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

Good morning, everyone. Thanks so much for coming. I cover the semiconductor space at Goldman. I am Toshiya Hari. Very pleased, very honored to have Cary Baker, CFO from Impinj with us this morning. I have a list of questions, but we'll try to keep it interactive. And thank you for coming.

Thanks for having us, Toshiya.

So in terms of the near-term, on your recent earnings call, you guided September quarter revenue to decline 25% sequentially, you talked about an inventory correction at your end customers, and I believe your inlay partners. You also implied that the correction may or could persist through the December quarter. I guess, what have you seen over the past couple of months? Any significant deltas or changes relative to what you had guided us to?

Yes. We guided the third quarter to a midpoint of $64.5 million. As you noted, it's down about 25% with most of the downturn in the endpoint ICs. The downturn is related to 2 areas: first, the channel inventory and the burn down that needs to happen there, and then just continued destocking amongst our end retailer customers. We're targeting to complete the channel inventory burn down by the fourth quarter. So we're taking as much as we can in the third quarter then catching the tail in the fourth quarter. At this point, it's really too early to say when retail might rebound. We're seeing mixed reports out there, but still expect that to be a challenge in the fourth quarter as well.

Got it. Got it. And I guess, the question that we often get from investors is, is the near-term weakness purely due to inventory/kind of cyclical factors? Or is competitive dynamics at play as well? I know you don't have perfect visibility into how things are playing out, but what are your thoughts on sort of the competitive landscape there?

I think it's mostly macro-related. Demand continues to be soft with retailer. We see that in the import data being down quite significantly since last September. On the competition front, I would say, competition remains intense, but I don't view that as meaningfully different over the -- than it has been over the last several years.

Okay. Got it. And then a question on your ability to gauge real demand relative to what could be inventory build. A year ago, at this conference, I think at the time, you guys were looking how demand is up here and supply is down here and the gap being above 50%. And I think a couple of quarters after supply caught up to demand, and now, we're going through a correction. So I guess, the question is, what kind of visibility do you have into customer inventory or partner inventory? And I guess, what sort of measures do you have in place to keep track of that?

Yes. The big difference between this year and last year is the retail demand is much softer than we anticipated. After a very supply-constrained 2022, our channel entered this year really on fumes. Our goal is to run the channel at about 6 to 8 weeks of inventory. So in the first quarter of this year, we began rebuilding that inventory and we rebuilt and continue that into the second quarter. Now, what changed in the second quarter was our view or our outlook on demand for the third quarter. We had anticipated a recovery in retail that just hasn't materialized. And as a result, we overbuilt the channel by about 4 weeks. So instead of ending Q2 at our targeted 6 to 8, we ended it at 10 to 12. This was unusual for us in that it was widespread overbuild. It wasn't isolated to a handful of inlay partners. It was pretty common across all of our inlay partners, even some of our larger, more sophisticated systems partners. So we're going to try to burn that down, as I said, as quick as possible. And we think we can take 2 to 3 weeks out in the third quarter and then catch the balance in the fourth quarter. Because this was widespread, it was hard to see. That's not to say there isn't room for improvement, and we're certainly looking at that. We're looking at ways to get even closer to the end customers, either the service bureaus or the retailers to understand the fluctuations in their demand expectations and then bounce that up against our inlay partners to get us a better view.

Got it. Taking a step back, the RAIN RFID market has been a very robust market. It's grown at a CAGR of roughly [ 30% ] over the past decade, and I think you and your partners or your peers, shipped about 34 billion ICs last year. Remind us, what is the value prop of RAIN RFID? And I guess, from an end demand perspective, which verticals are you most excited about?

