I think we're live here. We'll go ahead and get started. So, good afternoon, everyone. Welcome to the Impinj presentation. I'm Rob Mason, Senior Analyst at Baird that covers advanced industrial equipment. Very pleased to have Impinj Technologies at the Baird Industrial Conference this year. And just a quick background for those of you that may be newer to Impinj. Impinj is a leading developer and enabler of RAIN RFID, which is connecting billions of items, providing item-level visibility, and delivering real-time information to businesses about the items they create, manage, transport, and sell. So, I want to welcome Cary Baker, Chief Financial Officer from Impinj. We also have Hussein Mecklai, who's the COO of Impinj, with us. So, we're going to just kind of go right into Q&A up here.
I did want to mention, if you have any questions, feel free to send those up to the iPad here, and we'll work those into the conversation as well. Maybe I'll just start off, though, Cary, just thinking about coming out of the third quarter, kind of the state of the world. It seems, one, I heard normal seasonality more as you're describing your business, and we're getting past some of the, I guess, abnormalities following supply chain crunches and the like. Just kind of how are you thinking about the fourth quarter and as you go into the first quarter of 2025 under the guise of kind of normal seasonality as well, how that paints a picture for Impinj?
Yeah. So, I do think we're back in a phase of normal seasonality right now. Normal seasonality for us is down 5%-10% typically in the fourth quarter. That's because we ship in front of the holiday season to both retailers and logistics providers as they prepare for the holiday. This season, endpoint IC, I think, will be on the more favorable side of that down 5%-10%, so a little bit better than what historically would say. On the systems business, we have kind of the opposite factor. Systems typically increases in the fourth quarter. This is due to a couple of reasons.
First, just as a baseline, as systems providers are coming up against the end of their fiscal year and customers are coming up against the end of their fiscal year, we typically see a budget flush, and that results in higher system sales for us in the fourth quarter. In this Q4, we're also benefiting from an outsized delivery to the loss prevention deployment. That deployment will spike in Q4 and then go back to normal levels in 2025. And then the third benefit the systems has this year is we are still shipping a few last-time Indy reader IC shipments. This was our prior generation Indy reader IC. The Indy product was end of life was announced last year. Last-time shipments were originally scheduled in the June quarter of this year, but those have extended a little bit.
As you think about just kind of the end markets, you did have some upside in the third quarter as well. Just kind of relate to us what drove that and how that may be influencing the outlook as well.
Yeah. So, we saw broad strength across all of our key verticals in the third quarter, and this is what drove the strong performance and the overperformance versus our guidance. First, in retail apparel, we still see recovery happening in retail apparel, and we don't believe it's fully recovered yet. Now, this is our interpretation of the macro data. We track retail apparel imports versus retail apparel sales. And imports, we saw a nice uptick in the July data that held through August and September, but that still lags retail apparel sales. We also track a subset of public retailers that we think are indicators of RAIN adoption and RAIN volumes. And almost all of those retailers in their most recent print showed sales growing faster than inventory. So, our interpretation of this is there's still recovery left to happen in the retail space.
Beyond that, within retail apparel, it's a category that is approximately 30% penetrated, yet most retailers are moving forward, yet most retailers are not 100% deployed. In fact, you can count on one to one and a half hands the number of retailers that are 100% deployed. So, there's still plenty of expansion opportunity left within the retail apparel space, and there are net new opportunities within retail apparel as well, despite the high logo penetration that we see today. On the general merchandise, we saw continued expansion with the large North American retailer on their two phases of general merchandise. This was a program that, after a series of delays, really kicked off the tagging portion of this deployment in earnest this year, and it's been growing sequentially each quarter, and we expect that to continue into next year.
Then, finally, supply chain and logistics with our large second North American logistics provider. That program's been going like clockwork. They set up a target, they hit the target every quarter, and the predictability of that program's been in place for more than a year right now. It's kind of a combination of all of these three, plus just continued recovery and growth in the long tail that drove the strong performance in Q3. All three of those factors are prevalent in Q4. It's just Q4 we deal with the softer seasonal aspect.
Yeah, and then just getting some feedback here. I'm not asking you to predict when retailers would restock, but what would be the leading indicators that you would, I guess, how much advance notice and visibility would you have on that effort?
Yeah. We'd see it in our orders first. We'd get the confirmation and the lagging indicators of apparel imports, I think, is where we'd see it the most, and then in that subset of public retailer prints.
Any indications as to when they may take inventory levels to a higher level?
I wish I knew that one.
Go ahead.
We would have thought it would have happened a while ago, but it just hasn't happened yet.
