Advanced industrial equipment. Again, pleased to have Impinj Technologies at the conference. And just quick background: Impinj is a leading developer and enabler of RAIN RFID. The company connects billions of items, providing item-level visibility and real-time information to businesses about the items that they create, manage, transport, and sell. And so, happy to have Cary Baker from Impinj with us, the CFO. Cary may—he's going to open with just a couple of comments to level set us, and then, we'll jump right into Q&A.
Okay. Well, thank you, Rob. Thanks for hosting us today, and thanks for everyone joining the call in person or via the webcast. As Rob said, Impinj is a leading RAIN RFID provider. Our mission, quite simply, is to connect everything. To date, our platform has enabled connectivity for over 100 billion items across verticals like retail apparel, retail general merchandise, supply chain and logistics, food, baggage tracking, toll tags, and a number of items and opportunities beyond that in the long tail of applications. Our technology leverages the RAIN RFID technology. We deliver our connectivity on a miniature radio chip smaller than a grain of sand. Our partners take that chip and attach it to an antenna about half the size of your pinky. The chip is readable at up to 30 feet, at up to 1,000 items per second without line of sight.
It harvests its energies from the radio waves provided by a RAIN reader, making it effectively battery-free. Our platform includes both sides of the radio link. We deliver the chip, which we call an endpoint IC, and we deliver the systems that read the chips. Think of that as readers, gateways, and reader ICs. In terms of a revenue mix, approximately 75% of our revenue in a given quarter is made up of the endpoint ICs, which we think of as recurring revenue, and the other 25% is from the system side. Our endpoint ICs sell for pennies, our reader ICs sell for tens of dollars, our readers sell for hundreds of dollars, and our gateways sell for thousands of dollars.
Perfect. Perfect. And again, if you have any questions, send those up, and we'll work those into the conversation. Cary, maybe just start with touching on current tone of business. The, you've seen a couple of quarters now where you've seen sequential growth in the Endpoint IC revenues. You're guiding the third quarter, the June quarter, to be—or, excuse me, June quarter to be a third quarter, I guess, of sequential improvement. It appears, you know, you're seeing some improvement across multiple verticals also, but just, you know, could you dig it a little bit deeper in terms of what you actually are seeing in, in terms of the recovery, how it's playing out, how broad-based actually is it?
Mm-hmm.
just kind of, level set there.
Yeah, we're seeing the recovery really come from two areas. The first in general merchandise. We're beginning to see that ramp take hold. Q4 was the first quarter. Q4 2023 was the first quarter where I thought we met our expectations for the general merchandise ramp. Now, admittedly, those expectations have been lowered. In Q1, we saw continued progress, and the general merchandise ramp actually exceeded our expectations. And I, from this point, I anticipate continued modest growth throughout the back half of the year on our general merchandise deployment. This one we think is a multi-year deployment, so we'll, we'll look at 2025 as well. We're also seeing recovery in the retail apparel market. After a prolonged period of destocking, we're starting to see some rebuying. And it's really been those two factors that have driven the outperformance that we've seen through Q1 and into our Q2 guide.
What about kind of the status and partner inventory levels? We talked—that was a topic of conversation to a great degree. Last year, it seems like we've kind of moved past that, but can you state that every inventory levels are satisfactory across all your customers, or is there still some leaning out there that needs to happen?
We feel pretty good about inventory. We exited a channel inventory, that is. We exited 2023 with our large inlay partners, relatively healthy across the board. I think of our large partners as driving the bulk of the volume. Our smaller partners made progress in the first quarter. Their demand is typically project-based, so it took them a little bit longer. So we felt good exiting Q4. We felt good exiting Q1. We're in pretty good shape on channel inventory. From an end customer and end retailer perspective, we still think we're digesting some inventory. We track the retail apparel import data, and while it's up versus last year, we're still down versus 2019, which was the last normal year of retail demand.
You compare that with what we see in retail apparel sales, and even on an inventory or inflation-adjusted basis, retail apparel sales appear to us to be up versus 2019, which signal to us that there may be some destocking going on at the end retailer level.
Is there a point of view as to when that may resolve itself?
Not yet. I would say the signals remain mixed on that, and you can read any of the prints in there. They represent a mixed status as well. When we do start shipping to that, our shipments more closely aligned with end customer consumption, we might expect an increase in demand, but we're cautious on that. We're a couple steps removed from end customer demand. So we're going to take the tack of not predicting when that recovery might occur, but reporting the news when it does occur.