So years ago, the industry really started with the retail apparel vertical. And there, the base use case is inventory visibility at the retail store level. So the value proposition to retailers in that case is really moving from inventory visibility of mid-60% accurate without a RAIN deployment up to 90% or better accuracy. You think of the traditional way of counting inventory at a store, it's shutting down a store early one day and bringing in extra staff to count that item -- count all the items by hand overnight. And then the next day, it's less accurate. And the day after that, it's even further less accurate. We have retailers today that are inventory in their stores twice a day. And the value that they're gaining is the efficiency, of course, but it's also -- because they have the visibility, they're able to sell items, and we've seen increase in same-store sales. We also see retailers able to take inventory off the balance sheet because they're able to sell down to the last item that they have. So we see the play and the value proposition, both on the sales side, but also on the efficiency side. Today, retailers are asking us what else they can -- how else can they extract more value from the RAIN deployment. So we're seeing retailers move from the base use case of inventory visibility into self-checkout loss prevention. We're seeing interest in front store, back store automation and some -- and by some of the very visionary players, how to drive innovation in fitting rooms. That is the first use case, the first vertical to move forward. We're seeing that same base use case. Inventory visibility now translates into general merchandise. And big box players, such as Walmart, are leading the way as they're moving from an apparel deployment into multiple phases of general merchandise. Early days there as well. And then I would say, from a third vertical perspective, it's really in logistics providers. Logistics providers have historically been able to trade between throughput and accuracy, but never having both. With a RAIN deployment, we're seeing visionary players in the logistics place, such as UPS, drive significant efficiencies by no longer having to make that trade-off. But much like general merchandise, we're in the early stages of logistics right now.

Given everything that you just said, it seems like adopting RFID across these applications seems like a no-brainer. When you and your partners walk into a customer and they ultimately decide not to adopt an RAIN RFID, what is the typical pushback? What's the issue?

It doesn't happen very often. The typical friction point we see with the sales process is that it just takes time. You think about it, we're changing the way retailers or logistics providers count, track and trace the items that are most important to them. That's not a decision that they take lightly after -- it typically takes a lengthy pilot before a recommendation is brought to, not the procurement team, but typically to the C-suite or sometimes the Board. And once the decision is made to go, it typically -- the deployments typically go in a very pragmatic fashion. They start with a small portion of their business deploying, something that they can run in a controlled environment to make sure that the value demonstrated in the pilot can be replicated at scale. And then they'll make the decision to roll out more broadly. But again, that takes time. You can't go from 0 to billions of items tagged in a year. The benefit, however, for that patience is the customer has a solution that they know [indiscernible] can deliver the efficiency that it was promised. And we, as Impinj, are the happy customer that provides an opportunity to generate reoccurring revenue. Now, that's a very lengthy sales cycle, deployment cycle. The final piece that I would add is that, once they go down that process, no one turns back, and we've seen no real examples of any customer turning back from the RAIN deployment.

Got it. So given that and the value prop and how low adoption rates or penetration rates are, I mean, the opportunity set is massive, right? And I guess, as outsiders, it's always very difficult to forecast how fast or how slow the growth rate could be. Again, historically, the market has grown at a 30% clip.

Yes.

Is that the right sort of trajectory as we think about -- again, if you're going to have higher years, lower years inventory cycles. But on a through-cycle basis, is that sort of the right number to be working with?

Yes. So that 25% to 30% unit CAGR has largely been on the back of retail apparel. Today, retail apparel is roughly 30% penetrated to an $80 billion unit opportunity per year. We are now moving from apparel into retail general merchandise. Retail general merchandise is a much larger opportunity. I think 325 billion units per year. Yet today, as it stands, it's less than 1% penetrated. And we're also moving into logistics. Logistics is even bigger at 400 billion units per year. Again, it's less than 1% penetrated just because we're in the early days, much like general merchandise. So I think the benefit of retail, which will continue to grow because most retailers are moving forward, yet most retailers are not 100% deployed at this point, plus the benefit of layering on general merchandise and layering on logistics. I think that historical CAGR can continue to grow even as the base grows.

That's great. At the Investor Day, you talked about engaging lighthouse enterprises. For the benefit of the audience, what does that mean? How is the organization set up to go under these big customers?

