Welcome to the Impinj Session. I'm Rob Mason, the senior analyst at Baird, covering the advanced industrial equipment sector. Pleased to have Impinj with us this morning at the Baird Industrial Conference. And just for those less familiar, Impinj is a leading developer and enabler of RAIN RFID, which connects billions of items by providing item-level visibility and delivering real-time information to businesses about the items they create, manage, transport, and sell. And very glad to have with me up on stage is Cary Baker, Impinj's CFO this morning. Cary's gonna open with just a few introductory comments, and then we'll move into Q&A. And if you have any questions, feel free to send those up via the iPad, and we'll work those into the conversation. So-
Yeah.
Morning, Cary.
Morning. Thanks, Rob, and thanks for having us here today. Just to say a little bit more, as Rob mentioned, we're a leading provider of RAIN RFID chips and products. And just to say a little bit more about what that is, we make the digital evolution of a barcode. But unlike the barcode, we can be read up to 30 ft away without line of sight at speeds of up to 1,000 items per second, and our ICs are smaller than a grain of sand, so much more efficient. We're helping businesses track, trace the products they sell, transport, and make. Our number one vertical today is retail apparel. We think that vertical's anywhere between 25% and 30% penetrated today, well on its way to deeper penetration.
Beyond that, there's a long tail of opportunity for us, first in general merchandise and supply chain and logistics, but beyond that, in pharmaceutical and medical devices, in automotive, and many other applications, beyond those that I've just mentioned.
Very good. Very good. Again, if you have any questions, send those up. We're at session four at rwbaird.com. So, Cary, maybe just to start off, level set on where things stand in the business coming out of the third quarter. Certainly, as we've gone through, you know, this year, retail, retail apparel in particular, has, you know, that sector has had some challenges and slowed things, and you've gone through your own kinda inventory destocking around endpoint ICs. That seems to be, you know, maybe coming to a conclusion around the destocking aspect. Still, maybe at some of the smaller partners that you deal with, there's some inventory still there. But, you know, you talked about on the last conference calls, you know, some green shoots starting to emerge. So I...
You know, maybe just as we get to the end of the year, how should we think about destocking from a standpoint on a go-forward basis, you know, within the business and how you're possibly set up, as you enter 2024 on that front?
Yeah. So on the channel inventory destocking, we made progress in Q3, but I would say our progress was mixed. We had some players, or some partners on track or ahead of schedule and some behind. We'll continue to make progress in the fourth quarter, and I think our larger partners end the year reasonably healthy. I think our smaller partners are gonna take just a little bit longer, probably into early first part of next year, before they're healthy. As we look at the business and the industry, we are starting to see some green shoots, which is encouraging. And just to delve into that a little bit, we see green shoots in kind of the macro data that's available to us.
The retail import data, which is a measure of production for us, has improved for the last six months. We're still down 10%-20% from where we were a year ago, but this is much better than the down 30%-40% where we were at the beginning of the year. We're further corroborating that with our conversations with our end customers, with our service bureaus, and with our inlay partners, you know, throughout the third quarter into the early part of the fourth quarter. What we're seeing is that the U.S. market is stabilizing, Europe's maybe a quarter or two behind, and then Asia's mixed, with China still being pretty weak. It's probably still too early to call a bottom, but we're very encouraged that we're starting to make progress again.
Just around the inventory levels, if we think about that in terms of weeks of inventory, you know, a normal range that you would be in, where do you stand today relative to what a normal range would be?
We're historically a business that turns 50% in a quarter, so normal inventory would be in the six to eight weeks range, based on forward-looking demand. I've stopped talking about the number of weeks recently because it's based on a partner's estimate of demand, and that demand dynamic remains pretty fluid. But as I said, we're making progress, and I think that the larger partners will be healthy by the time or reasonably healthy by the time they exit the year.
You know, I think it was interesting you made this comment as well on the conference call. You know, with all the inventory destocking that has occurred, there's also been some larger projects that have shifted to the right. Yet you still said your volumes would be up, you know, close to normal range-
Mm-hmm.
- which is in the high 20%. So I guess just, you know, underlying that, you know, what's driven the business this year around all this, you know, peripheral noise?