You kind of understanding you only guide one quarter forward and not for the full year. And, maybe take this from an industry perspective, if you would. You know, historically, you've seen the industry grow unit volumes kind of close to 30%.
Mm-hmm.
RFID tag chip unit growth close to 30%. Just, where do you think we are, or where do you think the industry is, you know, laying out for this year relative to achieving a growth rate that's going to maybe similar, maybe not? I'm just curious what your perspective is, you know, for the industry as a whole.
Yeah. Last year, we were able to deliver growth above the historical unit CAGR, despite our number one vertical, retail apparel, being down significantly for the year. We were able to do that because we layered on volumes into supply chain and logistics for the first time, and we also rebuilt channel inventory. Our channel partners had entered the year after a period of supply constraint, relatively thin on their channel inventory. So that helped us deliver last year's growth rate that again exceeded the historical unit CAGR of approximately 30%. As I look to Q1, Q1 will be a tough year-over-year compare. Q1 was a tough year-over-year compare for us, and our volumes were actually lower because in Q1 last year, the bulk of that inventory rebuild for our channel partners occurred.
In Q2, we have signaled product revenue, so excluding the license revenue for endpoint ICs to be at or near record levels. That's still against a pretty tough compare because in Q2, our channel partners overbuilt inventory before taking that down in the back half of the year. Now, the comparisons get easier in the second half of the year, as you might expect, because we were taking down channel inventory, but it's too early to project at the full year or how, how the full year's growth rate will compare to the historical unit CAGR. What I would say, though, is that 25%-30% historical unit CAGR, we see no reason why that can't continue over the long term.
Yeah, you kind of touched on it earlier with the general merchandise comment, but, you know, there is one particular deployment that's ongoing there. I'm curious if, you know, the commentary around general merchandise ex-ex really extends beyond that, or are we mainly talking around that one deployment? But really, you know, the heart of the question was, you know, from a technology standpoint, deploying in general merchandise applications versus retail apparel, which is more, you know, a lot of the legacy volume. I'm just curious, what are some of the learnings there that have gone on because it has gone a little bit slower? Again, that's kind of a technology, as well as just, you know, the breadth of players involved there, challenge. What are some of the learnings, and do we get past some of those hurdles on future general merchandise deployments?
Yeah, great question. So the general merchandise ramp was slower than we, the entire ecosystem, and the end customer anticipated. It was slower because compared to retail apparel, general merchandise increases substantially the number of suppliers involved, the number of SKUs involved, as well as the complexity of the product packaging of those SKUs. And we had suppliers tagging products for the first time, and the industry tagging products or general merchandise SKUs for the first time. So there were some natural learnings that needed to occur that we underestimated. What's the best way to tag a board game with the barcode embedded in the packaging? What's the best way to tag a package of sticky notes that are wrapped in cellophane? Those learnings had to happen.
And I'm happy to say that it appears that a lot of those learnings have happened now. As a result, we're seeing the ramp in general merchandise, and we expect it to continue ramping. But those were some of the challenges that we saw. So even though the store employee motion is exactly the same, you're using a handheld reader to count the items on a store shelf, we had to figure out how to tag items for the first time. Now, as I look forward, the learnings that we've made today or up through today will be applicable for anybody else in general merchandise that wants to tag their inventory. So I would anticipate future deployments have the ability to go faster as a result of the learnings that we as the ecosystem have made so far.
What have you seen around just the, you know, the business case around general merchandise? How has that evolved, and how has that, maybe inspired some, you know, new prospects, new customer interest in that application?
The business case is very similar to retail apparel and in the initial use case of retail apparel, and that is counted items at the store level in a much faster way. You think about how retail stores have historically counted the items on the store floor. They typically shut the store down early, bring in a bunch of extra labor, and hand count everything. With a RAIN deployment, you're now able to equip a store employee with a handheld reader that can walk up and down the items, and depending on the quality of the reader, can read 1,000 items per second, again, without line of sight, up to 30 feet away. So you've dramatically improved the ability to count and track the items on your store floor.
And what we've seen in apparel and what we expect in general merchandise is that without a RAIN deployment, inventory accuracy at the store level is mid-60% accurate at any given time, given the complexity of hand counting the item. With a RAIN deployment, you can improve that accuracy into the high 90%. And we actually have examples of some of our end customers that are 100% deployed, are inventorying their store twice a day with a RAIN deployment. So we think all the benefit and value that was created by apparel will be replicated in general merchandise.