So we think of lighthouse accounts as visionary players in their respective spaces that recognize Impinj can bring a unique solution to them and that we think have the potential to move a vertical or an industry or category forward. And that unique solution we can bring is our competitive advantage. And our competitive advantage very simply is we're the only player in the market that has both ends of the radio link. We deliver the endpoint ICs. We deliver the readers. And we deliver the algorithms that go on to the readers that enable some of those more complex use cases. So when we talk about a platform opportunity or a lighthouse opportunity, it's a customer that's leveraging our entire solution set to drive a solution to a complex problem that hasn't been solved before by mix-and-match solutions in the market. Typically, our engineers are on site working alongside the customer to bring that solution to life. And then when we deploy our platform for those visionary or those lighthouse accounts, we expect to have a larger share of the endpoint IC opportunity once the system is up and running.

Got it. And with this lighthouse account strategy, if you will, what sort of verticals are you after? You've mentioned a couple already, but anyone to really highlight?

Yes. So think of retail inventory visibility. As I said it before, the base use case, it's counting the items in the store using a handheld reader. That's designed for a mix-and-match solution. When we move to loss prevention, that's a more complex solution because now you're putting a reader at the point-of-sale terminal to read all the items and enable that self-checkout environment. You're also putting an algorithm in the reader telling the tag to turn off so that when you walk out the door, it doesn't trigger the alarm. And then you're also enabling the returns desk to have a similar reader with the algorithm there to send that 8-bit pin back to the chip telling the tag to turn back on, so the item can be reshelved and sold to the next customer. That entire experience relies on our platform to deliver that solution I just described. So it's much more complex than the base case. So we talk about 2 platform or 2 lighthouse accounts right now. One is our visionary European retailer that's doing the loss prevention deployment and now moving into self-checkout as well. And then we've talked about a visionary player as our second large North American logistics company that's deploying the RAIN infrastructure throughout their DCs and are beginning to run tags right now.

Got it. And again, the goal here is for that one lighthouse customer to adopt and others to follow suit.

That's a great point. We're a company that's a little over 400 people. We can't deploy a solution like that across a multi-brand company like the visionary European retailer. So what typically happens in those cases is we work with the end customer to design the solution. We design in our lab, in their lab, we work together collaboratively on that. And then once the solution is ready to be deployed, they typically pair us with one of their systems integrators or their value-add partners. And we train up that partner to deliver the solution first to that customer, but ultimately, to train up that partner to deliver the solution to other players in the space. So hopefully, we can -- it can drive multiple sales out of that effort.

The North American logistics customer, I mean, they've been very public, right, on their calls about the efficiency gains and the cost implications. Are you hearing from other companies in that market due to that transparency from the leader, if you will?

Yes. I would say, we see a lot of interest in our pipeline. When you have a player of that size talking about the benefits that they're going to achieve and the level of benefits that they're going to achieve, it has garnered attention from a lot of players in the space. And while they're the biggest and most vocal, they're not the only player that has adopted the RAIN in the logistics space. We've got examples in the U.S., and we've got examples globally on a much smaller scale, to be clear, but they're not the first one to understand the benefits of RAIN, but they're certainly the first one to take it to this extent and be as vocal as they are.

Got it. Shifting gears a little bit. Maybe talk a little bit about the Authenticity engine. Obviously, you've got -- at the corporate level, you've got systems, you've got endpoint ICs. How does this sort of play into that business model? Is there another revenue stream that you could potentially benefit from?

So let me give an overview of it. Authenticity is a way for us to validate the authenticity of an item and eventually take fraud out of the market. So think of this as an endpoint IC, our M775 that comes equipped with a crypto engine and a unique key in that. A reader, it could be R700 that is equipped with an algorithm that can enable a challenged response dialogue with that endpoint IC, and then a cloud service that will validate the authenticity of that item. So the way that, that will work is that we'll launch the endpoint ICs, think of these as higher ASPs than our typical endpoint ICs. The readers in the market can be equipped with the algorithm. And then the new service is the authentication service. So think of that as more SaaS-like than our typical business to enable and validate the items' authenticity. We launched the product maybe 9 months ago at this point. So we're very much in the piloting of use cases. We've seen a lot of interest early on, interest where we thought it would be in fashion, in performance apparel and footwear. We're seeing it in places we didn't expect it, specialty foods. We're seeing it in pharmaceuticals and medical devices. And we're seeing it in tax tracking, so think duty-free stores. So we've shipped primarily in the second quarter, but several hundred million chips into the market to go into these pilots that are just now underway. Still very early days. We're excited to see the results of these multiple pilots, and that will dictate kind of how the pull-through will work for really probably the beginning of next year. At this point, I'm not modeling authentication revenue in 2023, but that's an opportunity for us down the road for sure.

