Impinj, Inc. (PI) Earnings Call Transcript & Summary

All right. Thank you, everyone, for joining. My name is Dmitri Anissimov and I'm a Managing Director in the Semiconductor Electronics side at UBS. And today, we have Impinj. We have Chris Diorio, CEO, Co-Founder and Vice Chairman of the company; and we have Cary Baker, the CFO. And I think we can just go straight into the presentation.

Chris, Cary, could you start with an overview and the history of the company and kind of how the story and the vision evolved over the years? And where you see the company going in the next couple of years?

Sure. Okay. Thanks, Dmitri, and welcome, everybody. For those who are new to the story, I'm just going to start at kind of the base level, but we'll move quickly through the story. So think of what we do as providing a wireless 2D barcode for items, but we use a radio chip, a miniature radio chip smaller than a grain of sand that goes into or onto the item. And that miniature radio chip can be [ ratted ] up to 1,000 items per second at up to a 30-foot range without line of sight and no batteries. We have delivered more than 100 billion of those radio chips to date. And last year alone, our industry delivered 44.8 billion of them. We're connecting items that range from retail apparel and footwear to packages shipped in supply chain and logistics to airline bag tags to marathon bibs to global entry cards and an incredibly broad range of opportunities. Impinj, as a company, makes those radio chips that goes on to the items. We don't make them ourselves. We outsource the manufacturing, but we design it and sell them. We make reader chips that go into partner readers to read those items in enterprises, in stores. We make finished readers and gateways for high-performance applications. And we do some cloud services, work for authentication, for authenticating items as being genuine. We're the only ones in our space who have an end-to-end platform spanning those elements, the chips that goes on the items, the reader chips, the readers and gateways and the cloud services, really the only ones who have a significant presence at both ends of the radio link, the chips that go on the items and the readers. We pioneered this space dating back years and years ago, and we've been in this industry ever since. It's a form of radio frequency identification that we helped co-found the industry alliance in 2014 and branded itself as the name RAIN for radio identification, a bit more friendly name than RFID. And we're promoting, and the industry is promoting and advancing that RAIN name going forward with a view to the future where consumers, through their mobile devices, will be able to engage the tagged items, essentially the connected items and get value from that engagement with items they have on them or own or in their homes. Today, enterprises are using our products to track items they manufacture, transport and sell. But the future, the far out future measured in a good number of years, but still we see it coming, is where consumers can engage directly with their connected items and enterprises can engage people through those connected items. Anyway, that's who we are. That's our story.

And so you discussed your full solution with key building blocks. What's your typical entry point into a customer? Is it the tag? Is it the systems? Is it the software? Or is it kind of everything together?

You asked previously what has changed over the years. Historically, our entry point back years ago was just convincing enterprises that going forward and putting chips on their items made sense. Now with the scale of the industry, we're not making that pitch effectively anymore. The use case is proven, the benefits for inventory visibility, supply chain traceability, item freshness. The use case opportunity is known. So today, our primary entry point with our end customers is from direct referrals from other end customers where an enterprise comes to us and says, "We've got a particular problem. We haven't been able to solve it. Can you help us solve it?" I'll give a very specific example there. One was an enterprise that wanted to do -- our visionary European retailer who wanted to do self-checkout. They wanted their customers to be able to do self-checkout in the stores. But in order to do self-checkout, you need to get rid of the hard tags that are on apparel items, and you need to replace them with something else. So they wanted to use the same RAIN tag that they were using for inventory visibility for self-checkout. But in order to do that, you need to enable loss prevention using that same tag. They have been unable to implement loss prevention. They came to us and said, "Can you help?" We solved that problem, which enabled them to go forward with self-checkout in their stores.

Got it. And are there examples of where you sort of come in and land and expand the customer into other solutions, maybe into a recurring business model, if you will?

Yes. I'll use that same visionary European retailer as an example. They want full supply chain traceability from item manufacturing, where advanced shipping notice and ASN gets generated to receiving, to verifying that the items received in a box, match that ASN. And we can read right through the box, so you can read 100 items in a box and verify the match with the ASN through sorting in the back store; through front store to back store to know it's on the store floor; to store floor inventory to do automatic replenishment; to changing rooms to streamline the changing room experience; to the self-checkout, loss prevention; and then in the future, having an app that the customer can engage the enterprise and vice versa through the items that the customer bought; all the way through to end-of-life recycling, using the same chip to provide recycling instructions to the recycler, how to recycle that item. So that's kind of the ultimate land and expand. We've only built small portions of that vision today. With that vision, that end-to-end item life cycle traceability, including the consumer use is our long-term vision. And when I'm saying it's a long-term vision, this is a multi multiyear vision. It doesn't happen overnight. We've been at this thing for about 20 years. We're going to keep getting -- pushing at it, but I don't see anything in the way of reaching that vision.

