Affirm Holdings, Inc. (AFRM) Q1 2027 Earnings Report, Transcript and Summary

Good afternoon. Welcome to the Affirm Holdings, Inc. First Quarter Fiscal 2027 Earnings Call. [Operator Instructions] As a reminder, this conference call is being recorded, and a replay of the call will be available on our Investor Relations website for a reasonable period of time after the call. I'd now like to turn the call over to Zane Keller, Head of Investor Relations. Thank you. You may begin.

Thank you, operator. Before we begin, I would like to remind everyone listening that today's call may contain forward-looking statements. These forward-looking statements are subject to numerous risks and uncertainties, including those set forth in our filings with the SEC, which are available on our Investor Relations website. Actual results may differ materially from any forward-looking statements that we make today. These forward-looking statements speak only as of today, and the company does not assume any obligation or intent to update them, except as required by law. In addition, today's call may include non-GAAP financial measures. These measures should be considered as a supplement to and not a substitute for GAAP financial measures. For historical non-GAAP financial measures, reconciliations to the most directly comparable GAAP measures can be found in our earnings supplement slide deck, which is available on our IR website. Hosting today's call with me are Max Levchin, Affirm's Founder and Chief Executive Officer; Michael Linford Affirm's President; and Rob O'Hare, Affirm's, Chief Financial Officer. In line with our practice in prior quarters, we will begin with brief opening remarks from Max before proceeding immediately into your questions. With that, I will turn the call over to Max to begin.

Thank you, Zane. We had a blue tsunami on Tuesday and a red wedding today. We are not hanging up a white flag though. So thank you for tuning into our call. Hopefully, we'll provide some balm for your minds and some mirth for humors. So let's get to your questions. Over to you, Zane.

Okay. Thank you, operator. Let's go ahead and begin the Q&A session.

[Operator Instructions] Your first question comes from Rob Wildhack with Autonomous Research. We apologize for the technical issue. We do have a question coming from the line of Adam Frisch with Evercore ISI.

Going to start off with a wiseass comment. Like did I miss anything? But I think we're okay. So I want to talk about growth for a second, and that's the crux of my question. I actually have 2. And this -- follow me along here. If you took your business as it is today and didn't discount any new growth opportunities or initiatives that you're working on, things like Edge and international and all the stuff, Max, that you're doing with your crew, how long do you think you could grow the existing business in the mid-30s range? And then the follow-up, and then I have one more on agentic, if you take all the new and emerging initiatives and put them back into the algorithm, how much longer do you think you can maintain growth in the -- somewhere in the mid-30s or higher or something like that?

I think you're asking me to make a forward-looking statement. I'm telling the team all day every day that we are not to grow less than a 3 handle. It doesn't necessarily mean that we're guaranteed to do it just because I will it so. But I think a key part of my job is finding a way to grow at 30-plus percent. As far as I'm concerned, for a very long time, we are less than 4% of e-commerce. We are a speck on the overall commerce, and just in the U.S., we're live in 4 markets. We're going to be live in a few more. So there's not really a reason why this can't continue for a very long time. Like it's ours to screw up. So -- and there's a lot of really interesting things we're working on that are not in the picture. So yes, I think we have lots of outs in the game. There's a reason why we keep on counting the number of quarters. We've been growing by 30-plus percent consecutively. Just trying to make a point.

And obviously, Adam, we've tried to reflect that in the full year guide. We've set a floor for growth. That floor now starts with a 3 as well. We're signing up for at least 30% growth this year. So we're doing what we can. We've seen really good momentum in the business. Even in the first month of Q2, the business is growing at a high 30s rate today. So yes, I think there's a lot to be bullish about from a growth perspective.

Okay. Awesome. And then if I could just squeeze in one more on agentic. There's lots of speculation out there that certain forms of buy now, pay later, like pay in X and 0s are "better" for consumers because they're, in essence, getting free money. And agents will favor that, and therefore, there could be a shift in ad dollars over time to buy now, pay later platforms. Do you agree with this thesis? And is that one of the longer-term drivers that you're spending time on?

We're definitely spending a lot of time in agentic. So that's certainly accurate. We are in really productive conversations with some folks, maybe more than that with others. So directionally, very important, we care a lot about it. The 0% better than not 0% is a slightly simplistic way of cutting it. If you are looking at someone who's likely on this call, the answer is unequivocal yes because you're getting free money, especially in the world of slightly rising interest rates. If you are a salaried employee and are managing your cash flow month to month, you hear a lot more about a monthly affordability metric versus a true interest number you're looking at. And so there are plenty of consumers that we serve at lower income levels, who are perfectly great risks by the way. They'll gladly pay you back and do so time and time again, but they really care about longer terms. And at some point, 0s become uneconomical in longer terms, and they are completely okay with that. And of course, retailers are excited to subsidize to compel them to transact anyway, but they may not care about a 0% loan as much as they care about less than $100 a month. If it becomes a 50% APR instead of a 25% APR, they're excited, but even at 25% APR, if it fits into their monthly budget, they're excited. I think the thing about this is agents are reliably going to pick the right instrument for that consumer. For a lot of people, it will be 0 is better than not. For a lot of people, it's going to be the longest term possible is great. And there's plenty of in between, too. So as I think in general, agentic shopping will optimize for the most rational transaction pricing. We're the only game in town, broadly speaking, that knows how to price every transaction at an individual level in real time every time. So we think we're going to benefit significantly in the world of agentic shopping.