Great. I guess, on the competitive environment, we touched on this a little bit at the very top. But again, taking a step back, can you remind us what some of the key factors are that differentiate you, vis-a-vis, whether it be our systems, peers or your IC peers? What's different about the Impinj offering?

Yes. So I'll up-level a little bit and start with the RAIN offering for the industry in general. We provide a pretty unique solution set to the market in that our endpoint ICs are readable up to 30 feet away, without line of sight at speeds of up to 1,000 items per second. The endpoint ICs are energy harvesting, so they're battery free, and the additive cost to put an endpoint IC on an existing label was measured in pennies. At volume, it can be $0.035 to $0.04. When we go to market, we'll see other competing products in the space. So think vision, think NFC, RFID. But none of the competition provides the same set of unique features we can that I just outlined. So it really doesn't become a competition at that sense. They're used in different places. And we think, over time, that those solutions that we sometimes compete with today will eventually become complementary in terms of a suite of solutions offered to the customers. Now, as we go to market, Impinj, our competitive advantage has been the systems business and really how we can drive unique solutions to the market. And our competition doesn't have that. Our competition either plays solely with endpoint ICs or solely with systems. And it provides us a unique advantage to get to the end customer, even though we sell through the distribution for both -- of both of our product streams.

Got it. We often hear about outside of the other endpoint IC player, emerging competition in China. Is that something that you see in the marketplace or something that we should be concerned about monitoring?

We see competitors in China starting to come around, but think of it as not quality at scale at this point and think of it as really a China for China opportunity. So we haven't seen a ton of pressure on our sales as a result of it, but we're -- I constantly have my eye on the rearview mirror and see who might come in. This is a huge opportunity. We're selling billions of tags today, but we think it can grow to trillions of tags eventually. So I expect competition. But what we do isn't easy. All of those features I described just a second ago fit on a chip that's smaller than a grain of sand. Our M700 on a 300-millimeter wafer, we get 0.5 million die per wafer. And with our M800, which was just launched, it's over 600,000 die per wafer. That's incredibly challenging to do. And we've seen big players in the early days come into this market and get out of the market because of that challenge. So we feel we're in a pretty good space today. Eventually, there will be competition. I have no doubt, but I don't see it in the near-term, not at a meaningful level.

Right. I feel like Chris always says drop the pencil on a piece of paper and that's the size of your chip, right?

Yes. You can feel it. But if your eyes are over 40 years old, it's harder to see.

I'm with you on that one. Maybe going back to the financials. So at your Investor Day, I guess, at this point, several months ago.

Yes.

A couple of months ago. You shared a long-term revenue target of $500 million to $750 million, which was a little wider than what people had expected. And I don't think you had a specific time line. So I guess, how did you come up with that long-term model given what sort of transpired over the past couple of months? Is that still the relevant long-term model in your view?

So we certainly didn't expect the downturn in the back half of this year, but we did model variability into the year -- into those targets. And think of those targets as point in times, not necessarily a range. We covered historically that the unit CAGR for endpoint IC business has been 25% to 30%, and we've outlined why I think that can continue. So in terms of -- while we didn't provide a time line on that, use that as a starting point. Obviously, that can move up or down with a big systems deployment. But over time, we expect the systems business to become a smaller portion of the revenue than it is today. Today, it's 70% to 75%. But the endpoint ICs are the recurring revenue. So as we layer on that recurring revenue, it will shrink the mix of the systems business.