Yeah, and then you talked about retail apparel being only about 30% penetrated. Can you put that into context just where we are in terms of overall industry composition for that? I mean, you had roughly 45 billion tags shipped last year, 2023. Retail apparel, probably the largest percentage of that, but where does that sit as a % of the total right now?
A percent of the total we think of as the opportunity. We think of retail apparel as 80 billion units per year. General merchandise and logistics are each four to five times larger than that. In terms of what we shipped last year, we think approximately half of the tags went into retail apparel. It was a little bit of a down year for apparel, so that same 50% may not hold in 2024.
Okay. And then just in terms of sticking on retail apparel, because I know we talk about a lot of other use cases, but that's where you still have probably this more significant part of the business, and there's still a lot of runway there. So, what's kind of the gating factors to seeing more penetration just within retail apparel?
Yeah. So, I would say within retail apparel, the adoption is going pretty well. A lot of the brands have adopted, and it's a matter of just the penetration, and they're progressing well with that. What we do see is two dynamics. One is that there is a transition to going from tagging as a hang tag or as an attachment to tagging as embedded tagging. So, that we see starting to gain some momentum. And the second thing we see is that a lot of the enterprises that want to go past just handheld-driven inventory counting to be able to do loss prevention and also self-checkout, they're pushing to get to 100% tagging, because only when you get to 100% tagging are those things possible. So, those are two catalysts that we see, and we see a push, we see the pull from the ecosystem for both of those.
We see the good momentum in terms of continuing deployment there.
Yeah. You mentioned that, because going to 100% tagging, source tagging perhaps, but just 100% tagging to enable those other use cases, the European retailer that you're associated with has been moving down that path in successive phases. You talked about higher systems activity in the fourth quarter, I think related to that. Where is that particular customer on their journey? May have lost count which phase we're on, three or four, but just where does that stand, and does that project carry into 2025?
That particular parent company has gone through most of their brands, so they're just picking up the tail end of the brands, and then most of the geos have already been deployed, and now they're going out to the less densely populated geographies. What we see in 2025 is them finishing out that deployment in terms of loss prevention, and now their focus has turned more towards enabling self-checkout, and then there's a new set of use cases that they're also exploring together with us.
I see. The general merchandise category that you spoke about, Cary, that continues to progress. We talked about having a lead, how important it is to have a lighthouse customer in a new vertical, and others may fall in behind that. Where are we in that progress, I guess, to start to think about others falling in behind in general merchandise? Any sense of that, how close we may be to seeing some of those follow-on opportunities or customers enter?
Yeah. Well, I think we have the best lighthouse customer for that right now, and as I mentioned earlier, they're midway through their first two phases of deployment. Conversations are already happening about what other categories can go into future phases. Now, the interesting thing in general merchandise is there was a lot of heavy lifting that had to happen first before general merchandise rolled out. How to tag a board game where the barcode is part of the packaging? How to tag a package of sticky notes that are wrapped in cellophane? All of those things had to be figured out by the ecosystem before this program was allowed to move forward. Now, within general merchandise, there's a lot of sharing of suppliers.
So, a second general merchandise player that uses the same provider of the board game or the same provider of the stationery can now go leverage the first heavy lifting and start running those tags through their store on a much faster basis. So, we think general merchandise is the unique opportunity where piggyback deployments can go very fast. It also helps if we're talking in the big box space where a piggyback deployment also leverages RAIN for their apparel program. That way, the store employees already know the motion to read tags, and we think, again, that just helps provide an opportunity for an accelerated follow-on deployment.
In the logistics arena, you talked about that being in a pretty steady cadence with that second customer. Any sense to how that's influencing endpoint IC, I'll say volumes or maybe revenues this year or in the next year?
Yeah. That customer increased label consumption in 2024 versus 2023. They will increase label consumption in 2025 versus 2024.
Yeah. Have you given any kind of order of magnitude or?
No. The biggest increase probably occurred this year, which is what we typically see. The first phases of a deployment typically take some time. The second, call it going from 20% to 80% can typically go very fast, and then the final 20% takes a little bit longer. So, while I expect growth in 2025, it won't be at the same rate as the growth rate we saw in 2024.
Okay. Starting to get some inbound questions just around the food space, which is elevated as an opportunity set, perhaps maybe even faster than we would have thought. Just kind of give us your perspective, Hussein or Cary, just on what's driving the acceleration in that. Are there technical challenges also that have been overcome to enable that recently?