Yeah, it's a significant metric. The historical industry unit CAGR has been 29%, and we're tracking consistent with that this year, despite our number one vertical today being down 10%-20%, but through the early part of the year, down 30%-40%. So we're very pleased with that growth. Now, to be fair, some of that growth is related to rebuilding the channel. We entered the year with the channel inventory very depleted after being heavily supply-constrained last year. So part of our growth is related to that rebuild that occurred in the first part of the year. However, even if you strip that out, we're still growing on a year-over-year basis. We're growing because we've become more diversified in our revenue base. Historically, two-thirds to three-quarters of our revenue has come from the retail apparel market.
This year, we're starting to see significant volumes of endpoint ICs go into supply chain and logistics, and this year, that's provided an offsetting function to the softness that we've seen in retail. So it's exciting that we're becoming more diverse, and we'll continue to track down that path, next year.
I made mention of, you know, some of these larger projects as we've gone through this year have pushed to the right a bit.
Mm-hmm.
How are those shaping up as you enter 2024? You know, will they be on the agenda to be able to be addressed?
Yeah.
Next year.
So let me break it into three. First, with the general merchandise rollout. This one's gone slower than we or anyone in the industry have expected. It's not for lack of urgency from the customer or not for lack of effort from the ecosystem. What we're seeing is a steeper learning curve as we ramp up suppliers outside of apparel, which have been deploying for ten years at this point, but suppliers in the general merchandise space. Think, toys or board games, electronics, stationery, the hundreds and thousands of suppliers now that are tagging items for the first time, that ramp has proven more challenging. Now, it's an all hands on deck situation for the industry, and we're all working to get these suppliers up and running, and I'm confident we will.
There's no lack of effort from anyone in this part. And perhaps the silver lining is, as I said, we're gonna deliver 2023 unit growth rates consistent with historical averages, yet general merchandise, which is a vertical that we measure in 325 billion unit opportunity per year, the general merchandise ramp is still in front of us. And then the second project that we talk about often is our second large North American supply chain logistics company and their rollout of endpoint ICs. That project is on track, progressing nicely, and it's been a very much a bright spot for this year for Impinj. And that project won't be complete this year. We'll continue ramping next year.
I don't know that we'll be complete next year either, just given the size of that program. And then the third project that we talk about is our loss prevention deployment with the visionary European retailer. We're in the middle of phase three right now. That deployment is on schedule. There's more opportunity for additional phases to deploy, as that customer moves further into their brand, into their brands. And in that, we're now starting to see the tagging flow through from that opportunity, really starting in the Q4, but that'll pick up in 2024.
If we could go back just to the general merchandise opportunity that you were referencing or that larger project, maybe, you know, larger customer, I think. We're probably somewhat familiar with, you know, what the driver is there. Given there's still a lot of complexity or scope, scale around getting that, is there... Do you feel like there's, you know, good visibility on that project, not slipping in the future? Could that push out, or could that, you know, potentially, the, I guess, the deadline for that.
Yeah.
Move again.
Yeah, the compliance date for that remains February.
Yeah.
But as I stand right now, I don't think we hit the compliance rates that are targeted. I just think we're too far behind on that. I have not heard that the compliance date has shifted to the right. I suspect there will be leniency granted for those suppliers that are making progress, and maybe pressure applied to those suppliers who have not yet made progress. But that's speculation at this point. We'll have to wait and see.
Mm-hmm. Well, you, you did mention it's kind of an all-hands-on-deck effort. Is there any sense that that's pulling partner resources away, away, away from other project opportunities, or does that factor into some of the, the slower inventory work down, that you referenced at smaller partners?
Yeah, I don't think it's related. I don't think it has an impact, particularly on the smaller partners. Those are more project-based, whereas the general merchandise ramp is with the larger players, predominantly. I think we, as an industry, underestimated the complexity of ramping a new supplier. If a new customer in apparel were to roll out today, chances are the apparel suppliers that they're working with have been tagging for years already, so that new program can go much quicker. But if we think back to 10 years ago, when the first apparel manufacturers were tagging items, it took a little while to get those suppliers up and running, and I think that's what we, as an industry, overestimated, and it's just taken a little bit more time.
Yeah. Yeah. And then with respect to the second-largest logistics customer, you said that will ramp through this year, maybe still not be complete and ramp further. I guess, what is the, w e know it's, you know, a large, you know, in terms of volume of tags. What is maybe some of the constraints around that project from moving faster? Any insight?
It's just a big project.
Just big.
They have their ramp this year has been nothing short of phenomenal on how fast they've gone, but this is a massive project, and even projects that we see that are in the 1 billion-2 billion unit range, they can't do that in a single year, and this project is multiples bigger than that. So, you know, I'm not at all surprised that we're not fully ramped after in year one. I'm very pleased with the progress, and I look forward to more progress next year in year two.