Your next question comes from the line of Rob Wildhack with Autonomous Research.

Can you hear me?

Yes.

Okay. Great. First, a quick one for Rob just on funding costs. The letter called out that they're up 27 basis points quarter-over-quarter. I think the old rule of thumb had been 40 bps for every 100 bps in broader rates, but you're already at 27 here. So can you just talk about the drivers there and then your ability to offset higher funding costs elsewhere in the P&L since it didn't really have much of an impact on margin?

Yes. I mean there's seasonality, right, in terms of how the balance sheet is composed from quarter-to-quarter. So I guess I would point towards the year-on-year declines in funding cost as a percentage of loans on the balance sheet. So we're pretty proud of that fact that were down at 6%. I think that's a really, really healthy level for the business, and that's, frankly, showing up and helping us drive to the 4.3% RLTC take rate that we posted in Q1. So I think funding costs, they continue to be something that we feel like we have really good line of sight into. We did call out in the letter that of the on-balance sheet funding, more than 70% of that funding was fixed rate in nature. So I wouldn't read too much into the sequential uptick in terms of that being a pure flow-through of rising benchmark rates. I don't think that's the right read here. So we've done a lot to, frankly, make sure that we take advantage of what has been a really compelling ABS environment for us. We've been in the market with several deals, and the good news about the ABS market is those deals are on the revolving side. On the on-balance sheet side, those are fixed rate deals in nature. And we've even gone out to 3 and now 5-year deals recently here in the market. So I think the combination of all of that is that we feel like we have a really good line of sight into funding costs, and that's being reflected in all the guidance that we set, obviously, both for Q2 and probably more importantly for the full year.

Great. And then maybe one for Max. In the letter, you talk about the transformer-based architecture underwriting model. And I think I can understand what that's doing, but would just love to hear you talk more about how that works and what you think a new model like this might unlock in terms of accuracy, lower loss rates, faster growth, et cetera.

So there's actually an excellent -- if you search Affirm technical blog, I think that's the headline, you'll get like a 2,000-word essay on the exact architecture design almost exactly. We took care to remove some trade secrets from it. But it's a pretty disclosive document describing both the process of building this thing and how it really works. So if you want real details, they're all there. The extremely shortened version is there's a bunch of information both in external data sources and our own that encodes really subtle patterns or that represents really subtle patterns, in particular in repayment but also just other relevant metrics. And when you're designing features, like the thing that you actually use to predict off a model, up until recently, the only way is you have really smart humans and they sit around, and they're like, so if we look at the patterns of when this person may have been paid, we don't really know exactly. Like -- so you do a lot of hypothesizing, and then you say we think we see a pattern and we're going to try to encode it in a feature. And then you put it through a bunch of things, including things like, well, would this possibly correlate to prohibited basis; is this too close to something that could correlate to race or where you live. And so you do a lot of manual work and then eventually end up with a model that predicts risk and you back test, et cetera. And so we built this really cool transformer model that learns -- figures out how to learn these subtle patterns without any human involvement at all and that produces essentially feature candidates and then trains on that full -- entirely in a single pipeline. And so one, you're picking up information value that may be just too subtle for a human to detect to begin with but more importantly, just doing it a lot faster. And again, I'm oversimplifying. If my AI people are listening, they're rolling their eyes and like, my God, this guy's an idiot. But I promise you, we have much, much better people who actually design this thing, and the net result is improves our approvals roughly twice better than the next candidate in a traditional -- we have families of models that we've been training for 15 years now almost. Every time we launch a new model, we initially compare it in its ability to separate good risk from bad versus the prior candidate. And so when we look at the traditional model in this new architecture compared to the performance of the future candidate and this new idea, the new idea performed twice as well. It doesn't mean we're approving twice as many people at all. It means we're approving probably 2, 3 points more people, but those are people who are -- we are able to lend essentially with no incremental loss whatsoever. And that is an enormous booster to -- like what -- how do we get to this 30% growth? One, we build new features, new product. The other way, find better risks without taking incremental loss or incremental delinquencies and that's exactly what this model does with a whole pipeline of more models in that family coming out. We're pretty excited about them. Yes. And by the way, I strongly encourage you to read the blog post. It does a better job than I can in 5 minutes explaining this.

Your next question comes from the line of Timothy Chiodo with UBS.

I'm hoping we could talk a little bit about some of the progress being made with the adjacent verticals. The TAM that you gave at the Investor Day, it's multiples of e-commerce, and you've had, I know, a lot of progress with Intuit, ServiceTitan and many others. I was hoping we could just touch on that for a little bit.

