Shopify Inc. (SHOP) Earnings Call Transcript & Summary

Excellent. Thank you, everyone, for joining us this morning. I'm Keith Weiss. I run the U.S. software research franchise here at Morgan Stanley. And really pleased to have with us from Shopify, both President Harley Finkelstein; and CFO, Jeff Hoffmeister. Jeff and Harley, thank you so much for joining us.

Always great to be here.

So a lot to talk about in Shopify. You guys came off an astounding quarter, and it's sounding a year really in terms of overall GMV growth, overall revenue growth and good profitability. But there's one topic that everybody really wants dig into and it's Agentic commerce. And it makes sense, right? And that's Shopify has led in a lot of commerce transitions in the past and has benefited from commerce transitions in the past. We see a big pace of change on a go-forward basis and investors are trying to understand how that change is going to evolve. So Harley, maybe we could start out at a high level and just give us the Shopify perspective, like maybe the Agentic commerce 101 of how the e-commerce experience is going to change, how Agentic commerce is going to roll out over the next couple of years? What should we be expecting to see in the marketplace?

Yes. Thanks, Keith, and thanks, everyone, for being here. Maybe I'll start kind of a really high level and I'll go kind of get down to the ground level. I think at a really high level, first of all, just to level set for those in the room that are not as familiar with Shopify, we power but, at least last year, we powered about $380 billion of GMV. So we are about -- just in the U.S. market, we're about 14% of e-comm in the U.S. market. So we see things, I think, at commerce is sort of a terrestrial level and an atmospheric level, I should say. And I think the business model of Shopify is actually very simple. We make money when merchants make money, the more they sell, the more we make. And that business model has worked really well for us for almost 2 decades. From a product perspective, the product actually -- we do a lot of different things, but the product in itself at its core, is very simple. We make it really easy to build, to sell, to manage all of your retail, all of your commerce operations regardless of whatever channel you're selling on, whether that's online or offline or on social media platforms and marketplaces. And we think Agentic actually could be one of the most exciting new trends or new paradigms for commerce, maybe since the Internet. So when we see Agentic commerce is that we think we see it as sort of this new front door to commerce that for those -- for consumers, it's going to act as sort of this personal shopper. It also has actually the opportunity potentially to increase the TAM of e-commerce by bringing some of those laggards who typically don't buy online into kind of the digital realm. But ultimately, Agentic commerce at its core is like having some sort of personal shopper. And it's going to help you do research, it could help you compare -- it's going to help you potentially even transact. And whenever there's a new kind of front door, a new surface area for commerce, that's actually when Shopify's value increases. And 11 years ago, Jeff was actually our banker at Morgan Stanley when we did the IPO. We talked about what we were trying to build to Shopify is this retail operating system. This idea that there's going to be all these different spokes where commerce can happen. And the more spokes that exists, again, online, off-line and social media everywhere else, the more the value of a centralized back-office system of record matters. So the more front doors, the more the house actually matters. And ultimately, that's how we see Agentic. We see it as being a really potentially really wonderful new front door. Now that being said, we think that for anyone who is going to lead in kind of this Agentic commerce era, 3 moats are going to be really, really important to do this really well and to be at the epicenter of it. The first moat that is going to be really important is going to be the transaction layer. Now that is not just the payment. That is the checkout itself, that is subscriptions, that is the inventory, that is shipping taxes, all the different merchandising options that come with any type of transaction. And we do that really, really well. The second moat that we think is really important sort of this Agentic commerce era is having the data layer, this proprietary data layer. And the way that Shopify is set up, we have millions of stores, millions of merchants. We also have billions of transactions and trillions of data points, which means that over time, our system actually gets smarter and it sort of creates this flywheel effect. The third moat that we think is really important for the agentic sort of era is this network effect layer, meaning the more that happens, the smarter the system becomes. And for us, again, back to the business model, more checkouts mean more merchants come on, means more customers are buying mean more merchants decide to come to Shopify means we're able to make more GMV, we're able to invest more in our product. And this flywheel really starts to spin. And in all 3 cases, we think Shopify has an unfair advantage around these 3 moats. If you take it actually down to the Shopify layer itself, the business of Shopify is actually -- and many of you in the room, you especially Keith, know Tobi. We've sort of always been looking towards the future of where things are going. I mentioned on the road show, we didn't describe Shopify as an e-commerce provider. We called ourselves a retail operating system and some of you in the room gave us a weird look to say, well, what exactly does that mean? Well, we were anticipating that commerce was going to be done everywhere and the company and product that brings it all together, that organizes it will be the most valuable company in commerce. So in many ways, we've sort of been planning for this for probably a decade. And in terms of the actual financials of Shopify, the fact that we can grow 30% top line revenue and have free cash flow, which was about $2 billion of last year, means we actually -- we can see this out for not just 5 years, but for decades. Now at the ground level, I will say that in the last 2 years or so, you've seen us create a couple of different products, which on their own may not have seem profound. But when you aggregate them, you begin to really see the picture of what we're building with agentic. The first one was this idea of agentic storefronts that directly from Shopify's admin, any of our merchants can click a button and say, push our products to, we have 3 main partners now, the first one being Gemini, Google and AI mode. The second being OpenAI with ChatGPT and more recently with Microsoft and Copilot that with a single click, our merchants can push their products and syndicate them to every agent service. The second thing we built, which we built with -- we announced about 5 weeks ago at NRF was UCP, Universal Commerce Protocol. What we're trying to build with UCP is a way an open-source language every agent tool to speak to every one of our merchants or every merchant in general, so that the experience on the agentic application feels just as good as it does buying from an online store. And then the other one we built was this thing called the agentic plan, which in a similar vein to Shop Pay off platform, this allows non-Shopify users to actually begin to use Shopify. And then there's some really other cool stuff which you can get into like Sidekick and SimGym, which we think is the greatest co-founder ever created for every one of our merchants that knows everything about Shopify, knows everything about your business that can help you make better decisions. So when you kind of zoom out from all those things, starting with the agentic kind of the meta store and then exactly what we're doing at the ground level, I think you begin to see that we've really been building the rails for agentic now for a number of years, and we are really well positioned to take advantage of it, whatever permutation may actually be the one that ends up being the go forward. So in many ways, this is kind of what Shopify has been built for. And I think it's probably -- I mean, anyone who's been studying it, is one of the most exciting times of retail over the last couple of decades.