I wanted to shift over to logistics. That's the, that's another, kind of lighthouse customer activity, or project that's been going on. Can you give us any feel for how, you know, this would, will influence endpoint IC revenues in 2024?
Yes. So, logistics, as I mentioned earlier, ramped for the first time in a meaningful way last year. This was with our second large North American supply chain and logistics customer. They're not complete with their Endpoint IC ramp, and we expect their label consumption to continue in 2024 as well. As we look broadly to the supply chain and logistics space, we see other players looking at the comments that this company has made very publicly about the ROI that they've already achieved and the ROI that they're yet to achieve. And we think this bodes well for more opportunities in the supply chain and logistics space.
How should we think about, in an application like that, how should we think about what your share position could be? I don't know if you've commented on that or what your, you believe your share position, you know, will be in that application?
We haven't commented specifically on that, but what this is, a platform opportunity for us. Think of a platform opportunity as a customer who uses our endpoint ICs and our reading solutions, whether that be a fixed reader that we build or a reader IC that goes into one of our partners' handhelds or fixed readers. When we have a platform opportunity, we're also usually deploying engineering resources to help bring that solution to life and work out the kinks, if you will. When we engage in that type of an opportunity, we expect high endpoint IC share.
Yeah. I'll touch on another, you know, high-profile project that, you know, is often mentioned, the European retailer who's been, you know, pretty forward-thinking in their use of RFID and even to go beyond kind of the typical inventory accuracy use case, with self-checkout loss prevention. That one's gone about in several phases. Where are we right now in terms of the rollout, closing out a particular phase or starting another one? Or, just how do you think that can bridge itself this year?
Mm-hmm. That program's gone very well. We're wrapping up the third phase of that program. Think of this as the customer starting first with their smallest brand in phase one, seeing success there, rolling it out to their flagship brand in phase two, and then phase three encompassing brands three and four. Now, originally, I thought phase three was going to complete in Q2. I now think it's going to spill over into Q3 on its completion. There will be a phase four we have been notified of. Phase four will kick off, I believe, in the second half of this year. I'm modeling phase four, though, at a little bit of a slower pace. Phase four has more deployment logistics complexity. It's deploying to four smaller brands, and I think that will have an impact on a slower pace.
But we're still building with the customer that what that project timeline looks like, and I hope to have more information on that timeline when we announce Q2.
Have they, has the use case itself broadened?
Yes. The use case started with loss prevention. It also includes self-checkout now in some cases. And this customer is on their journey to start embedded tagging. So they're taking the tag off the item or off the hang tag, and they're embedding the IC into the garment.
I see. You know, just along those lines, and this, you know, I recall it particularly at NRF, the National Retail Federation show back in January. RFID had a, you know, frankly, an elevated profile, I thought, this year than what it had in the past. And listening to some of the customer panels, you know, one thing that seemed to resonate with me was, inventory accuracy was kind of the gateway use case that gets everybody started. But then once they start down that path, they seem to find so many more benefits or variables that can factor into an ROI.
Beyond those couple of that you mentioned, it, you know, do you have any other examples that customers have come back to you with and said, you know, we didn't expect to receive this benefit, but now we're receiving it and it, you know, has really supported further rollout?
Yeah. It in retail, apparel, to your point, they the retailers always start with inventory accuracy at the store level. It the ROI is well understood for inventory accuracy. The path deploy is well worn. And at this point, it's relatively easy to do that. Most of the retailers are then on a journey to get to 100% tagging. They can extract value from the deployment with a partial deployment, even before they get to 100%. So they start receiving the value of the RAIN RFID, but most of them are on a journey to 100%. The reason they're on a journey to 100% is because 100% unlocks all the other use cases: self-checkout, loss prevention, loss analytics, loss detection, front store, backstore management.
Once they get to 100% tagging, they can start looking at those use cases, and then they're extracting even more value from the IC.
Yep. That seemed to be the case. You know, particularly the conversation around, you know, shrinks obviously been very topical of late. Just to be clear, I mean, your solution is not necessarily, and maybe in some cases it is, providing physical, physical security, but it's more around the tracking and the information that comes off, movement of that item, as well as maybe identification of where that item is leaving the premises, I guess can be, you know, a benefit as well that they're coming, they're deriving.