Yes. You can see in the supplement that, by far, the outperforming growth segments of the ones that we break out other than the nondescriptive other segments, the best one is services, and that's exactly the brands you just named. And so that's going really well. It's still super early. Like that is a major -- every time we launch a new vertical, it takes us a bunch of quarters to figure out the perfect product market fit. You have to do it carefully. You can't just show up with a "hey, borrow money from me" sign and see what happens. But that's one way -- or too irresponsible, and we don't do that. So we're still learning, but we're very excited about the progress. We're seeing really good uptake. It unlocks things like home improvement, which we were never really a player in. It unlocks access to off-line services, which isn't really just a hard thing to access unless you have these platform partnerships, which we now do. So doing really well there. We are always looking for more. I learned not to preannounce, but this is not the last new segment we're going to go into in a big way.

Your next question comes from the line of John Hecht with Jefferies.

Really big progress, ongoing progress with more direct-to-consumer activity. But it also looked like there -- the merchant revenues relative to GMV were lower as a percent. Is there some tie-in there? Does the merchant discount rates vary from DTC versus directly at the merchant?

The short answer, John, is -- this is Rob. The short answer is, yes, it does. Most of our D2C program runs on either a physical card or a virtual card. And so in most cases, we're collecting interchange, and that interchange is going to be a function of the merchant where the transaction occurs and in some cases, the industry for the transaction as well. So on average, the interchange that we collect is probably lower than the average merchant discount rate that we would have through an integrated merchant. And so there is a little bit of downward pressure on that network revenue line. The good news is our direct-to-consumer business is very interest-bearing heavy, Affirm Card, for example, has historically run north of 80% interest-bearing loans. And so the all-in compensation to Affirm is on average much higher than the business at large. But it's that mix of merchant revenue or network revenue in this case that is going to be a bit lower. But we think that, that product still works really, really well for us, and we're really proud of the growth and the profitability that it's thrown off.

Yes, just to sort of -- I'll add a little bit to it just because I think it's important to kind of see the long-term version of this as well. As we integrate with more and more merchants, we're able to do what we currently call the universal financing programs, where the financing programs we offer to consumers at the point of sale, which sometimes is 0% financing, sometimes longer terms, but inevitably subsidized by the merchant or the brand they're selling, is also available to consumers on the card. So you can almost think of the card and really all of our direct-to-consumer products as a little bit of a trailblazing device. We go to a merchant, I mean -- and literally, this is a sort of a very, very true story. I personally and our sales team goes to retailers and say, hey, we have already driven a couple of hundred million dollars of volume for you using Affirm Card. But this is just half of the story. We can do amazing things for you. We can double your volume if you start participating in the 0% deals that we have and all the other tools that we've built like BoostAI, et cetera, that allow you to optimize your, basically, marketing dollars. And the way to do that is a direct integration. As soon as the direct integration happens, you will see those 0% deals show up not just at the point of sale but also on the card. And we're investing in the card even more, making exclusive card deals happen. You'll see a bunch of that in this quarter we're in now. So today, what Rob described is exactly how it goes. Over time, each merchant where card gets a lot of use will get a call from one of us saying, hey, you really should integrate directly. You will get the other half of Affirm product, and it's amazing. And we're more than successful [ in doing it ourselves ].

Absolutely. That all makes total sense. And then just I'm wondering what you learned so far in terms of the card usage. Like how much of it is, I guess, the in-store component of usage? Does that change the transaction type? How are people making payment decisions? Are they paying in 4? Are they migrating it to a loan more commonly? Maybe just any insight on what you're observing at the transaction level with the card product?

Sure. The fraction of in-store transactions, I think, remain pretty stable around 30%. The frequency is a lot higher, but sort of we all know that. It remains quite high. The preplan, so you create a payment plan before you swipe, before you tap. It's still the dominant use case. So consumers are using us with a card similar to the way we use at point of sale. But there's definitely a non-0 element of post-transactional splitting where you first transact with a card using it as a debit card, which is what it is underneath, and then choosing to turn it into like post facto, which we explore as well. So it's a little bit more, let's call it, diverse usage pattern, but fundamentally, it's access to Affirm products and Affirm programs with a piece of plastic.

Your next question comes from the line of Bryan Keane with Citi.

Just wanted to ask, Rob, a little bit about provision expenses as a percentage of GMV went up. Delinquencies went up a touch. I know that's probably not a big deal. I think you guys made some comments on it in the shareholder letter, but I just wanted to ask about that. And then on the GAAP operating margin in second quarter guide was a little below Street. I guess it was similar to the guide in the first quarter. Just thinking about the pluses and minuses in GAAP margins as we go to second quarter from first.