And, yeah, I think this customer's even talked about additional, you know, opportunity or places that they, you know, potentially could attach, you know, some reading capabilities. Is that included in your commentary around this ramp, or would that be on the, perhaps even the tail end of-
That would be the, what-
Fundamental to what you're speaking to?
Yeah. What I'm speaking to right now is the endpoint IC ramp. So this is putting tags on packages that you or I would receive at our house. That ramp will take another year at least. They have announced multiple phases on, from an infrastructure perspective, on how they want to expand their reading capabilities, and, they have talked about additional opportunities there.
Okay, so that could insert in-
Mm-hmm
... along with the IC ramp. Okay.
Potentially, yes.
And then maybe just to hit on all three of these projects again, the visibility around the self-checkout project, you know, moving ahead in 2024, just given the overall state of the retail industry and, you know, as we get into early part of 2024, you know, the presumption is that's when CapEx budgets get set. Is there a good visibility line of sight on this, you know, moving forward in 2024 around-
Yeah
... being funded, I guess, in other words?
We have, you know, we have not yet won phase four opportunity, but, you know, obviously we feel confident in that, given the success that we've been able to demonstrate to this customer in phases one, two, and three. They have just to back up, phase one was a rollout to one of their smaller brands. Think of this as the controlled test of RAIN after they were done with the pilot. Phase two was to their flagship brand. Phase two was 3x the size of phase one. Phase one was $6 million, and then phase three was two additional brands. They have more brands in their portfolio that I believe they have every intention of rolling out loss prevention to.
Okay.
We haven't won it yet, so I can't comment on the schedule, but it's certainly an opportunity that we're staying very close to the customer on.
This self-checkout capability overall, where have you seen the, I mean, this customer's been high profile, where you've talked about a lot. But beyond this customer, are there others behind it that are adopting as well, following this customer? Just give us some perspective on what the overall uptake has been around-
Yeah
S elf-checkout.
So this is the first customer deploying what I would call a high-end loss prevention solution. And I would say we're still in the early days of the self-checkout and the loss prevention deployments. And there are multiple solution, multiple RFID-based solution in addition to the high-end one that we talk about, that we deliver. The reason we're in the early days is the prerequisite for self-checkout and loss prevention is every item in a store has to be tagged. That's the only way that system can work. And there's a little more than a handful of retailers around the globe that are 100% tagged at this point. They're all on the journey to get there, but as we talked about earlier, the process takes a little bit of time.
So while we're in the early days, we see a lot of interest for self-checkout and loss prevention to be more efficient with their retailers' labor dollars. So what I think happens is it accelerates the path to 100% adoption in these stores as we go further. And I would say our pipeline reflects a lot of activity to that end.
Well, that maybe plays into my next question. Just, you know, this, this, dynamic that retailers are speaking to around shrink overall being a, you know, having elevated in terms of a challenge for them, how can RFID play into that, as a part of the solution? You know, does it require 100% tagging for it to really be, you know, part of the solution, or you know, can they make some, more, I guess, strategic, insertions of RFID to address that problem?
Yeah. So shrink is definitely a hot-button topic amongst the retail community today. Retailers wanna know what was stolen, when it was stolen, how many times it's been stolen, and ideally, who stole it. They want visibility to shrink. Visibility is exactly what RFID provides. Even in the base inventory management use case that retailers use, we provide visibility. And by that, I mean, think back to recent retail reports, and you probably heard retailers announce that they had a gross margin hit because shrink was larger than they anticipated, and they didn't realize the extent of that shrink until they did their annual inventory count. You don't have visibility when you count your inventory once a year. Contrast that to some of our customers who have fully deployed RAIN, just in inventory management. They're inventorying their stores twice a day.
So yes, shrink still happens to them, but they realize it much earlier in the process, and this is before we even start talking about a loss prevention deployment. With a loss prevention deployment, the added visibility is you see exactly when an item was stolen, you see exactly what item was stolen, you can track back to other stores to see how many times that item had been stolen to identify any type of organized retail crime, and then you can also identify who's been stealing it. So now you have the visibility to really start combating shrink. And as I mentioned, it's a hot button in the industry right now, and our pipeline activity reflects the opportunity that we could provide value to retailers in this, in this, in this area.
What is your sense in terms of, you know, where these, you know, prospective customers would be in terms of their ability to move forward, are they still in evaluation phase, or is that kind of the stage that we're in right now?