Excellent. So we had this framework in Morgan Stanley, not more so the team software, we call it our best athlete framework and Shopify was our first best athlete. And one of the attributes that we align to is a speed and speed of innovation. And you sort of -- to talk a little bit about that with some of these new products. So from a fundamental perspective, Shopify has always been about enabling these new channels for your end merchants. Merchants need some off-line capabilities, you develop that for your merchants. Social commerce becomes a thing, you develop that. Can you dig in a little bit into the initial innovations like the new product innovation that enables the agentic commerce product catalog, you talked about, there's universal checkout that you talked about. Where are we in terms of like are these products ready for prime time today? And two, are they being adopted? How broadly are they being adopted by your end customers?

So I mean the products are ready, the product out that we probably have about a dozen merchants that are already using it. The only reason it's gated is we're just waiting for the agent applications to continue to open the doors. I would say, in the next couple of weeks, you will see thousands of more merchants end up being live on these agentic applications. And then over the next couple of years, maybe in the next year or so, even you'll see millions. So the products are out there already. If you look at companies like Skims or companies like Gymshark or companies like, what is a good example, Fashion Nova, for example. We probably have put a dozen stores on Shop that are already live on these agentic applications. And effectively, running all of those things, UCP. The catalog has been fully built, which, again, provides an aggregated every single SKU on Shopify price, an aggregated collection and delivers them to all these agentic tools. So this thing is ready. I think what you'll probably see in the next 1 to 2 years from Shopify perspective is you'll see more agentic source GMV happen. I mentioned this on the earnings call, but maybe it's worth repeating. If you look at traffic to Shopify stores coming from agentic applications, if you compare January 2025 to January '26, it's about 15x. It's on a very small base, but you're already starting to see these thing begins to grow quite a bit. So I think in the next year or 2, you'll see more agentic source GMV, you'll see tighter integrations from Shopify to all the agentic players and you'll see our own products develop even further back to that moat that I was talking about that proprietary data moat, things like Sidekick will get even better. So it will be able to predict the first thing that a merchant should do and the tenth thing that they should do. It will predict where they should advertise. It's going to predict what the next product you should sell to any customer. So I think while it's still very, very early days, the ramp-up of this is unlike what you described with social commerce, which I think took a little bit longer. Social commerce ended up becoming more of a discovery engine rather than a checkout tool. But what we're trying to build is this real -- this antifragility in that whatever permutation ends up becoming the standard for Agentic commerce, Shopify and our merchants are best positioned.