Yeah. I think the number we call it loss prevention, but the number one value we provide, and it doesn't have to be in a 100% tagged environment, is loss detect, loss detection. What item was stolen, when it was stolen, how many times has it been stolen? If it's paired with a camera system, who stole the item? With that missing information, the retailer can adjust their security posture and better protect themselves against theft. We saw in one of the panels, the head of security for a large retail store that typically operates on multiple levels in a mall environment found that they had security on the first and second floors because that's where most of the items were located that had been stolen.
On the third floor, they didn't have security at the door because that's where the furniture was and no one was walking out with a couch. But what they found when they put readers at the doors is items were being grabbed from the first floor, taken all the way up to the third floor, and then being walked out of the store and stolen. So what they did was they, with that information, they were able to adjust, as I said, their security posture. They were able to put security at the door. And while they don't have a 100% deployment at this time, they were able to positively impact shrink.
Yep. Yep. That was pretty compelling cases that I heard. You've got a new chip that's recently been introduced. It's the M800, follows the M700, which was introduced not too long ago, which was a, you know, pretty significant advancement as well. Just, you know, kind of update us on the M800 introduction. And I'm curious, you know, more extensively, just kind of the capabilities that it brings you and what it unlocks, and then also the margin, you know, potential behind that and how that might scale.
Yeah. So the, the M800 is the next chip for us, in the 65 nanometer node. It's the evolution of the M700. It will address the same markets that the M700 addresses. So retail apparel, retail general merchandise, supply chain, and logistics. The M800 has improved sensitivity, which means a smaller antenna can be used to get the same read range. So our inlay partners can take cost out of their model by, by leveraging that improved sensitivity. We are at the early days of the ramp. We're shipping production volumes today. In Q1, we announced and we'll increase those volumes in Q2. They're still very small relative to the billions of chips we ship every month. But, as we make our way through the, the quality and the certification process, we expect the M800 to ramp.
Historically, when we've launched a new IC, the new IC goes into new sockets first and then gradually over time replaces existing sockets. So historically, we've had a multi-year ramp. We're trying to accelerate that. We think there's a lot of value in the M800 for our inlay partners and for our end customers. So we're trying to accelerate that curve. One of the other reasons to accelerate that curve, in addition to benefit to our customers, is the M800 is a much smaller die. We get 25% more die per wafer. That is a huge cost advantage given that the wafer cost is the majority of our IC cost. So when fully ramped, we expect the M800 to deliver approximately 300 basis points of gross margin accretion to the corporate average. That's a reason for us to try to accelerate that adoption.
It's too early to say whether we'll be successful. I need a little bit more time to understand how this ramp is going to happen, but we're very excited for the chip. We're very excited to get it in the hands of the end customers. We're on our way.
And, the levers, or some of the levers that you're using to accelerate that adoption would be.
So first is in the chip. So this increased sensitivity. So it allows our inlay partners to take cost out. 2, the M700 was the predecessor to the M800. The M700 was the smallest chip launched in the market. There were a lot of learnings on how to industrialize that smaller chip that we went through. We've built or baked all of those learnings into the M800. So this is an easier chip for our inlay partners to industrialize and turn into a tag. So we think it's easier for them to handle. And then finally, we've been, tried to be smart about the pricing on the M800. And we've priced it, even though it's a higher performing, easier to handle chip, we've priced it slightly below, emphasis on slightly, slightly below the M700 to try to drive adoption.
You've also introduced not too long ago, you know, a pretty unique solution around authentication, which is a unique use case in and of itself. I guess, there was some pilot activity, that had gone on. Where, where is that currently stand in terms of its adoption? Maybe, I don't know if you can speak to how many customers are piloting it, how that's broadened out, but just give us an update on that.
Yeah. So first, the M775 is a cryptographic authentication chip. So it's an Endpoint IC with a crypto standard ISO Crypto Engine in it. It's a reader with an algorithm that can engage a challenge response to that IC. And then it's a cloud service to point to that holds the key to validate the authenticity of the item. The goal is to take fraud out of the market, where I would say at a high level, we're still very early days. We've got very interesting pilots in performance apparel, footwear, in specialty food, in tax tracking, and in medical that we're excited about. I would say we're still in the pilot stage of this. So it's too early to call where this might go and how quickly it might take.