Sure. Yeah. So, actually, you articulated well. It was really a pull from the enterprises and the grocers, and primarily it was driven by their desire to address the perishable goods. So, one of the large grocers in North America had come and told us and shared with us that they were throwing away $1 billion worth of food annually. So, there's clearly a huge incentive for them to try and tag to address waste and identify and address waste. And what we've seen is that they start with the highest turnover perishable items that have the least amount of people touching it from an ecosystem perspective, and that led to bakery as an example.
You get a frozen pie into the store, you heat it in the store, package it in the store, tag it in the store, and then you deal with the sale or the disposal of it if it expires, so that means that most of the tagging ecosystem is contained within the store itself, so it makes it a lot easier to go because you're only touching one set of activities. You're not having to push further into the ecosystem. Plus, with bakery, you have daily turnover, so it's a high turnover area, so they started there, and to your point, we had expected food to be much further away because there are technology challenges that have to be overcome when you're trying to tag food items.
And in particular, so if you started with the bakery, if you look at the pies, the pies are in a tin pan, which makes it pretty hard to read the tags if they're attached to the tin. Then the pies have a plastic case, so you could put it on the plastic case, but then they're stacked one on top of another, which means now it's touching another metal pan. So, as a consequence, you stare at that and go, like, you have to rethink how you do tagging. So, the pull has basically caused the ecosystem to be much more innovative.
So, the solution that they came up with was pretty clever, which is you attach the adhesive portion of the sticker on top of the casing, but then it hangs over the side of the pan, but then there's no adhesive on the side, so it doesn't stick to the side. So, then now you create an air gap between the conical shape of the pan and then the tag itself. So, now it makes it readable. It's no longer touching the pie above it, and now you can stack them, and now you can have the same store operation, but now you get the benefit of the technology. So, what we see is the migration starting there and then moving through the perishables.
So, think milk was another good example of where we had a leading grocer come to our lab and say, "Hey, we want to tag milk," and we all stared at him and go, like, "Great, liquids, that's going to be a really hard thing." And we stared at him for a while, and then we determined, like, you know, you could actually put the tag in the cap of the milk carton, and the cap doesn't touch the milk, so now you've got an air gap and you can make the technology work. We did that, and then we brought them in and showed them the demonstration. Now they're all kinds of excited. Now they want to do juices and all other liquids that expire.
So, you start to see this flow through, and it makes sense the way they're approaching it because they're taking where they have the most waste, and they're pulling us in that direction and going, like, "Help me eliminate the waste," and then working backwards through it to items that have a longer and longer shelf life. So, I would expect that proteins would be next, and that's going to be challenging too, and then eventually getting to all the items in the grocery store because long term, where they want to get to is they want to enable self-checkout and loss prevention in an easy-to-do way. So, fundamentally, labor costs are going to be the last consideration, but the first consideration starts with eliminating waste and then working backwards through the efficiencies.
That's pretty insightful. The ability for Impinj to participate in that opportunity, obviously on the endpoint IC side, what efforts are you going to make on the system side or any software around that as well?
So, what we like to position ourselves as is the last resort in the ecosystem. So, if there's anyone else in the ecosystem that can solve the tagging problem or the enterprise use case, we encourage others to go do that because we're ultimately the technology provider in the space, and so we benefit from that. Where we come into play is typically when the ecosystem says they can't do something, then the enterprises tend to come to us and go, like, Hey, the ecosystem, nobody's able to solve it. And if we look at the problem statement and go, like, Okay, yeah, this actually requires us to advance the technology to solve it, and it's not something that you can solve with the existing state of the technology, that's where we get involved.
And typically, one of the benefits we have is because we participate on both sides of the link where we have the technology for the read point solutions and then also the technology for the tagging, we can jointly optimize and we can therefore advance the technology and open up the envelope of what's possible to do. And then once we do that advancement, then we deploy that out and enable the use case. So, we're very judicious about which enterprises we work with. We like to have an enterprise that's a leader in the space because they're usually coming to us because they're convinced that the ROI is there, they're convinced that this is a use case they want to deploy.
So, we know our efforts are going to be monetized quickly, and then they typically tend to be leaders in the space that once they do it, everybody else kind of follows along. So, where we see ourselves being pulled in, for example, today we're working with a particular grocer that wants to do, has a very innovative thought process around how they want to do point of sale, and they are heavily private label, which also makes it a lot easier, right? Because if you own your own supply chain, you can make the migration happen a lot quicker than if you have to deal with tens to hundreds of brands that you have to convince to go tag for you.
So, those kinds of opportunities we see as being really good because it gives us. There's a clear pull. There's a clear understanding of what the problem statement is. There's a clear direction that the technology has to go that we can enable, and then it has a very broad applicability. So, that's typically where we get involved, and we are therefore advancing the point of sale solution utilizes both the read point and the tagging capabilities at the same time because in some cases it's a challenge to tag and in some cases it's a challenge to read.