Yes. And Keith, the only 2 points I'd add to Harley's comment. Part of what we're doing is agentic is enabled by catalog, right? And we set catalog in motion more than 2 years ago. It's also enabled by Sidekick, which we set in motion more than 2 years ago. And I think that's one of the things that we've done for a very long time is be able to kind of look around 2, 3, 4 corners, see what merchants are going to need and then start building that before people even realize they need it. And so everything that we're going to merchants right now in terms of the agentic plan is a combination of all these things.

It also helps on the agentic side. The fact that I think we've become the core partner for so many of these -- so many of the most important genetic application companies is because rather than showing up with a presentation or some sort of documents, we show up with a working prototype. We showed them catalog years ago. We showed them agentic store fronts years ago. We started building UCP last year, and we announced it on stage with Sundar at Google in January 8 at NRF. So a lot of the things you're seeing now, yes, we're only starting to talk about it now, but we've been building these rails for 2 to 3 years.

Right. So product catalog, I think, is relatively easy for us to understand. You guys are presenting the products of your merchants in a way that agents can -- will discover and kind of utilize in their algorithms. UCP is a little bit more complex. Can I start with a dumb question? In terms of 15x increase in overall agent e-commerce, but UCP only rolled out in January of this year. Like how was the transaction taking place last year, right, that you're 15xing against? And what is UCP change in that? Like how does that evolve that agentic transaction for the end customer?

So if you go back 6 months ago, and any of you using any agentic application typed in looking for sneakers, you would effect you see a photo, which would then click into the online store. It was very rudimentary. I didn't talk about any of the attributes of it. So one of my favorite stores on Shopify is AG1, Athletic Greens or another good example may be ButcherBox. These are incredible businesses on Shopify that have been built -- they're category leaders. But if you actually look behind the curtain of these businesses, the subscription side of it is incredibly sophisticated. Some people doubled their subscription 1 month or if they're on vacation, they may want to pause it. They may decide that they want to purchase in the loyalty program. In other cases, there's bundling, for example. So if you bought this particular product, you may also want to buy these products. The original iteration of Agentic commerce didn't anticipate any of those nuances and actually, commerce on the agentic application has to feel as rich as on the online store. And so we built this protocol with Google because we felt that what was missing from the discussion was the experience. It's not just about a checkout or a credit card of transaction. It's actually but all the other attributes of it. Another example would be discounting. If there are merchants who feel a particular time of year they want to add a coupon code or I mean, loyalty is massive for a merchant in about Alo or Vori or On Running or some of these great -- Mattel, for example, some of these great Shopify stores their loyalty programs are a huge part of their business. And without a proper protocol that anticipates the actual nuances of commerce, it becomes just a dumb transaction. And so we feel like our positioning, again, we have 14% market share in U.S. e-com, the only check out larger then Shopify in America is Amazon. We feel like we had a position in the -- we were in a position to go ahead and build a protocol to better anticipate what commerce will become when it's fully fledged on agentic.

Got it. So UCP enables you to really take the richness of the commerce transaction that you've already enabled on the merchant's site and bring that into ...

And then open the sources. So even if you're not on Shopify, you can be used there. And actually part of -- that's the position we've always taken. Even when it came years ago when we were beginning to see what was happening with things like Stablecoins or some of the new payment transactions that were happening, we've tended to always been at the forefront of it because we simply understand commerce in a deeper way. And I think that gives us the right to at least go to the table and say, we think this is the way it should be done.

Got it. So OpenAI and Stripe came out with an ACP or Agentic commerce protocol. Can you help us understand UCP, is it a competing standard versus ACP? Can the 2 work together? How should we think about that dynamic? And on multiple, I'm thinking back to VHS Betamax not very many people in the room understand what that is. But like is there a potential for competing standard to potentially freeze the market and as we understand how like these are going to be built in the future.

Yes. No. From our vantage point, UCP is -- they're not competing standards. And I think to a large degree, in terms of where commerce is going to go or agentic commerce is going to go there's going to be essentially an evolution of morphine and kind of where these are headed. Because as you know, one of the things that Shopify has always done a really good job and Harley just touched on it, is anticipating where commerce is going to go, anticipating what merchants are going to need -- and we want to make it sure it's as easy as possible for them to do this. So they don't have to, for example, deal with a bunch of different competing technologies if I'm getting a lead from this channel or lead from that channel. That's not what we want. We want it to be easiest possible for the merchant to get on the Shopify admin, their dashboard run all these transactions, see side by side, all the analytics and make commerce -- Agentic commerce as easy as possible.