Got it. And is there a recurring aspect to your business model going forward? I don't know if it's the software aspect of it or maybe you view your tags as kind of like a reusable recurring element?

You should think of the -- the endpoint I see is the little radio chips that we provide to our inlay partners who attach them to an antenna. That's a recurring revenue stream for us. When a customer buys an apparel item, that item they may take home. And the chip is either part of the item or it's part of the price label when they cut it off and their retailer orders another item with another chip in it. So we believe we're the first market for recurring silicon. . And we also see an opportunity in the consumer side actually before we get to the consumer with some of the enterprise things, but then heading into the consumer side with cloud services around item authenticity, consumer privacy, security and item resolution. There's a very big data play associated with the consumer engagement model that -- but it begins today with the enterprise engagement model. We are not a big giant data player. We're not going to be competing with the giants in the data space, but there are entry points for us in cloud services that we believe makes sense for us and will be a long-term recurring revenue stream for us as well.

Got it. Now on the market side, just like shifting gears a little bit. I think we can start with your TAM opportunity. I think you guys look at it from a number of tags or dollars, whichever is easier. But as we go from retail apparel to retail general merchandise to logistics and supply chain, how big is your TAM? And sort of where are we in terms of penetration in some of these markets?

The industry did $44.8 billion last year. I take that as less than 1% of the total opportunity. I kind of wag it is maybe 0.5%, if you think of all items that can be connected. Sure, all retail apparel and footwear, but that's small. What about all pharmaceuticals? That's a little bit bigger. What about all bio specimens, blood draws and everything in the health care space? All automotive parts? Everything that's shipping and logistics companies transport. All food items and everything in between. We don't have an exact number there. But I'm like, yes, this is on the order of 10 trillion items a year, and I believe every one of them can be connected. So we are still in the early days of adoption. And I'll even say that the only vertical that we have that entered mainstream adoption, really truly entered mainstream adoption is apparel and footwear. And all the other categories, supply chain and logistics and all the others we talked about, we're really in the early adopter phase. Fortunately, our early adopters are Fortune 100 or Fortune 500 companies who are proving the use case for -- another example being food, who are proving the use case. So when you get a Fortune 100 company driving forward, they pull an industry with them. So I have little doubt about the adoption potential. I'm excited about the enterprises that are adopting as the early adopters, these gigantic companies. But in terms of the answer to your question, we're still in the early days. 20 years in but we're still in early days. We've got a gigantic opportunity ahead of us.

Yes. We size apparel opportunity as 80 billion units per year. We think we're roughly 30% to 35% penetrated to that opportunity. The logo penetration is much higher. Most retailers have made the decision to move forward with RAIN, but most are not 100% deployed. They're just somewhere on their journey to 100% deployed. We think of the general merchandise opportunity at 325 billion units per year, low single digits penetrated right now. We think that logistics at 400 billion unit opportunity per year, low single-digit penetration right now, very low single-digit penetration. So we've got the big players in each of those verticals moving forward. But to Chris' point, we're just in the very early days.

So on the retail apparel side, can you talk about what's been changing in that space in the last few years? I guess as we were moving from like manual checkouts to autonomous reading, things of that nature and how you're addressing some of these technological challenges?

Yes, two fundamental changes. One is the current use case for the chips as they go in the price label that a person cuts off after they buy the item. Going forward, we're going to see more and more of the ICs being embedded as an integral part of the item. And then after that, being able to survive multiple washing cycles and being used for recycling at end of life. The second one is the initial use case was inventory visibility in stores, and that inventory visibility in stores is expanding in both directions of the self-checkout on loss prevention at one end, and then supply chain visibility at the other end, including visibility as the item moves within a store and an automated replenishment.

And so technologically, what do you have to sort of do, change or address with the radio technology, I guess, or the software around it, doing make work, right?