Yeah. Okay. Just still within food, I mean, if we rewind 12 months ago when we were thinking about food, talking about food, at least more, it was around QSR, it was around tracking inbound product. This is a much more high volume opportunity, is that fair to say? But I'm just curious, any update around where the QSR opportunity is and kind of food safety drivers around that opportunity as well?
You can take the volume.
You go ahead.
So, on the volume opportunity, right, we had anticipated that food deployment would go in that direction, like, you know, it would start out with pallet level tracking and then get to a carton level, and we hadn't anticipated that item level would be a pull. But just where the grocer sees waste and where they see the biggest financial opportunity is going straight to the item level, so we can kind of skip past for that, and we started with the item level, and that's where the ecosystem has been pulled. And so we do, and to your point, it's a huge volume opportunity, and that's why you see so much of the ecosystem rallying behind it to try and enable it to go.
But we still see it being a very long deployment because you have to solve the challenges associated with a particular category, and if you can do it category by category, you know, starting with baked goods, then going to the short-term perishables and so on. We do see that being a long tail of deployment, but along the way, each of those stages are pretty significant volumes even compared to what the industry does overall when you look at all the other segments combined. You know, in its full deployment, we anticipate food being trillions of items a year, right, as compared to the other opportunities, which are hundreds of billions of items a year.
Yeah. Within the food category, we see opportunity in terms of pilots and deployments active right now across three vectors. The grocery, which Hussein just mentioned, QSR, it's a different use case. It's for food traceability so that they can identify, isolate, and contain contaminations, and then it's in food logistics, gaining the same type of efficiency that you see in other logistics places. And there are, as I said, there are active pilots and deployments across all three right now. Yeah. I want to shift gears just a little bit. Just you've got the M800 chip that's now out, you've talked about, which is a progression from the M700 just to level set. It has new capabilities, it has different potential margin profile differences better than its predecessor. Just kind of talk about how that transition, you know, has flowed.
You've taken some efforts also to try to stimulate demand and adoption there at a more accelerated pace. Just provide us a little update on that effort.
Sure. So, the three, you know, if I look at the M800, first of all, it's a high performance endpoint IC, and what that enables you to do is it's solved for two different types of issues. One is we see it being utilized in cases where they're hard to tag items, and food's a good example of that. So, you need the performance to be able to just get to the readability that you're looking for. The other place we see it being used is that the increased performance can be utilized to build a smaller inlay to get the same performance that you previously had. And for a lot of the inlay manufacturers, the material cost is the primary cost driver. So, if you can build a smaller inlay, it reduces the cost of their inlays and makes them more competitive.
So, we see a pull on both of those cases and in both cases, and our M800 deployment's gone similar to our previous generation deployments where we see it primarily starting out by targeting new use cases and new opportunities first. So, people use the M800 to try and win new opportunities first. And then typically in the second year of deployment, as the existing opportunities, the ones that are high volume and recurring, those are the ones where they will go in and try to requalify the M800 because of the cost benefit to replace the existing volumes and cannibalize the previous generation. But we still see a very long tail of the older generation products where it's a very fragmented ecosystem, and so the qualification efforts and the redesign efforts are just not economically worth it. So, people just leave it alone on the older generation.
In fact, you know, our Monza R6, which is now two generations behind the M800 and was first introduced in 2014, we still ship that. And we ship it because of exactly that reason. They're going to a long tail of applications that are small volumes each and just not worth requalifying. So, we kind of expect M800 to follow the same pattern, and we're seeing the early deployments now. For the M800, we expect 2025 to be primarily new opportunities and then 2026 to start to migrate and add substitution and cannibalization.
Okay. And then the gross margin improvement path that you've talked about, Cary, what's the latest update?
The foundation for that is the M800 has 25% more die per wafer, with the wafer being the primary cost for our ICs. When it fully ramps to volume running status, basically replaces the M700 as the volume runner, we expect 300 basis points of gross margin accretion due to that cost advantage.
I want to make sure before we drift too far from food, I've got a few inbound questions, so maybe cover that. Just the competitive backdrop within food, does it differ than what you encounter today in apparel and logistics? Is there any differences there? Or will there be any differences there?
I would say, first of all, just given the sheer volumes and the low margins associated with food, we do expect to see continuing price pressure, and we expect to see competitiveness there. We feel we're well positioned with our offering both to overcome the technical challenges and be able to enable the ecosystem to be able to be cost competitive because ultimately it's the overall solution cost, it's not just the Endpoint IC cost. As I mentioned earlier, there's things we can do on the Endpoint IC side that helps with the overall solution cost. Right now I would say that that's going to be a much bigger factor in terms of the choices of what goes into food tagging. If you can get to the performance level, then price is going to be the next consideration.
