17 different places, I can do it with stablecoins really efficiently per the settlement thing. The reality is once those businesses and consumers come into possession of stablecoins, they didn't want them because they like stablecoins. They wanted them for other reasons, right, the efficiency and effectiveness. Then what they want is to be able to use them just like they were fiat money. And that's where we come in. And so we've been supplying stablecoin-linked and stablecoin native card credentials. And that's what users want. They trust it. They know it. That's how they're used to paying. That's what they do. But who likes stablecoin-linked cards as a mechanism to enable stablecoin payments even more than consumers? Exchanges, wallets, fintechs, financial institutions...

We've been hearing multiple companies at our conference talk about stablecoin cards.

All these guys...

Now it's an opportunity.

If you think about it, put yourself in their shoes, you're custodying these stablecoins on behalf of your users and your users want to use them. You could go tell your user, give it a shot, try and spend your stablecoins on a wallet-to-wallet-based transaction where almost nobody accepts stablecoins and you've got to go through a hassle experience or you could say, look, I'm going to take Visa's trust network and place that on top of the stablecoins, you know it, you trust it, you know where you can use it. Merchants have to change absolutely nothing. And I don't have to worry about as the custodian of these stablecoins, building acceptance network, managing risk, managing repudiation post transaction. All that stuff just gets taken care of by the Visa stack. And by the way, they like the commercial model, too.

That makes sense. That makes sense. Last one, and then I'll take a couple of questions from the audience. But if you looked at all the different deals you've done and investments you've made over the last 1 or 2 years, what are you most constructive on? What are you most excited about in terms of investments or even M&A that you've done that's turning into some real fruit? And what do you need? Or what do you want to do in the next year or 2 that [indiscernible].

Yes. I mean we've been talking about a lot of it. I mean we are investing heavily in AI across the board. We talked about it in the context of the market-facing product set. But we've invested heavily in equipping our entire workforce from an AI standpoint. And on the product development and engineering fronts, we're just continuing to see the benefits from that. I mean in the early stages of AI coding tools, it was sort of a productivity enhancer. You get a little bit of juice out of that. You can add a few more agile squads to your -- the most recent versions of AI coding tools are starting to create more discontinuous benefits. Like just a couple of weeks ago, one of our teams said, "Hey, I've got a concept, and I want to take this concept from idea to working prototype with documentation and code that I could expose publicly. And one team came in the office over the weekend. I think they had this idea like Wednesday, and they had code ready on Monday. This would have taken weeks or months in the past. And it's now sort of distilling down to a single team in days. I'm just like wildly excited about the output of that Yes, it gets back to like you were talking about the opportunity on the VaaS side of adding new features into our VaaS portfolio. Now we're starting to see the ability to invest all those -- all that productivity and throughput into new product development. Kind of creating a flywheel out of AI.

That makes sense, Jack. Guys, any questions from the audience? We have about 2 minutes, and that's -- there's one in the back, if we can get a mic.

Thanks Jake, I had a question about your BNPL Flex credential, right? I mean that seems like a pretty major shift from how you think about issuing. How are you seeing uplift and take-up both in the U.S. and outside of the U.S.?

Yes. We're seeing a lot of enthusiasm for Flex credential. We've been out there with it for a while now. the sales pipeline is incredibly full. But what I'm more excited about is we're really starting to see the implementations grow. We've got our anchor client, SMCC in Japan, continuing to see spectacular growth. We've got Affirm. We've got Klarna. We've got Emirates MDB in the UAE. So the clients are really starting to grow. But the more fun thing is like when you start to look inside of those portfolios that have launched, you're really seeing a lot of growth and engagement on the part of consumers. So excited about the commercial momentum that we have on it. But I'm also just excited about it as a long-term new way to think about card credentials. You can almost think about the Flex credential as an identity credential, right? It identifies you and then has subcomponents of it that can do different things. You can fund in different ways. It's going to really relate, I think, back to agent and back to stablecoin in meaningful ways, too, because you can imagine having subcomponents of your Flex credential that are bound to agent identity so that you can create controls from within your Flex dashboard for all of your different agents that you've created tokenized credentials for. You can have subaccounts for different stablecoin...

[indiscernible] relationships with the consumer directly then? Do you see then some type of...

Yes. No, I wouldn't say that. I mean we're really offering tools to enable the issuer to expand the breadth and depth of the relationship that they have. Now -- and by the way, this goes back to the value-added services piece. Flex Credential, we offer it in 2 different ways. There's a sort of completely self-service way where we're just exposing API endpoints and an issuer can actually do everything I just said by consuming our API endpoints. But we also have a more full-service version where we're offering issuer processing. But the branding and the design of the experience is all the issuer.

Okay. I think we're going to stop there in time. Jack, thank you so much for joining us. Great presentation.

Yes. I really appreciate it..

Thanks for joining.

Thank you.

Guys, we have the CFO of Affirm up next in about 4 minutes or 5 minutes on stage. So stay in your seats, if you can. And thanks again.

Thank you.