Yes. Sure. I'll take them in order. I think in terms of the provision dollars in the quarter, we were in a position where we did choose to run some experiments within the approvals and the origination, and we made the decision to keep those loans on balance sheet. And the experiment itself was pretty localized to the month of September. So those loans are on the balance sheet. They're showing up in the allowance because they're experimental. I think they're more than properly accounted for in terms of allowance. So that did have the impact of driving a little bit more provision expense, a little bit more allowance in the quarter. But when we look at the profitability and the revenue less transaction cost take rate that we were able to drive in Q1 of 4.3%, we felt like that was the right thing for us to do, and it was overall accretive to the network and helps drive on the new user front, particularly. So we feel good about that decision, and that -- the byproduct of that was a little bit more provision expense in the quarter. But again, that felt very comfortable just given the profitability we were seeing in the first couple of months of the quarter. And then on the GAAP operating income on the guide relative to external expectations, I mean, obviously, I think on our side, some of the GAAP-only expenses that would be the drivers between adjusted operating income and GAAP operating income, I think as I've looked at some of the external models, I think maybe there's a little bit of under modeling around the warrant expense in particular. As most folks know, we have a program with a large partner where there's a performance component, and that partner earns warrants through new-to-program users. We're seeing actually a nice trajectory of growth in new users through that program, partly driven by some international expansion that's happened. So I think that's an area where our internal forecast is maybe a bit heavier on expense than the external estimates. I think, ultimately, that's a really good thing for the business. We really are excited about getting this toehold in the U.K., and I hope that it's the beginning of several successful international markets. So again, I think that's the one area I would point to that maybe the Street has under modeled for us in terms of expense. I think that's driving most of the delta in the GAAP operating income margin.

Is there any seasonality in that in first and second quarter just based on different trends in volumes.

I think the seasonality probably looks a bit like retail. And then there is also this maybe added complexity where in a new market, we see a higher rate of new users than we would in a more mature program or a mature market. So I think it's a combination of those 2 things. But yes, there's definitely some holiday period seasonality.

Your next question comes from Harshita Rawat with Bernstein.

I have 2 questions. So one, I think in the letter you referenced kind of bank account for down payments instead of debit cards. I guess my question is what are the challenges to getting more borrowers to do that, I guess. And also, like, I think, help us kind of quantify some of the kind of savings. I know obviously, you save on interchange with debit cards, but there's also some costs associated with bank accounts. And then second, maybe talk about digital wallets. You referenced that in the letter. You're live with Apple Pay, Google Pay. What have been the learnings so far from the rollout as well?

Yes. I mentioned the down payment with ACH, a little bit in the same context as the AI gateway. Like there's a lot to save still inside our system. We're fairly well optimized in terms of managing operating expenses. But at our scale -- there's some serious fire engines going by. Hopefully, you cannot hear it. They're clearly in a hurry. So anyway, so these are -- and I guess one way to think about the -- the way to analyze these savings is the approach in anyway. Every time you switch from a real-time network to a non-real-time network, you are introducing an element of risk. We are very good at risk management. At a certain scale, you have enough data where you can essentially self-insure using debit card networks to take down payments. You buy -- you get a little bit of an insurance of good funds, but you're paying interchange. By switching to ACH, you are introducing a delay. You don't necessarily know you're getting good funds in real time. But if you have a really good underwriting model or you're able to build a specialty model for that particular form of payment, you are saving money and effectively just putting it towards giving consumers better deals and saving on OpEx. So that's kind of the general thrust. And I mentioned one, but there's many such initiatives going on. And they're always going on. This one was just particularly dear to my heart because it was a long time in making. And then on the wallet side of things, we are -- we were trying not to talk in great detail about any one particular product, so I'll keep it general. It's been a -- generally speaking, it's been very accretive. We find that folks that prefer digital wallets really stick to them. So having a direct integration is a great unlock for repeat usage. We definitely get some new users through those integrations as well, but the most compelling kind of benefit for those deals is always just being there when the consumer is. If you look at some of our merchant partnerships, you'll find our logo. You'll see many digital wallets right next to us. And any door you go through, you will find Affirm available. And some folks love going through our flows. Others like to do that inside a digital wallet provider, and we see it as an important part of our business to be in all those places. And so that's been kind of the most important thing. And that strategy was something we said we were sometimes -- as we embrace channel conflicts as sort of a shorthand for this and we've embraced it for a little while, and it's paid dividends. We're seeing those users come back through those doors over and over again, and they become accustomed to finding Affirm exactly where they prefer to shop.

Your next question comes from the line of Will Nance with Goldman Sachs.

I was going to dovetail off that last one there, at least the answer. It was great to see the DTC and wallet driving around half of the volume this quarter. And when we think about ways to kind of drive that even higher over time, obviously, you have things that you can do that I would kind of put in the network effect category, adding more merchants, more nodes to the network and more places where you can use Affirm and that just driving good things for the overall business. If we think about it more from an issuing perspective, what are some of the things that you're contemplating just to enhance the overall experience of the Affirm user and just drive kind of more repeat usage, more habituation, I thought the example that you just gave on the wallet partnerships was great. Anything you could talk about just on kind of user experience and kind of things on the road map there?