These things are going to converge. I mean that's the simple way to put it. There's going to be a bunch of products that come out and they will eventually all converge and eventually, they'll be open source in this way, that will be the new language, for every agentic service talk to every single merchant. And what we're bringing the table is, we think, a much richer version of that. But we like the ACP, UCP and if there's another protocol that's going to eventually be created, these things will converge in very, very soon.

Got it. And are you currently having those conversations with your customers, do they understand the progression of these protocols?

Merchants don't care about that. They don't want to hear about open source protocols. They don't even know what ACP might be or UCP might be. What they want from us -- what they're hiring -- what these millions of stores are hiring Shopify to do is to simplify all these things. If this was 30 years ago, we would -- it would be brick-and-mortar. That's what the business would be. We've helped with that. Now online store, e-commerce, it's still sub-20%, but it's obviously growing very, very fast. It's now a big part of all retail. Then you add social commerce. Then you add the ability to cross-sell and marketplaces. We integrate with Amazon and Target to make it really easy to push products to different service areas. So the idea should be over time, depending on your particular brand and your particular business, you should sell on every single service area where your consumer is spending their time. We have a Roblox channel where -- which most of our merchants are not actually using. But some of our merchants are seeing a ton of growth in the Roblox channel because their particular consumer is spending time there. And how do they access it? In the Shopify Admin, they go to the list of channels and they click Roblox and now they get to push products there. When a product is sold in Roblox, it's fed back to the Admin. So we are -- the business is simplifying all these things. And that's what we're trying to do with the agentic -- excuse me, with the agentic storefronts with UCP. We want to make it so that the merchants don't have to think about that. And I think that puts Shopify in a position. I mentioned in last earnings call that we brought in some really big brands lately, brands like Estée Lauder, for example, or Birkenstock for example, or Starbucks and Burton. When I meet with the CEOs of these companies, when I met with Stéphane, who's the CEO of Estee Lauder, the main question that they ask of us is, will you future-proof our business, meaning I don't want to have to migrate ever again. And the answer to these incredible merchants and retailers for me and from Shopify is unequivocally, we got you. Whatever happens, whatever permutation happens, you are better off with Shopify than anywhere else.

Yes. And the only point I'd add, Keith, too, is just think about from a -- when you think about a macro investment thesis or kind of high or low investment thesis, Shopify has always had its strength in online commerce. And everything has happened with UCP, everything that's happening with Agentic commerce, probably at a minimum, continues to help drive the percentage of commerce, which is online from 20% to some higher number. And so to the extent we can help merchants do this, and there's a bend in the slope of the curve in terms of how quickly we see online commerce happen that plays right to our strengths.

Yes. So we did -- in conjunction with our retail analysts and our Internet analyst, we did a big deep dive on Agentic commerce. And we came to the same conclusion, right? That Agentic commerce is going to shift more commerce into online channels. In addition to that, we did a deep dive on Shopify and Shopify's role in that. And that's what kind of we came away with. Like this is a positive for Shopify in terms of GMV growth, right? This is going to inflect the opportunity, inflect the TAM in terms of online retail, but also Shopify tends to gain more share when we get to these like kinks in the curve. I think investors are still worried about the economics of these transactions when it comes to Shopify. And so 2 questions in there. Like, one, is there more to the economics for Shopify and the upside in terms of stuff like take rate that we're not thinking about in our analysis? And two, are there risks on the other side of the equation? Is there -- and specifically, I think the risk people are most worried about, is there a potential that the transaction comes up into the agent layer, right, that you guys no longer are conducting the payments behind that because that's been a big part of the take rate story for the past decade for Shopify?