Got it. Got it. Maybe on pricing of endpoint ICs. So I guess, historically, you are on a fairly consistent cadence, right, where wafer costs would come down, your pricing would come down, tag pricing would come down and your customers or the end users would benefit from that. I think over the past 2 years or so, it's been more inflationary given higher input costs for you guys. How are you thinking about cost inflation going forward? How should we be thinking about pricing of your endpoint ICs going forward?

So that's exactly right. The pricing increase over the last couple of years was driven, first, by inflation environment. Our wafer cost went up, but we also had mix playing into that. So we sold -- particularly last year, we sold some higher ASP SKUs at a greater mix than normal, that drove the price point up. We'll go into our price negotiations with our foundry, think of it as our cost negotiations with our foundry partner in the next couple of months. I've heard reports all over the board, hoping for flat, but we'll wait and see what plays out. In the last couple of years, when we've received cost increases, our strategy has been maintain the margin model, so pass those costs on to our customers. And if we get another cost increase, expect us to do the same. Now, if I go back historically, our prices have gone down. Our cost has gone down and our prices have gone down as well. We typically had ASP reductions in the low to mid-single digits. But as I sit here today, I have some of the biggest companies in the world adopting both in retail and logistics. I have one of the biggest logistics players touting the savings that they're going to get from a RAIN deployment. They'll take out $0.5 billion of cost from their structure, take out 20 million manual scans per day. So the value that RAIN provides is much more transparent than it was 3 years ago or 4 years ago when we last saw the historic ASP declines. And instead of ROI being the focal point of the sales conversations, ROI is still important, but it's understood at this point. And now retailers are asking, okay, I'm going to deploy with inventory visibility like all the other retailers, how do I extract more value from my RAIN deployment? How do I move to loss prevention and self-checkout and get even more benefit from it? Which is to say, going into our typical annual price negotiations with our end customers, I feel stronger about the value that we can provide to not fall back into that cadence of every year we're dropping price. I think the price elasticity for retail logistics is pretty balanced at this point. Now, in my long-term targets, the $500 million and the $750 million, I have put in modest price declines to be conservative. So I'm talking really specifically about the near-term right now, where I believe our hand going into those conversations is stronger than it has been historically.

Okay. So just to clarify, we don't need to worry about pricing being flat to up and that causing a decline in demand, price elasticity kicking in.

I don't think so.

Okay.

I don't think so with the value that [indiscernible]

Right. It's pretty close, yes.

So let's talk UPS, who has said publicly that they'll take mis-shipments from 1 in 400 to 1 in 1,000. According to their 10-K, UPS shipped 6.2 billion packages in 2022. So you run the math, that's 10 million fewer mis-shipments. Pick a cost of what a mis-shipment is. And that plays pretty well to the pennies it takes to add an endpoint IC on everything.

Got it. Yes, that's clear. We have about 7 minutes left. I'm going to pause here and see if we have any questions from the audience. If you can wait for the mic. Yes.

Is there any reason to believe other supply chain and logistics company shouldn't see similar kinds of cost downs that your premier customer is seeing?

No. The logistics companies have historically always had to trade-off between throughput and accuracy, which is why you see the temporary labor spike in the fourth quarter to support the holiday volumes. With our solution, you no longer have to make that trade-off. That's why our second logistics provider can say, 20 million manual scans fewer per day and the mis-shipment rate improving. So I believe that other logistics providers can achieve similar type savings. Now, it depends. They've got to go in at it and approach it with the same visionary approach that the second provider has and not everybody has that same type of visionary approach, but I think there's savings to be had for anyone in logistics that wants to deploy.

Cary, you have a very strong relationship with TSMC as your foundry partner on the endpoint IC side. You had suffered from very severe shortages during the pandemic. Diversifying your foundry supplier base, is that a conversation that you guys have internally? Or I know it's expensive to have multiple routes, if you will, right? So how do you balance those two things?

So I would say, first, that our relationship with TSMC is very strong. And that relationship has actually got tighter during the supply-constrained environment. And to the point, we have regular conversations up and down the leadership change. However, it does make sense to consider multiple sources given the supply-constrained environment we just came through. It's not as easy as flipping a switch for Impinj. Everything we do is custom. We leverage none of the standard libraries, supporting our technology to a new foundry is an 18- to 24-month process. So it takes some time to do that. There's uniqueness in every foundry and you've got to work out different kinks and so forth. So which is to say that, yes, we'd be naive not to consider other options. There's no plans at this point, but we will do what's in the best interest of Impinj.