I think for inventory management and how that can help with shrink, I think they're past the evaluation stage. It's a matter of how quickly they can ramp. Most retailers are moving forward with RAIN for inventory management. As they move from inventory management into loss prevention, I think there's more of an evaluation process, because now you're starting to talk—you're changing the reading capability, or you're increasing the reading capability. Inventory management is typically accomplished with a handheld reader and a store employee walking up and down the aisles and reading all the items in the store. When you move to loss prevention, you are now putting pedestals at the door that have a fixed reader in there, and that reader has to house a very complex algorithm.
That algorithm identifies, out of the sea of thousands of tags that are at the front of the store, the one that's moving through the door that hasn't been paid for to trigger the alarm. So it's a, it's a more complex use case. It requires more infrastructure, deployment, and that's why they're, they're still in the kind of call it the pilot or testing phase of it.
As they move down that towards more of a loss prevention solutions, what are the impact on the economics to your business from an endpoint IC standpoint, or are there any impacts?
Yeah, from an endpoint IC, what it does is it accelerates the adoption. So we're, you know, as I mentioned earlier, we're 25%-30% penetrated in retail, yet most retailers in the apparel space, most retail apparelers are moving forward with RAIN. So I believe we'll get highly penetrated there. It will help accelerate our drive to higher penetration within retail apparel. There's also functionality in the chip that is proprietary to Impinj, that helps enable that loss prevention solution that we've discussed. That's our Protected Mode, the ability to turn a chip on and off, so it can flow through a point of sale, through the door, and then a return process.
So there's an opportunity for share gains with us, and then ultimately, there's an opportunity for infrastructure systems deployment as we're putting pedestal readers and pedestals for the front of the store.
Is there any difference in the margin profile of the chip itself?
Nope.
Okay.
Every one, every M700 and every M800 now have Protected Mode in it natively.
Very good. Just, I'm curious, you know, in this environment that we've been in, of late, just how sales cycles on some of the larger, RFID proposals, you know, are trending, if that's started to. If it is elongated, you know, if you're seeing any change in that, you know, as we think about finishing out this year, moving into next year, what might be in the pipeline?
There hasn't really been a change. You know, the first deployment of a use case or the first deployment in a new vertical always takes longer, and subsequent verticals or subsequent deployments based on the learnings from the first few, can always go faster. But at the end of the day, the sales and deployment cycle is long. It's long because we're changing how these businesses count and track, and trace the items that are most important to them, which is to say, no one makes this decision lightly. We typically see a long pilot to prove out an ROI and to work out the kinks, and then, based on the successful pilot, a recommendation is brought to the C-suite and sometimes the board to move forward.
And when the move forward decision is made, they first start with a small or controlled deployment, where they can prove out the ROI at scale before rolling it out to their entire business. And then they make the decision to go with the broad deployment, and as we talked about earlier, that broad deployment can take multiple years. So it is inherently a long sales and deployment cycle. The benefit is that we've seen no one turn back, and when we've got a customer deployed and seeing the benefits of RAIN, we've just created a long-term recurring revenue opportunity for Impinj to sell endpoint ICs into. So our patience is rewarded, and yes, we'll try to make that faster, but I don't, I anticipate it always being kind of an elongated process.
I wanted to turn back to... You made the comment around, you know, your chips, M700, M800. M800 is new to be-
Mm-hmm.
Ramping this year, or into next year, I think is, is the expectation. Just, how should we think about, one, maybe the pace that it could ramp, and then, two, what the benefits rare, how quick, and how quickly some of those benefits could be achieved?
So the M800 is our newest chip. It's like the M700, is geared towards retail and logistics applications. It's like the M700, built in the 65 nm process node. It is more sensitive than the M700, which means our inlay partners can use a smaller antenna to get the same read range. It's a smaller chip, more performant chip. We're obviously very excited about that. Where we are in the process right now is our partners have M800s. They've built inlays, and those inlays are now in the qualification or certification process. I anticipate them exiting that process in the Q1 timeframe, and then we'll start to see the M800 ramp. Now, when we have historically launched new chips, they go into new sockets first, and then eventually, over time, they'll replace existing sockets.