Yes. So that's a little bit of a hard question to answer because every particular retailer is different and every use case is different. The challenges involve everything from dealing with consumer privacy issues. For example, if you're going to have sewn-in tags, you've got to protect consumer privacy to environments where there's a lot of readers. Think about supply chain and logistics, where there's multiple trucks backing up the dock doors, right adjacent to each other and identifying the package and which truck it goes into. And that it's the right truck, not the adjacent truck, which is just a couple of feet away. To everything in between. So we've been building a toolbox that we use to solve those problems. And that toolbox includes not just software. It includes changes to the endpoint I see to enable us to sell these use cases. And we just released that toolbox broadly yesterday in the form of something we call Gen2X. The industry radio protocol is colloquially known as Gen 2. We have enhancements to the protocol, which we call Gen2X, which we just announced and released broadly to the industry with support from all the leading reader manufacturers. So yes, we will continue innovating in our space every way we can to address these enterprise customer problems.

And as we go from apparel to general merchandise, what is different? Obviously, as you mentioned, it's a bigger TAM potentially. But what is different there in terms of how you have to reengineer the tag itself and the technological aspect? And sort of what has been taking, I guess, a bit of time for that segment to start shipping?

Well, if you look back at apparel and footwear, it took a bit of time, too, because Macy's and Decathlon and Marks & Spencer got going in the year 2010. So it seems like it was just yesterday, but it's actually been 14 years, and a lot of the [ kinks ] have been worked out of the system. So as we move to general merchandise, the form -- the number of form factors that need to have an IC on them is immense. And some of the -- I'll just give a specific example. Imagine puzzles. Okay, it's in a box, it's all cardboard. It should be easy to tag, right? RF wave is very easily readable. But there's no tag on the box. So where do you put this adhesive sticker? And how do you avoid covering the part of the box, which is the puzzle? And what -- how do you modify a running assembly line to add the thing? Or maybe you don't add the tag at all. Maybe you stick it inside with the puzzle, do you do that? Or do you make it part of the corrugate? So that's a simple example of the questions that need to get answered for essentially every item of general merchandise. Some have metal in them. Some are consumer electronics, and you need to be separate from the consumer electronics part and not be shielded by the consumer electronics. So for every category of general merchandise item, the retailer and their supplier needs to basically work through and say, "Okay, how are we going to do this one? How do you tag a fishing lure?" By the way, fishing lures are being tagged today. So they work through that one. But if you think of it, how do I tag a fishing lure? And how does it go on the shelf? So there's nothing fundamental here about the pace. Nothing fundamentally negative about the pace of general merchandise rollout. It's just there's a lot of details that have to get worked through. And once one category gets solved, the industry moves to the next one.

And would food and perishables be part of this general -- or is it a little different with the supply chain angle as well? But in general, can you talk about your food and perishable effort? I think it's pretty interesting to [indiscernible].

Yes. So I had originally believed that food tagging would go through the same difficulties. And in order to get food items, we're going to have to figure out how to put tags on cans and on bottles and all these things that are really hard because cans have a lot of metal. The radio waves don't go through them. And so I have envisioned food as being far out in time, but I was wrong. Where the food opportunity is materializing is in items where the price label is created in store so you don't have to tag through the supply chain. Think of all the fast replenishment perishable items: bakery, meat, fish. You can even think florist. They're perishable items, a store needs to either sell them, mark them down or donate them before they become food waste. Food waste is a huge problem today. The retailers used to lose -- and the grocers lose a lot of money on that food waste. If they could just donate this stuff, they'd be better off. And on top of that, they have the infrastructure in the store that already prints the label. So you just take that label, you put our little chip with the antenna inside the label, you change your printer so it prints the label and encodes the chip at the same time. It gives the employees a handheld reader to inventory the store. And all of a sudden, you can make an efficient inventory visibility for perishable items in store. So I had not anticipated that use case. The grocers came forward with it themselves, but it has dramatically accelerated the pace at which I thought food adoption would happen.

And then going back to the general merchandise, you're working with one large North American retailer there. Can you describe how you sort of, strategically perhaps, have gone after one big player in the space and how maybe others are starting to piggyback off of that and you're getting some of the smaller customers as well?