Never preannounce the road map. It takes longer than you think. I'll bite, but I'll try to keep it not overly disclosive. So there's a lot of work being put in right now into the card. Like we really think it's the best way to use Affirm or it should be the best way to use Affirm. We think it already is -- there's so much more we can do. The card should be unequivocally the best, most protective, least difficult. It should be the thing that you prefer. And the reason I'm so excited on that is because it fits naturally into both kind of normal ways "point-of-sale integration" as a natural process of your card, also fits natively into all the digital wallets. And so we're very, very focused on that product, just continuously adding benefits to that. So there's a lot of work being done. You'll see some of it actually launching literally in days, if not weeks. So you'll see the answer -- an answer to this question in a very immediate future. There are some card-specific financing programs coming. There's a card-specific Affirm offers that are going to start appearing that will make the Affirm Card the preferred way kind of just through the pure financial benefits to consumers. I'll stop there, but I think that's kind of the -- that's the thrust. Some of this is, by the way, what we internally call the mainline development. And then there's even more coming a little bit later from the zt1 team where we're really thinking in the sort of the next iteration of the card.

Your next question comes from the line of Andrew Jeffrey with William Blair.

Max, you always speak so eloquently about credit and your ability to -- I think you talked about the we take risk comment in the letter, which is great, but credit and the ability to manage risk and increase approval rates across the credit spectrum. I wonder if a little bit of tonight's reaction to another really good quarter is just the -- Michael explained the delinquencies in this quarter. But just a reaction to the higher provision and concerns that in a more uncertain economy, credit becomes a problem or credit weighs on RLTC. And I wonder if there's just a misperception fundamentally about Affirm's ability to compound GMV at 30-plus percent while maintaining the excellent credit outcomes we've seen.

I'll start, but I think Michael is the one who's going to give you the really eloquent answer, but I'm going to steal a line from him before we get anywhere. So I'll tell you the story. In May of 2020, we were not a publicly traded company. We did not have a lot of cash, and we had no idea what this COVID thing was going to be all about. The stimulus checks hadn't arrived yet, and Michael and I were staring at our book asking a question, if the world goes fast with hell of a hand basket, no one can pay their bills, what are we going to do? And he gave a very picky line that remains true to this day full 6.5 years later. This was in May that whatever happens, by August, it won't matter because the entirety of the book at risk will have been paid off. We lend very, very short weighted average term. You can -- I'll leave it to Michael to repeat. I think actually Rob was the one who explained that a lot of the delinquency changes are entirely seasonal. The year-over-year change is irrelevant. The provision update has to do with some -- we think is very intelligent experimentation. But it's all entirely in our control. But the bigger point, if you don't believe any of the eloquence that I bring to my storytelling around modeling and risk management, just note that the average term is still very, very short. We get to deal with whatever overexposure we might take on today in less than 90 days. So that should give a lot of people, I think, great comfort in our ability to recover from whatever mistakes we might make. That said, like 1/3 of the company is involved in some form or another of managing credit. Credit is job #0, used to be my call sign in every one of these letters until I got tired of repeating myself, but it hasn't changed the way we run the business. So if that's the reason the market misunderstands us, let us stop us this misunderstanding.

And then just 2 things to add. First is the credit outcomes that we are getting now and have gotten over the past 6 years of our public life have been in line with what we wanted to have. Our credit outcomes are a choice point that we make in our business, where we want them to be. Everything we do is intentional with respect to credit. And we can do that because we're looking at hundreds of millions of decisions every quarter when we're making underwriting decisions. And you can tune those models to the appropriate amount of losses you'd like to have based upon that just incredible volume of decisions that you're making in the quarter. The second is the capital markets have spoken very loudly on this topic, whether it's the ABS markets who both understand, appreciate and have the first risk in those -- ton of our loans but also the forward flow market, really sophisticated, world-class, blue-chip loan buyers who understand the credit losses in our business exceptionally well continue to be extremely interested in exposure to our assets. These folks are, honestly, some of the smartest credit investors in the world, and they are still extremely excited to participate in our assets. The ABS deal that we are in market right now on and priced today was over 4x subscribed despite doing it in 1 of probably the worst -- 1 of the worst macroeconomic -- certainly in the equity capital markets days we've had since COVID. To be able to price a deal like that without volatility, you can't do that without that market giving you 2 big thumbs up on your ability to control losses.

Okay. I appreciate that very much. Sorry about misrecognizing you, Rob.

Yes, [ it's a ] compliment. We believe that [indiscernible] if you've been tuning in for a while, you know how we feel about credit and how much we care. It sometimes frustrates us that not everybody has gotten the message yet, but we'll just keep compounding at 30% and keeping the [ DQs ] in the right range. And at some point, it will become apparent.

Your next question comes from the line of Nate Svensson with Deutsche Bank.

Max, one for you. I was hoping you could talk a little bit about the time you're spending now with the zt1 team. Clearly sounds like it's positive for your mental health. But I guess, in terms of what your role in that team opens up for Affirm as a whole, now that you're able to refocus your time a bit and I think in one of the prior questions, you said you're not going to preannounce road map, so I won't ask that. But I think more of a flavor on how you're spending your time with that team, what you're working on would be helpful.