Got it. That makes sense. Gross margins, you have a target of 55% to 57%. You're around 50% at the moment on a quarterly basis, I believe. Obviously, you're going through this pretty significant correction. So I'm guessing, I'm hoping you're underearning, but what's the bridge from 50-ish today to 55%, 57%? What are some of the key drivers that we should be looking forward to?

So we guided Q3 down for 2 reasons. One, the lower revenue scale. And then 2, we have a softer reader IC mix. I view both of those as temporary. And when they recover, I think we're back into that 53% to 54% range that we had targeted. As I look to the longer-term targets, the difference is the introduction and ramp of the M800. I mentioned earlier that the M800 has 25% more die per wafer. It's also a more sensitive chip, which means that our inlay partners can attach a smaller antenna and get the same read range. So we are now in the qualification stage for the M800. Think about the M800 as being priced similarly to the M700. So our customers get the technological benefit of the more sensitive chip, and we get the cost benefit of 25% more die per wafer. So what's going to drive us to those longer-term targets is that M800. When we launched the M700, we were able to drive a 300 basis point increase in gross margin when it fully ramped. I expect a similar 300 basis gross margin lift when the M800 is ramped. Now, the M800 ramp will go slower. Historically, when we've launched new products, we see the take on new sockets, not replacement of existing sockets. Once the chip is qualified, you're talking about a penny part. It doesn't make a lot of sense to change immediately. Eventually, it change, but there's not a lot of incentive to do that. So we'll see the M800 go into new sockets. The M700 was a different ramp. We launched it in 2020. So think of depressed retail environment for COVID, and then we went straight into a supply-constrained environment. And there was a period where we can only get 300-millimeter product. We had to tell our customers, look, if you want endpoint ICs, you've got to migrate faster to the M700. So it's really a forced migration and that would allow us to ramp to get that 300 basis -- the first 300 basis point lift on a much accelerated scale. For the long-term targets, I'm assuming a normal scale, our normal ramp where we win new sockets, but we don't have a lot of replacement.

Got it. Okay. That's very clear. How to think about operating leverage going forward or OpEx leverage going forward? I think a couple of years ago, several years ago, you were still very focused on making the growth investments. And I know you still are. But I think you've talked about the pace of increase moderating going forward. So as hopefully, we come out of the cyclical correction in '24, how should we think about the OpEx leverage and the fall through?

So our primary focus for OpEx investment is in engineering. Previously, our goal was to grow engineering at the same rate as revenue growth. We found we couldn't do that. We couldn't -- we were growing faster than that. And we were still able to deliver the innovation that we needed to. So we'll continue investing in engineering, but think of it as slightly lower than the rate of revenue growth. Sales and marketing, we take our products to market through a channel. So there's leverage in that channel. Yes, I think of it as requiring nurturing investment to keep the existing channel healthy, trained and so forth. And then as we're moving into new verticals, if there's not channel overlap to bring on new channel partners. But there's a ton of leverage in the sales and marketing line, just given the -- given our sales approach. And then there'll be significant leverage in the G&A line item, of course. So overall, leverage across all 3 lines, but at different rates with our primary focus for investment in engineering.

Great. Maybe in the last 45 seconds, anything about the Impinj story or anything about the broader RAIN RFID market that you feel like we collectively underappreciate, underestimate through your conversations with [indiscernible]

Yes. I think some folks still think of us as more of a digital play where we have to go and resell every year. We're -- I think you should think of us as consumable silicon we sell and our endpoint ICs that go on perishable goods, and they have to be replaced every year, every quarter, depending on where we're selling. So the endpoint IC business is really a reoccurring revenue stream, and we're adding more and more customers in new verticals on top of that.

Great. Thank you so much. Really appreciate the time.

Yes, thank you. Appreciate it. Thank you.

Thank you.