But we're pleased with the fact that we've shipped into trials, into pilots, you know, hundreds of millions of chips at this point.
Okay.
Now we're waiting for those pilots to materialize and see what we need to do to accelerate adoption.
I'd probably be remiss here with all the discussion around technology. Just, you had a victory here of some magnitude with your largest competitor around some patent and IP infringement, you know, which you prevailed and, you know, just there'll be an upfront, there was an upfront settlement payment to Impinj and then they'll pay you annual royalty or license fee, I guess, you know, for that use of that technology as long as they continue to use it. And I think that $15 million is annual payment that could scale from there, you know, to the extent we think about that as kind of a recurring in nature payment, what is the, I guess, how difficult is it for that competitor to move off your technology, shift to a different technologies given the nature of the patents?
Yeah.
You know, not speaking necessarily for their own roadmaps, but just curious how durable that is.
So let me first give the details of the settlement. So we settled with NXP, after a prolonged patent litigation. They paid us $45 million upfront, which we recognized in Q1 as other income cash. And they agreed to pay $15 million per year, flat fee that increased by a modest fixed rate annually for as long as they use the IP or 10 years. They have the ability to early terminate by designing out our IP. Now, the question that you've asked is how long will they use our IP? And I, unfortunately, can't answer that question. That's a question better suited for NXP. What I would say is we've given a reasonable royalty rate that continues to be reasonable as time passes.
It will be up to them to determine whether they accept that reasonable rate or go through the effort of designing a new IC that designs out our IP and then migrating all of their business to that IC. That's a question only they can answer.
Sure. Understand, you know, a couple of things that have been a little more topical of the last couple of quarters. They're longer term in nature, but you know, you've been more willing to speak about them. That's both in the, maybe just real quickly, why would that be impactful to Impinj from an RFID standpoint, its passage as that plays out? And I'll let you expand on that one and then another longer term.
It's still very early days, but DPP is an initiative in Europe to track an item basically from cradle and grave and give end customers or end users the ability to engage that item. It's first going to go to lithium- ion batteries and then textiles. Textiles is interesting because our number one vertical is retail apparel. It's exciting for all of those reasons that you might guess, but what's most exciting about it to me is the consumer engagement requirement. That means that you and I, it has to be meaningful to you and I, and you and I have to interact with that. I think the most likely way for that to happen is for our phones to become readers. And when that happens, it opens up a whole bunch of use cases for us.
It is way too early to project which way this will go, whether RAIN will be the technology selected, but it's exciting because it's one of the first opportunities out there that has the potential to drive end consumer engagement.
Just as an ancillary benefit, if phones started to embed RFID reading capability, would that be an opportunity for you to provide that technology or is that more with the phone chipset providers?
It could be with us. You know, it could be with the phone companies. It's way too early to say that. What's exciting though is the Endpoint IC opportunity from you and I having a reader in our phone.
Right.
The information that we can either put in the chip or allow the chip to point to just opens up the aperture for RAIN adoption significantly.
Yeah. Maybe in the interest of time, I just wanted to touch real quickly around margins. There's still, I think, below, there's still below maybe target margins, gross margins that is. What are kind of the pathways to get that to target level?
Yeah. So our normalized gross margin today pre M800 should be 53%. To your point, we're below that. We're below that for a couple of reasons. One is we're still subscale from an operations and internal operations perspective. But we're closing that gap quickly. And once we get back to first half revenue from our first half last year revenue levels, we'll have reached the scale that we need. The second piece is what we've seen is, you know, coming out of a downturn, Endpoint IC recovers before systems. As a result, our mix of Endpoint IC revenue is higher than normal right now because we've yet to see the systems recover. Endpoint IC gross margin is just below our corporate average, whereas systems is above. So we've got some overall mix headwind in that perspective.
Then the final piece that's pressuring really Q2 and Q3, and then we'll be done is we have a higher mix of 200-millimeter ahead of the M800 ramp. 200 millimeters on an older process node, it's a much larger die, therefore it has a lower gross margin. We are moving some of that 200 millimeter before it is two generations old relative to the M800 when it ramps. So we just have a higher mix of 200 millimeter in Q2 and very likely in Q3.
Okay. But Q4 though, we'll, we should be past that. Okay. Very good. Well, we're at time. We could keep going, but I hear happy hour also starting to get cranked up. So maybe we'll in there. Thanks, Cary.
Yep. Thank you very much.