Sure. And I know some of them are listening. Shout out to my zt1ers. They are a genuine delight of my daily grind here. So I guess, some sense for what we're actually working on. So I'll give you a -- this is maybe a true look behind the curtain. So the MO that I set with zt1 roughly try to remind us what we're here for every quarter or so, is to work on projects that have either a risk-adjusted or probability-adjusted path to $1 billion in revenue in a span of 5 years from launch or to move the needle to the tune of 15% to 20% on an existing line of business that is already $1 billion. In other words, we either want to take a bet on an existing thing that we believe Affirm proper is misprioritizing and we have both the audacity and the capacity to try to move the needle there or just frequently -- in fact, more frequently, we actually want to go into kind of the 3 big approach where we build something, and we have no idea if it's going to work or not, but we're going to try. And if we see a heartbeat, we go all in and build it until it gets to a point where we hand it off to the main development team and then they pick it up. So one practical thing. This is not a big reveal because we talked about this product before. So every once in a while, we mentioned BSPs or brand-sponsored promotions. This is the idea that if you are a multi-SKU general merchant, you're probably selling a lot of things that you are not marking up more than 10%, maybe 15%. And so when Affirm comes in and says, "Hey, you ought to do a 0% sale. Big Nothing is a great idea. Black Friday is coming up. 0% drives the volume. Let's do this." If you are a vertically integrated seller, like, of course, sounds like a great idea. Here is my 6.5% to 9.5% MDR. Let's give lots of people access to [ those ]. If you're a general merchandiser and you're reselling a branded product that someone else manufactured, your markup is basically the ask that we're making of you for the MDR. The 0% is not going to happen unless you're just trying to have empty calories. However, the original equipment or the original item manufacturer retains a lot of margin. Many of them are motivated to put their marketing dollars to work. In fact, they already have existing relationships with the retailer, and so helping the retailer and the brand tie all those funds, which are typically plentiful, there's a lot more margin to go around, which usually gets spent in sort of silly promotions like driving -- did you know there's a 20% sale at retailer X on brand Y. There's an extremely, extremely sophisticated way of doing it. A much better way is did you know that this particular SKU is on a 0% loan at exactly retail price. There's no discounting, but this retailer has this brand at 0% available right now. So all of this sounds kind of obvious. It's braindead that nobody has done it. The reason no one has done it is because it's really hard to scale. This actually happens in retail today. We didn't invent the genre. If you talk to any major player, any general merchandiser, they'll tell you, yes, we do this all the time. It's called vendor promotion. It's vendor-funded promotion. So you go to the vendor and say, "Hey, give a couple of hundred thousand dollars. I want to put you into my circular. It's going to go out to people." And it works okay, but it's run off of spreadsheets and people have to make phone calls. And it's quite a lot of manual labor, and so they do this maybe once or twice a year. And it's, generally speaking, applicable to every item for a brand, and it's very, very undifferentiated. We've been building a system that allows for all of this to be managed basically entirely electronically where you can really have a large number of SKUs at any given time with these brand-sponsored promotions. This requires things like budget management because nobody wants to run these promos forever. They need to be launched and wound down at exactly the right time. They have to hit at the same time as marketing promotions that happen. Typically, within the general merchandisers, retail media networks, all of them now have their own advertising arms. So this has to echo through that and reach the consumers that way. So slumming all of that together and making sure it's still very profitable for the retailer is just an incredible amount of work. That's what we've been building at zt1. If you want to see what that looks like, if you actually play with the Affirm app, you will find that we've scaled DSPs out very significantly. I'm not going to drop any names in this call. It will be a good scavenger hunt, but there's about 1 million different SKUs in the Affirm app that are brand-sponsored -- funded. So that's very real. No, it's not anyone's forecast, discount heavily everything I say here, but I'm personally very excited what this will do for Christmas sales. And hopefully, we'll scale it out to hundreds of millions of SKUs before too long. Hopefully, half of this made sense. This is like -- so if you want a flavor of what it's like to work with me, these rants occur in zt1 on either regular basis.

Your next question comes from the line of Jason Kupferberg with Wells Fargo.

All right. Great. I wanted to just come back on GMV growth for a minute. The 34% in Q1, obviously, real strong. We're doing at least 30% according to the new guide for the full year. So on the surface, that implies some deceleration during the balance of the year. I know your compares get a little easier because you're -- got past the Walmart headwind, et cetera. So I mean, any factors that would actually drive a slowdown off of these Q1 levels? Or is this just the typical conservatism that we've come to know and love?

Yes. I mean, obviously, we've given a floor for growth for the full year at about 30%, so it's not a range. There's no ceiling to growth. We're doing everything we can to grow the business responsibly. We're seeing growth rates early in Q2 and certainly in September as well that were much north of what we posted in Q1. So no reason to -- nothing to point to specifically in terms of a deceleration in the business. We just wanted to set a responsible floor for growth for the full year. And that's what we did with FY '27 guide.