Yes. Well, it always works that way. Our industry is just a little bit strange in that we tend to have an industry giant or 2 or 3 lead a new market opportunity. We engage with them. We develop the technology enough so that they can move forward. And then others follow in their footsteps often without fanfare. And so it happened in the apparel space. So like I said, it was Macy's, Marks & Spencer and Decathlon. They're a sporting goods retailer but they also sell a lot of clothing. They started on the clothing side. Supply chain and logistics, one lead customer. And retail general merchandise, one significant lead customer who we partner with closely. But there are others already issuing [ RFUs ] for items where the suppliers already know how to tag them. So as I said on our last earnings call, we have other general merchandise retailers piggybacking on those efforts.

Got it. And then on the supply chain and logistics, you just mentioned there was also 1 or 2 key focused customers for you. Can you describe the challenges in that space? Is it somewhat related to sort of tagging on the retail side? Can you combine the 2, one day where you can do work on the logistics side of it and then to work on the store side of it? But just described the challenges there because it sounds like the TAM there is also fairly large.

The TAM there is very large. The challenge is there are item density. You've got conveyor belts. Some with just items that are spaced one at a time, but they're really close together. They've got to get routed the proper way. Some conveyor belts for the truckloading that are just densely packed with items. You've got trucks that are right next to each other that need to get loaded properly. . You've got metal trucks with boxes all in them, and you've actually got to get good readability where the RF waves are bouncing all around. And you got to make sure the right item is being shipped to the right location through the conveyor system onto the right truck. So it's very different from the retail use case. It's very much a supply chain and logistics doubling. It's actually the supply chain, for the supply chain and logistics company, the routing of those items and dealing with clutter, significant amount of interference. If you think you've got 100 trucks, you've got 100 readers operating simultaneously. You got an RF quota problem. You've got an item density problem. You've got a readability problem. And you want to get a very high degree of accuracy. So for retail apparel, when a person goes out and does a handheld inventory, they never get to 100%. It's not worth the labor time to do that. So typically, retailers specify an amount of time to spend or they actually have some way of counting. And when the person doing the inventory gets to about 98%, they're done for that day. They get the next day missed ones next time. 98% is good enough for each inventory cycle, especially if you do it twice a week. For supply chain and logistics, 98% anywhere near close to being good enough because the truck's got to be loaded with the right items with better than one in 1,000 accuracy. So there's an accuracy problem as well, and that makes it hard and that makes it fun.

Speaking of fun. Here's a fun question. Sort of looking at these 3 segments, where do you see them in the transformation in the next 5 to 10 years? And I think the fun part that you mentioned in the past would be this technology being embedded into, say, smartphones, right, or into things of that nature where it's in your phone, you're walking into a store and you can sort of automatically go through the whole retail process. Can you give me the vision in the next 5, 10 years, maybe 15 years where this could take, where this could be?

I can. I'll probably get it wrong. But the way I like to think about it and the way I like to envision what we have to accomplish is in tracker tags, like you could buy it to put on luggage or things, that come for free with the item and with 10,000x the footprint. Every item in the world will have a tag on it. Every phone will crowd source reading as you're walking around, and you'll be able to have a Find My app, enterprises will be able to engage through you, through your connected items, but not you through somebody else's connected items. The data back end for that connectivity associated with everybody's phone reading items and crowd sourcing as well as using your phone for buying, as well as using it in your house, as well as using it for your own inventory is gigantic. It's just a gigantic opportunity. And the shortcoming I have is, if you go back in time, I think when mobile phones, when we first got touchscreens, who thought of the apps that we would have? Who got it right? I mean I certainly can't raise my hand. So I can't even think of all the apps that we're going to get based on item connectivity. All I can say is, with a fairly high degree of confidence is, I know it's going to be big.

If there's any questions from the audience, feel free to send them through the app. On the product side, just more mundanely I guess. You're moving from one generation to another, from M700 to M800 in the next few years. Can you just describe that sort of technological advancement that you have to do on a continuous basis? And what that means in terms of pricing, performance, footprint, et cetera?