Yes. I think it's all take in 2 pieces. One is the attach rate and 2 is just kind of how we fit within the transaction overall. We still are fundamentally the only ones really building the commerce stack for the merchants. And commerce by definition, still need to involve a transaction. We as consumers are going to buy something. We're going to buy it from a merchant. And whatever form we do that with the merchant, we are providing the tech stack for the merchant. So whether you're going to the website, whether you're going through an agent, however, you may do it, we're still the ones that are providing them everything in terms of all their transaction history, their payments, their inventory, their shipping, Shop Pay, all that context is something where we have essentially built through the combination of all these things, Harley alluded to the 3 most earlier, and it kind of comes back to this in terms of how we think about everything that we can do for the merchant and therefore, capture the economics, which we think are deserving of kind of what we're doing.

Part of the reason also that I think a lot of these agentic applications want to work with us is, yes, they want to work with this because of the catalog because we have the products that every consumer wants from every one of your favorite brands, pretty much everyone who's wearing sneakers here, those brands are being sold on Shopify. So that's part of it. The other part of it is the trust factor. Part of why Shop Pay is so popular. It's so well penetrated is not because we force it or defaulted on. It's because it is fundamentally the highest check -- the highest converting checkout, the highest trusted by consumers, if you look at unaided awareness, we have like Shop Pay is successful based on merits. And so we actually think part of why these agent applications want to work with us is for things like our checkout is for things like Shop Pay is for the catalog as well. To answer your question directly, the economics whether something is sold on a marketplace or in the online store or even on something like an agentic application, the exact same thing to Shopify. In fact, we think, as Jeff mentioned, we can see a higher increase in GMV because now this new service area exists. But in terms of checkout on directly in chat, the way that social commerce actually happened was everyone anticipated that the checkout would happen directly in Instagram, which we were powering. In the end, it turned out that social commerce was a great discovery tool, but the checkout still happen in the online store. We were indifferent to where that was going to happen because we powered the check same thing on agentic. We're agnostic to whether they check out directly in the agent tool, we're powering that. It doesn't bypass Shopify check out, it flows through a checkout or if they just use it as a way to find a new brand and then they go to the online store afterwards, and they use Shopify checkout natively in an online store. It's agnostic to us.

Got it. I want to shift gears and talk a little bit about genetic plan. So this is 4 brands that aren't currently on Shopify. You enable them to syndicate their product through these AI surfaces and it feels a little bit like a top of funnel most marketing initiatives for Shopify there, you're getting people into your ecosystem, enabling them to sell -- or potentially you to sell them eventually. What's the go-to-market strategy for agentic plan? How do you get this into the marketplace? And how difficult it is for particularly the larger brands to engage with something like agentic plan?

Part of the reason that we created commerce components, as I think fortunately, you invited us on stage here a couple of years ago to talk about that, Keith. There are some brands in the world who -- very large retailers, traditional retailers, Hunter Douglas, which 1919, Mattel, 1945, Birkenstock 1776, plus or minus a few years, who want to sort of come to Shopify, they want to migrate to us, but maybe they not ready for a full migration yet. Years ago, we did this experiment where we did the thing called Shopify Commerce components, where we took all the different parts of Shopify, checkout, our inventory engine, Shop Pay, all these different sort of tools and modules inside of Shopify, and we said, we actually, you can pick any of them on their own. What we noticed was it was a really wonderful from a go-to-market perspective, it was a great way for us to start conversations with brands who may not -- might not otherwise have been ready to move to Shopify. We're doing the exact same thing with agentic plan. We think that now there are -- almost every retailer are -- I'm on the board of NRF with the largest retailers in the world. Every one of them has told me that their board, their executive team is asking them what is their vision, what is their strategy for agentic when it comes. And so we wanted to basically show up to the market and say, even if you're not ready entirely to move to Shopify, just start with the agentic plan now. We can help you syndicate your products right away, right now with all the merchandising information and all the attributes of the product you require to these agentic tools. Ultimately, what we hope will happen in a similar vein to the commerce components is that starts a commercial relationship. And ultimately, they do a full migration to Shopify. So that's the -- that's where we're going with the agentic plan. It's being rolled out right now. The go-to-market strategy is that we're being very loud about it. But we also have this massive funnel of merchants that we know have wanted to talk to us about agentic, but weren't ready maybe an entire forklift upgrade. And we think the idea is to help them get started early and fast, but eventually, how do we help them fully migrate to Shopify, which has been incredibly effective on the commerce component strategy.

Yes. And the only thing I'd add is it's in their self-interest, and make sure that they use catalog because why would you rather have someone go in search and find something which is maybe a screen scrape versus they actually go, they do a search, it connects back into the Shopify catalog and everything that the merchant spend time on in terms of the product images, the descriptions, the specifications, all that stuff is captured in the way it was intended to by the merchant. It's just a better solution for them.