Okay. Understood. I also saw in the shareholder letter you've got the first Affirm Edge pilots are going live in the first half of the next calendar year. Talk a little bit more about how many banks that will be with. What's the general size of these institutions? And any other color because, obviously, based on what you talked about at the Investor Day, it seems that could be pretty exciting over a multiyear period?

I'll take this one. This is Michael. We're really excited about Edge. It's a chance for us to distribute what we know is a better way for consumers to access credit to a really very wide audience by partnering with the issuers of debit cards across banks of all sizes. It is also really early. We've been hard at work, working to develop the product. We're like really close to being wrapped up with some of the development work. And then we will need to figure out the right set of partners to get live with. Nothing to announce here today, but I think patience is the code word here. We have a set of partners, both the platform partners as well as the end customers here who don't move exceptionally fast. And we have to be really mindful of just the pace at which they can move, and we have to pace ourselves to them. And they're going to govern that. So it's going to take a little bit of time. Don't have lots of partners but are and continue to remain very excited about the products.

Your next question comes from Dan Dolev with Mizuho Securities.

Appreciate it. I want to ask the question about agentic commerce in a different way, Max. Like maybe just help us understand, and this is what everyone has been trying to figure out. So if a merchant cares about agentic conversion, shouldn't they subsidize more Affirm 0% loans or any other Affirm loans? Or does it not make a difference? Like what's the key hook here for the merchants in an agentic world?

Couldn't have written it better myself. Yes, they should. And part of why we've been so hard at work on things like these fees is to create programmatic APIs. So initially humans with calendars and tools but soon enough, agents on their own can deploy dollars to compel a buying agent from their side to say yes. Agents aren't going to be shopping on your behalf. They'll certainly be looking for deals on your behalf. Affirm is really the only scaled player that's capable of managing longer-term 0s that are differentiated, that are driven by SKUs, that are driven by targets or targeting. And we should expect at some point -- like this is very, very early. I just saw or heard Dara talk about how there have been very few agentic rides booked on Uber. So it's all very, very early days. But it is coming. It's coming. It's exciting. I'm sure all of you have tried all the various agents, and some of them are really, really compelling now. Each one of these will be programmed to find the best financial transaction. It's going to be a combination of cost, which is going to matter, and 0% is going to be the best cost because there's no cost to borrow. It's also going to matter that they are fully disclosive, that the price doesn't change. If you were somehow a day late, that -- it doesn't compound, so you can estimate your dollar, so all the things that we talked about as really, really important. A high percentage of our consumers pick us because of those features. Some pick us because we're there and they don't necessarily grasp the full kind of value that they're getting by choosing Affirm. 100% of all agents will grasp exactly the value they're getting. We do offer the best credit product there is. That's actually true. And so as agents find us, it will help merchants sell more. Merchants can help merchants sell more by funding more promotions. And so I expect 0% volume to go up a lot. In agentic, I expect agents to care about the ever-changing catalog of 0% loans or 0% deals inside the Affirm ecosystem. I think that's a really important service that we're going to provide to every agentic shopper tool out there. So there's a lot to like about where this is headed. But again, like I would just caution anyone to create a spreadsheet that shows going to the moon [indiscernible]

Your next question comes from the line of Dan Perlin with RBC Capital Markets.

So I just wanted to follow up a little bit on that actually, but from the context of like cybersecurity investment cycles, it's definitely becoming a little bit more of a topic this quarter and especially in a world that you just described, Max, around increasing agentic opportunities. So the question is how are you guys thinking about where you are in that cycle. And are you able to throttle up in or down given potential risks associated with that. And then the other part of the question is, is there any risk to you guys in terms of ownership of liability to the event that you start to see more agentic fraud as a result of pointing in your direction. I don't know the answer, so I'm just trying to just kind of figure out what kind of, I guess, position that...

It's a great question. So on the cybersecurity side of things, obviously, we take it super seriously. This is not -- and by the way, so should everyone if you just -- like there's a meme going around. I wake up, there's another vulnerability. That is the life of our security team and every security team who has a job right now. And so I think we're all constantly watching out for new disclosures. We're trying to keep up with all the patches, trying to make sure that we're keeping our data safe, our users' data safe, et cetera. So -- and this will be a -- I'm generally speaking quite bullish by way -- we're definitely not in security, but it is a little bit of my past life, so I have strong opinions. I think defenders are getting access to better models quicker than attackers. And so we will go through an unpleasant period of patching and fixing and learning how to write better software. But in some fairly short period of time, we're going to just be better software builders with AI and generally speaking, better protected. So we're in a little bit of a shooting war right now with these bugs that are being discovered every day, but it will end hopefully relatively quickly, and then we're all going to be better for it. So not unconcerned, not even a little bit. In fact, I text our security team in all kinds of hours saying, "Have you guys seen the disclosure yet? Like what are we doing to make sure we're not affected?" So that's daily life right now. But we're technically very, very strong, and we're definitely benefiting from the excellence of our engineering and in particular, security engineering teams. On the fraud and liability side, that is a fantastic question. Generally speaking, you can't expect agentic commerce to be more than a little bit messy in the early days. Your agent will hallucinate. You must have said yes, and here's an 85-inch screen TV for your tiny 1-bedroom apartment, and we don't care what you're going to do with it. You bought and it's yours. And you'll have to return it, but the headaches and the hassle will be there. So as all of that gets figured out -- and by the way, AI is improving very quickly. Your agent is going to learn. The models that power your agent is going to learn very quickly, including from the mistakes your agent made. So all of that is also going to resolve itself in a reasonably short term, but the messiness of consumer interaction with agentic purchases, unwanted purchases not as described and then as, correct, very aptly pointed out, agentic fraud is going to be a thing. Generally speaking, if there's risk to manage, we like that world better than the world where there isn't. We are very good at it. We're very good at capturing the right data. We're very good at rapidly scaling model building and training things that allow us to mitigate price and ultimately manage risk. And so I don't want to say I'm excited for the new challenges that are coming our way. But to some extent, every time there is a disturbance in the force of payments, whether in a form of risk or rates or fraud or credit or new modalities, new channels, we tend to win because the solutions are inevitably engineering and AI and ML-based, and we have very strong advantage than the rest of the industry in that space. Just give me something to build is the shorthand answer.