So I'm going to say, I'm going to give a little of the technological side. I'm going to let Cary take some of the rest. But before you say rather than have to do, I'm going to say that we get to do. I have this fun curve that I used at [indiscernible] conference maybe 2 weeks ago. And it plotted the sensitivity of the IC as a function of time going back to 2006. And we've been able to improve the sensitivity of the IC by about 25% per year, which is an exponential. . It's actually kind of fun because we've got multiple exponentials behind our industry. That being one of them. Another one being the industry unit volumes have been increasing, growing by about 25% to 30% per year for the past 15 years as well. So it's very rare for a company to have tailwinds that include multiple exponentials, but that's a separate story, yes? Okay. So what does that really mean? Well, it means we've been able to shrink the size of the antenna, the area of the antenna by 25% every 18 months. Yes, it's a 25% improvement in sensitivity every 18 months or alternately, we shrink the area of the antenna by 25% in 18 months. So now we're at the point -- back in the old days, we had antennas that were this big. They were 4 inches by 4 inches. And now something about the size of the end of my pinky, which means they can go on to anything, and we're still on that curve. So we get to continue innovating on the sensitivity and performance side on the size of the IC and the features we put in the IC, the smarts we get to put into it, and we're just scratching the surface in its wireless technology can do.

Yes. And the other benefit is by shrinking the size of the die, we're essentially spreading the largest cost in our bond, the wafer cost amongst more die. So with the M800, we get 25% more die per wafer, which translates to a significant cost savings. So we're able to deliver the more sensitive IC, which allows our partners to take cost out of their bond by using a smaller antenna to get the same read range and we're able to do so at a lower cost than the prior generation M700 chip. And when the M800 ramps to our volume runner, we still have cost advantage left over for Impinj that should drive 300 basis points of gross margin accretion. So that's down the road when the product is fully ramped, but this is why we invest in multiple avenues.

And what does it mean for ASB pricing? Has it moved from one generation to another? I mean it sounds like the gross margin uplift is there through the cost side.

Yes. The gross margin is there. We would like to accelerate the adoption of the M800, Historically, when we launch a new IC, it takes time to ramp because once the socket is working with an IC, there's not a lot of incentive for our customers to change. And what we see is we first go into new sockets with a new endpoint I see and then gradually over time, replace existing sockets. We're trying to accelerate that curve with the M800 because we want to get to the gross margin goodness. So we purposely priced this higher-performing IC at a lower cost than the M700 to accelerate or in hopes of accelerating the adoption.

I'm going to step in with one fun factoid. So just think about this. We make a radio on a chip that's roughly 250 by 350 microns, roughly the length of 50 bacteria end to end, and it is a full radio on a chip. That's the fun.

I agree. I agree. On the financial side, you had a couple of [indiscernible] in the last couple of quarters. Things are looking really, really good. Broadly speaking, how do you look at into '25 as we're seeing, I guess, normalization of inventory levels? Have all of your segments sort of normalized over the last quarters?

We think -- so retail apparel is the largest impact on that. And I would say in terms of inventory levels normalizing, it's mixed. We track a subset of retailers that we think of as a -- public retailers that we think of as a RAIN indicator set. A quarter ago, most were growing sales faster than they were growing inventory. The most recent quarter, it's more mixed. So some are growing faster, some are growing at the same rate. When I look into 2025, I see the same factors at play for '25 that we saw in '24. We see continued expansion in retail apparel. There are some net new opportunities, but most of the apparel is going to be expansion related. We see in general merchandise, the large North American retailer continuing to expand on their Phase 1 and Phase 2 deployment. So that will continue into 2025. And we also see continued growth with the second large North American logistics provider. They're not 100% deployed yet. They're continuing on that journey. I think they get closer to it in 2025. There will be net new deployments in 2025. But what we don't see is the large Fortune 100 opportunity ready to go in 2025. We don't see that at this point. That could change, but there's not the big player. And then I would say, the final piece that we're all trying to figure out is what are the impact that tariffs have on our business, on everybody's business on the economy? It's just -- it's too opaque for us right now to fully understand that.

So to close it out, a super quick question from the audience, which I think would be interesting is, what are you doing that your competitors aren't yet doing? And how do you see that in the next couple of years?

Enterprise Solutions. We're engaging directly with the enterprise, our primary competitor. Really the only entity in the space that I think of as a competitor is NXP. They sell endpoint ICs, we sell solutions. So we're engaging directly with the enterprise. We're innovating with the enterprise. We're solving the enterprise problems. We're driving those innovations into our platform, which includes into our endpoint IC. We're innovating features in the endpoint IC that give us a differentiated opportunity in the market. Kind of the goal is to create those solutions and have partners step in and repeat them using our endpoint ICs.

Fantastic. Well, thank you, Chris. Thank you, Cary.

Thank you.

Thank you.

My pleasure.