Got it. Got it. And then can we look to any historical analogs like perhaps Shop Pay, right? Like it's something that you push out to either -- even non Shopify merchants as a potential indicator of conversion rates or what we should expect on conversion rates of taking these agentic plan customers and making them full Shopify customers?

I mean I think ultimately, agentic will operate similar to the Shop Pay commerce component or the checkout component as well, whereby, in some cases, there was an era of retail and commerce and Keith, you said this as much as I think anyone in this room has where they believe their unfair advantage was building their own stack. That part of the reason they were able to substantiate their multiple or their valuation was they were -- we are a technology company and a cosmetics company. That area is over. There is no cosmetic company, most of them are on Shopify, the ones that are not, we're in talks to who believe that their unfair advantage is having their own homegrown commerce stack. So I think we're past that point where people believe that to build it themselves. And to be honest with you, a lot of the enterprise -- the traditional legacy enterprise software companies are not doing a very good job keeping up, that's an opportunity for us. We see this wedge, we're taking it. And part of the way we're able to take it is to say, the migration might take in some cases, 6 to 12 months if they have a lot of ERP integrations and supply chain integrations to do. But what if we can get you up and running within 3 to 4 weeks, simply getting you syndicated right now across every agentic application. That is a very compelling pitch to make. Over time, we over deliver for them. And now we get to go back to them and say, if you like what we did with your agentic plan, we do you see we can do with point-of-sale or helping you build a much faster online store or helping you with the Shop app, for example, you're not in there, you should be. So I think the way that you're talking about it in terms of the metaphor of Shop Pay to agentic plan, that's the way we think about it also.

Let me double that there for a second. In that -- a broader investor concern right now is cogeneration tools, right? And the ability to use these solutions to develop much more quickly, much more customized to your specifications and then exactly what you want. So why isn't this the easiest time for these brands to say, "Hey, listen, we could create exactly the front page that we want. We can create exactly the experience that we want." Because we have these cogeneration tools, and it's now just so easy for us to develop this ourselves rather than having to rely on a partner and pay rents to a part like Shopify.

It's a good question. Look, for years, basically, since the beginning of the company since I've been the company for almost 2 decades, you could replace Shopify with an open source e-commerce widget plus some sort of payment provider plus some sort of point-of-sale instance plus some sort of cross-sell marketplace tool, you were asking merchants to do the thing that is the opposite of what they're all looking for. You are pushing complexity to them. At the enterprise level, no enterprise wants to vibe code to check out. They want to actually have the greatest checkout, highest converting most secure on the planet. They want to be able to do -- we power flash cells for Supreme or for Taylor Swift or for the largest brands in the planet. They don't want to vibe code a checkout that falls apart within the first 5 minutes. So the enterprise, they're just not -- they're not looking to cobble together a bunch of single-point solutions and create their own version of it. And then even at the small business side of it, you could have very easily probably for the last 1.5 decades, use any of the CMS tools, any of the content management tools to create your own version of an e-commerce widget. It is no concept of inventory. It is no concept of loyalty or e-mail marketing or supply chain or taxes or any of the things that you actually need. And so the reason that no one has actually built a large-scale store on any of these CMS because that's not what anyone is looking for. So could you replace Shopify with a bunch of vibe coded tools? You could. But that -- then you're pushing exactly the opposite of whatever wants, you're asking them to effectively become human synchronization tools. That means they don't get into the right agentic tools. It means they can't fully integrate into Spotify or Roblox or Pinterest or Snap or any of the channels that we power. It also means you don't get things like Sidekick, you don't get things like SimGym So for $39 a month, you get everything you need and you don't have to vibe code it yourself. I think from a value proposition that is very, very compelling. And that's the reason why I think even well into this kind of vibe coding era, you're not going to see people cobble together 5 different systems and effectively become that human synchronization tool.

Got it. I opened up the conversation talking about the majority of the investor conversation or the investor concern coming out of the most recent quarter was on Agentic commerce. There was a little bit of conversation on margins. So bring that into the conversation as well.

Finally, something for Jeff.

Finally, something for Jeff. So you guys delivered $2 billion in free cash flow in 2025, 17% free cash flow margins. How should investors be thinking about free cash flow margins into 2026? And how are you guys thinking about that trade-off between the potential of driving more margin expansion versus the big nascent opportunities that you're seeing with Agentic commerce and the expansion of the business?