Your next question comes from the line of Matt O'Neill with Bank of America.

Well, maybe, Max, to cherry-pick a word from the shareholder letter, it would be fatuous of me to double down on some many of the questions that have already been asked and answered. So maybe for a closing remark, is there anything that none of us managed to ask that you would have liked to have been asked or you'd like to conclude with?

I applaud your attentive read of my Lebowski references. That's one. There are several more. No one ever asks how many there are? There are many in this list. Some of them are very deep cut. So that's a good homework, assignment if you want to take your mind off the bloody market. I think most of you really understand the business pretty well, and we're always happy to engage in these calls because it gives a chance to brag a little bit and the questions have gotten better and better and smarter and smarter over the years. I don't know, Michael and Rob, if you guys have something that should have been asked. I guess -- I feel like this is -- the last few quarters, if you want any feedback, analyst community, you've done really well. Like we look forward to these conversations. We like the questions. We feel appropriately challenged, and we don't roll our eyes at any of them.

And our last question comes from James Faucette with Morgan Stanley.

Wanted to ask on some of the larger ticket sizes and transaction opportunities, thinking about what you can be doing with ServiceNow -- or ServiceTitan, excuse me, and others. How should we think about how that may impact MDRs, et cetera? It seems like 0 percentage, you were talking about, Max, could be really attractive part of the sales toolkit for those kinds of service providers. But I think one of the things that we grapple with is like what kind of impact would we have on MDRs and the overall economics of transactions compared to what you're doing today.

I mean, our large ticket items, large ticket bills, services included, is our bread and butter. There's no one else in the industry that's figured out how to do that at scale in a differentiated manner. So we're always excited for more of that volume. You're totally right. There's a really healthy demand. I remember when I answer, the largest transaction we've ever written was $7,000, and that was like, "Oh my God, I can't believe I said that out loud." And that is not even close to our largest transaction. So we are carefully, carefully increasing our aperture for risk for only the deserving borrower or only the typically quite high credit quality. But there's definitely room there. It's actually hard for me to immediately reason about MDR changes because there's 2 opposing forces. The merchant is -- if you're selling an HVAC system, so making it up as I go, but typical HVAC is like $20,000-$29,000 range. That's a typical sort of replacement for a single-family home. A merchant trying to move that merchandise is more than excited to pay a substantial MDR to make the answer as compelling as possible. You don't want me to come back and shop again. You just want me to buy this thing. And so MDRs are naturally easier to get from merchants because they understand that they're only going to sell a dozen of these per quarter. One of them not being sold makes a huge difference to their bottom line. So specialty shops that sell these things are particularly drawn to 0% promotion. So all of that is really, really good for MDR. On the flip side, the buyer, given the size of the transaction, is actually typically very low risk. And so from our point of view, to extract a reasonable margin of safety on the loan, the requirements are meaningfully lower. And so you kind of have dual -- 2-directional pressure in each direction. And I don't know if I can instantly handicap whether this means the net result is higher or lower. It feels like it goes a little bit higher but maybe not quite as high as a 4-year 0% loan would compel or something like that.

Yes. And I think for a larger ticket transaction, James -- this is Rob -- we would also go longer on duration. And so in terms of the overall total revenue quantum that we would get from the transaction because it is going to be a longer duration loan, it's probably on the higher end of the spectrum for us as well just given there's going to be more funding costs associated with that loan on our side.

This now concludes our question-and-answer session. I would like to turn the floor back over to Zane Keller for closing comments.

Well, thank you, everyone, for joining today. Apologies again for the technical difficulties at the beginning of the call. And thank you for your time, and we'll talk to many of you again shortly. See then. Talk to you.

Ladies and gentlemen, thank you for your participation. This does conclude today's teleconference. Please disconnect your lines, and have a wonderful day.