And so, I think fundamentally from a competitive landscape, that's not all that different from where we see the rest of the landscape. It tends to be the same thing. If you get over a performance threshold, then it becomes about price and that's where you compete. So, to first order, it's the same competitors, it's the same kind of dynamics. I would just say that the sheer size of the volume and the thinner margins has incented people to move a little quicker.
Yeah. Yeah. I think this relates to food as well. Explain RFID versus Ambient IoT, or does that relate to food? We come back to that question.
Yeah, I'm noodling on it. The way I look at it ultimately is it depends on the type of solution, and different people have used different types of solutions. You see solutions that use vision, you see solutions that use Bluetooth Low Energy, you see solutions that use our technology. Our thought process is there's two categories that we look at. One is we look at what is the balance between infrastructure cost versus tagging cost. So, if you're in an environment where you have high turnover of items and high volume of items, then we believe our technology is well suited to that because our tagging cost is in the single digit, low single digit pennies, and our infrastructure cost is much higher.
You're looking at hundreds to thousands of dollars per read point, but that's a once and done thing, whereas your volumes tend to dominate your operating costs. If you look at Bluetooth Low Energy as an example, what ends up happening is that the tagging cost is much higher, but the infrastructure cost is much lower. So, if you were doing like asset tracking, that would be a better, that could be a better solution for you because then you have a better balance in the cost structure. People who use vision, vision is great. We see vision as being something that's a complementary technology, not an alternative technology. Because ultimately vision is great for being able to see things. So, for example, if an item is damaged, you wouldn't be able to tell from our technology that item is damaged or not.
But by the same token, you know, if you're getting a shipment of a case of, you know, call it 20 milk cartons that is in a package, you have to open it to be able to see it through a vision thing. With our technology, you don't have to open it. Just the fact that it enters through the dock door, you read it, you see how many items are in it, and you get the overall solution. So, what we believe the future will be is a blended technology deployment where you choose the right technology for the right applications, and there will be multiple technologies to solve it because I don't believe any one technology is going to be able to solve all of them in an economic way.
Okay. Early views around 2020, 2025, Cary, just, you know, historically the industry's grown unit volumes near 30%. As we think about, you know, 2025, is there puts and takes that we should think about at a high level against that, you know, historical metric?
Yeah. We feel pretty good that the historical unit carrier can be achieved. And we believe that because there's still growth in apparel, there's a lot of inertia in apparel as we've discussed with high logo penetration, but moderate volume penetration. We think there's opportunity within general merchandise, not just within the first two phases from the North American retailer, but expansion phases within that, as well as piggyback deployments from others leveraging their work. We see growth in supply chain and logistics with the first or the second large North American supply chain logistics company. And now we have food coming onto the scene. Now, the food volumes will be modest compared to the rest of the business, but that's a growth vector that just hasn't been present in the past.
We feel very comfortable with three major verticals moving forward that unit CARG can continue for some time.
Yeah. I guess just, you know, lastly, this question as well that came in, just, you know, broadly, you know, the barriers to adoption. I mean, what are the risks still around adoption that customers are struggling to kind of get over the hump?
Sure. So, I would say the most immediate one we see is the capital expenditure to put in the read point infrastructure. And that's one of the things we're actively working on is our belief is we want to have ubiquitous read points, high performance at the lowest cost possible. And that's something we invest in pretty aggressively. One of the reasons we continue to have a reader IC product line is to be able to bring that to bear. The second one that we see is that being able to solve the hard to tag, hard to read use cases. And there's still a lot of headroom in the technology, and there's still a lot of demand to be able to utilize that headroom to overcome those challenges. And then the third one is being able to fully benefit from the deployment of the technology.
If you look at the early days of the technology, primarily people were using it as handheld driven inventory counting, and once they were able to do that, then they started to realize it's like, hey, this has a lot more potential, so you see some of the retailers who've gone to 100% tagging going, I can do loss prevention with this, I can do self-checkout with this, and they've even advanced beyond that. They want to do, you know, they have ideas around fitting rooms and so on and so forth, so there's a huge opportunity and an expanding opportunity of what they want to do and envision the technology could do. All of that requires innovation, all of that requires integration into their business processes, and we see all of that momentum still there.
Okay. Super. We're at time. We'll break there, but I want to thank you, gentlemen.
Thank you very much [crosstalk]. Thank you.
Very informative.