Well, I guess in reverse order, I mean, our financials, and this has been for 3 years now where we basically have been delivering this top line growth and these free cash flow margins in an even quarter have kind of been in the mid- to high teens to low 20s, which is obviously what we've seen in Q4 and a lot of quarters. And so what you'll see from us is that continued as we think about the balancing of the 2, right? We've not suffered in any way, shape or form in terms of our ability to invest in the future. You see this in additions, which we do twice a year and all the great products we've continued. And even as head count has come -- stayed flat to come down, we've still been able to crank out this many products, deliver this top line. There's not -- and I talked about this a little bit on the call, there's nothing in terms of how we think about it this year, that's any different from last year or the year before. We fundamentally think we've got a very good balance between delivering these margins and continuing to deliver this GMV growth rate, this revenue growth rate, this gross profit growth rate that we've talked about. In Q1, I mentioned a little bit that there were some tax dynamics in Q1. But for the year, and I was very explicit about saying that's an intra-quarter dynamic. There's nothing for the year that we look at and say this is any different than the excellence we've been able to put forth the last.

I'd also say just on the head count thing, I know that that's sort of the -- I don't know if that's the topic of the Morgan Stanley conference or not the head count in. I know there's talk happening around the corner from here on a company that's making some big changes. Shopify has not been growing headcount. In fact, just to Jeff's point, for the last 8 or 10 quarters or so, headcounts actually come down. I will also say, at the same time, the head count is coming down. This is -- and I say this with -- like with reference for the fact that I've been around for so long at Shopify. This is the highest quality talent team we've had the talent density of this company is as high as it's ever been. And we think we can continue to do this really incredible top line growth and provide $2 billion of free cash flow and invest in all the right places. I mean, for anyone that follows our addition releases, which is our product releases, which we do twice a year, these big events, if you look at the pace and just the velocity of product releases on a biannual basis, it's unbelievable. And we're doing that on a team that's gone -- we're below 8,000 now. So we've taken our medicine over the years. And I think the iteration that we're in right now is very strong, super high talent team, grow top line, 30% or so, $2 billion of free cash flow. That allows us to invest for the next 10 or 20 years and really not just take part of this agentic new era, but actually lead this agentic era.

Got it. So we focused a lot of the conversation on the agentic commerce part, but there's a lot of other innovations. What are the other 2 or 3 kind of product innovations, Harley, that we should be focusing on in 2026 that you think it start to really move the needle?

I mean I think we talked about agentic, but we really talked about it from the consumer side, generally, at least in the last 30 minutes or so, which is really the consumer side of how they purchase. I mean, anyone who's not paying attention, what Sidekick has built is truly unbelievable. From an entrepreneur perspective, from an enterprise perspective, I was in L.A. this past week meeting our largest merchants and just showing them like on their own, tell me -- ask a question, like as your largest customer, your individual customer and now ask Sidekick what you should sell the next. And the ability to get this data at this incredible pace, not just that it makes so that when we have a new feature that we think they should adopt. Sidekick Gives us an amazing tool to push product to them. So I think Sidekick is one of those things that doesn't get a lot of attention, but is incredibly important. I think what we're doing now around point of sale. We still have a lot of merchants on Shop app who don't use point-of-sale who should. Most stores and all of you walk into physical stores. When you look at their cash register and you look at their point-of-sale system, it is like -- it's terrible. We think it's a huge opportunity. The stuff we're building internationally, this idea that when you sign up for Shopify, you were default global, you should sell across every country. We think it's a really, really big, big opportunity. And maybe the last one, some of the stuff that we're doing around Shop, the Shop app itself, if you -- if any of you that have been tracking it, for the Black Friday, Cyber Monday weekend of the 4-day weekend of 2025, not just in the shopping category across every category, the #2 app was a Shop app. So the Shop app is not getting a little bit better. It's getting way better. It's becoming a lot more consumers' favorite place to purchase and it's an owned property of ours. And so those are some of the things, I think, that are -- we still have not fully recognized the opportunity there, and that's what the teams are focused on.

Outstanding. Amazing piece of innovation coming from Shopify, amazing execution was make Shopify best athlete for our Morgan Stanley research. So I appreciate you guys coming in sharing the story.

Appreciate it.

Thank you to you both.