Just to highlight the longevity of our endpoint ICs, our Monza R6 was introduced in 2014. It's still selling volumes today. It's still winning new sockets today. So we typically see our endpoint ICs roll out and ramp in a layering fashion. Now, the M800, as it is the smallest chip on the market and as the M700 was previously the smallest chip on the market, we learned a lot about the manufacturability of chips that small. And we've built some of those learnings into the M800 to make that process easier for our inlay partners. Now, that combined, that improved manufacturability plus the improved performance of the chip, may accelerate that adoption, but it's really too early to say. I probably need another six to nine months before I can see if we're ramping faster than historical.
You have talked about that M800 being able to contribute another 300 basis-
Yeah
... points to gross margin. And I just wanted to clarify, that's consolidated gross margin that we're talking about. What is a reasonable—I mean, it's, I'm sure it depends on the uptake and the adoption, but what is maybe a reasonable expectation, you know, to scale to that level of contribution?
Yeah. So, first, a little bit of background on how we get to the gross margin accretion. So the performance improvement of the chip will pass on to our customers and relatively at no cost increase. So it'll be priced fairly similar to the M700. The M800 is 25. We get 25% more die per wafer, so it's a significant cost advantage to us. Impinj will keep that cost advantage, and that's what's gonna drive the 300 basis points of gross margin accretion. In terms of how long the ramp could take, I think it, yeah, I think it's a—before it becomes the volume runner, if it follows historical norms, I think it's probably a two to three-year ramp before it crosses the 50% threshold.
Yeah.
Now, again, we're very excited about it, and we think we've made a lot of gains that our partners will appreciate in this. So there's an opportunity for the M800 to ramp even faster, but it's just too early to call right now.
Sure. Well, just around gross margin overall, obviously, sub-target at the moment.
Mm-hmm.
Volume has a lot to do with that, but, how do you see that migrating back just to previous levels? And then, you know, from there, you've put forth a longer-term target, you know, 55-57, I believe.
Mm-hmm.
Is that almost solely dependent on the M800 ramp, or are there other elements to get to that longer-term target as well?
So our today's target is 53% gross margin. I signaled last quarter that Q3 would be down by about 300 basis points, half due to the lower revenue scale, because our revenue has come down, and then half due to a softer product mix, particularly reader ICs. We came in right around there for Q3, and then I signaled Q4 gross margin was gonna increase. The revenue scale isn't that different, but what we are getting is a better product mix. We're starting to sell a little bit more reader ICs as, as we're seeing some progress there. You know, as I look, getting back to the 53%, all we need to do is get back to the revenue levels that we were at the first part of the year. There's nothing else required there.
It's just scale against the fixed cost in our COGS line. The M800 is what is gonna drive the gross margin from the 53% to, call it, the 56%, right in the midpoint of the range you provided, and that's just the ramp of the M800. Beyond that, there's opportunities for further gross margin accretion with services like authentication, where we can add another revenue stream to the business.
Could you just give us, you know, a quick comment on authentication, just in terms of what the uptake has been? That was, is a relatively new introduction.
Yep.
How much, you know, if any, it's contributing right now?
Yeah. So authentication is our service to cryptographically authenticate items as genuine. Think of it as three different components. It is the M775 endpoint IC, so it has an ISO-standardized crypto engine in it and a unique key. Think of it as an algorithm that is in our readers that enables a challenge response to that. And then the third component, which is a brand-new revenue stream for us, is the authentication service, which will be SaaS-like margins for Impinj. We're in the early days here. We have been shipping chips, think of it in the quantities of hundreds of millions of authentication chips into the market for use cases in specialty food, performance apparel, footwear, pharmacy, medical devices, and tax tracking.
Those use cases or pilots are underway right now, and we'll see how they progress and how we can ship into those. Very encouraging to see that type of broad interest for authentication at this point. I still think it's too early to model significant authentication revenue in 2023, but hopefully, as those projects and pilots progress, we'll start to see that in 2024.
We'll conclude here with this last question and maybe lightning round for you as well. Just a thought on competition, and if any change there, any shift there, you-
Yeah.
Major competitor, but anything, you know, we get asked, even though I don't cover it, you know, how we see, you know, China competition.
Yep
... you know, potentially emerging.
So, the landscape today is, it's a duopoly, us and NXP, in the endpoint IC market. We have probably together 90% share. We expect competition to come, but we haven't seen it yet, not in a meaningful way. What we do is hard. You know, all that functionality is built onto a chip that's smaller than a grain of sand. We leverage none of the standard libraries of our foundry partner, so, and we have a very significant patent portfolio. So the competition will come, but it's gonna take some time, I believe.
Very good. We'll conclude there. Thank you.