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

Good morning, everyone. Thanks so much for joining us. I'm Toshiya Hari. I cover the semiconductor space at Goldman Sachs. Very excited to have the team from Impinj with us this morning. We have Chris Diorio, Vice Chair -- CEO, Vice Chair and Co-Founder; and Cary Baker, CFO, with us this morning. I've got a bunch of questions, but it's a nice group here, so I would love to keep it interactive. So to the extent you've got questions, please have them ready. Chris, Cary, first of all, thank you for coming.

Thank you, Toshiya.

Thank you for having us.

Great to see you. I was hoping you could give us a little bit of an intro, Chris. We're aware of Impinj and what you guys do and what your differentiation is. I think there are many people, particularly on the webcast that perhaps are relatively new to the company. So perhaps such on the history of the company and where you play specifically in the RFID market.

Okay. I will do my best, and I'll try and do it concisely, but give you an update of who we are as a company and where we're headed. So think of the products that Impinj makes as being a wireless replacement for barcode. We make miniature radio chips, smaller than a grain of sand, that go on individual items, especially the market led by retail apparel, but today by general merchandise, automotive, supply chain and logistics. And inter radio chips that go on the item that wirelessly communicate that item's unique identity and do so at up to 1,000 items per second without line of sight and at up to a 10-meter or 30-foot free range and for a few pennies per item. Impinj as a company makes those miniature radio chips that go on the items. We sell them to partners, some of the likes of Avery Dennison or a Checkpoint, who attach that IC to a small antenna with a IC antenna either in the price label of an item or sometimes sewn into an item or otherwise attached to an item. We make reader chips that go into partner readers read those connected items. Those reader chips go into handheld readers, point-of-sale readers, loss prevention, exit gate readers and a whole host of other reading opportunities. We make fixed readers with our own algorithms and software that solve difficult problems, for example, loss prevention in stores. And going forward, we, as a company, have talked about other opportunities for us, for example, to put unique features in the, what we call the chips that go on the items or an endpoint IC, so unique features into those endpoint ICs that give differentiated value above and beyond just you identifying the item. Today, we have a feature called Protected Mode which allows you to turn that chip invisible at point of sale with an 8-character pin. And it's so doing for tech consumer privacy, but also facilitate checking for loss prevention. Going forward, we've talked also about the opportunity to do cryptographic authentication on one of those chips, so you can cryptographically authenticate the item is genuine. So think more broadly, if you have a semiconductor chip that goes on an item to replace a barcode, it offers enormous opportunities that are offered by the entire -- everything the semiconductor industry has done for so many years to take thousands of learnings and those abilities and apply it to each individual, uniquely identified item. Impinj has been at this thing for about 20 years. We do a particular type of RFID thing. RFID is about as broad as saying radio. It just means that it's -- I know it's a certain way of communicating over the air. We do a certain type of RFID, which the industry is branding as RAIN for radio identification. The ubiquity of the term, we say cloud and the term rolls -- just rolls off the tongue, so the idea here where the industry is to use the name RAIN for radio identification, but it really means uniquely providing the identity, location and going forward, the authenticity of every individual item that companies manufacture, transport and sell. And that is our vision. It provides that unique identity, location and authenticity for every item in our everyday world.

Awesome. Thank you for that intro, Chris. A question on how to think about growth going forward. Based on our own estimates, we have your business growing about 30% this year despite supply constraints and all the disruptive things that are going on in the world. And I think you'll be roughly double the size in terms of revenue vis-a-vis 2018. As you mentioned, Chris, you've been going at this for many, many years, but it seems like the rate of adoption of RFID is accelerating. And the pushback that I get as a relatively new analyst covering Impinj is you're new to the story. You've been talking about this for ages, and what's different this time? Can you sort of address that question?

So maybe I'll address the latter part of the question first, we're a little bit about what's different this time. And then I'll let Cary take the former part of the question about growth. So what's different this time? Well, actually, if you look back over the past decade, the unit volume CAGR for our industry has been running between 25% and 30%. Started at a small base, the original opportunity or the first opportunity out the gate was retail apparel because apparel items are perishable. They're seasonal. And the retailers need very good inventory visibility given the number of SKUs, different sizes, types and colors. So retail apparel was the first opportunity to go forward and drove industry adoption with the first vertical across the chasm. As we entered pandemic, companies that had deployed RAIN RFID and had visibility to their items could rapidly pivot to omnichannel fulfillment. And it became readily apparent that RAIN RFID was no longer an optional capability that would give a retailer visibility to their items. It was a necessity to drive the efficiency of their operations to drive their ability to manage their inventory and their ability to basically pivot and adapt to changes in the market environment. Subsequent and out of COVID, there's also labor shortages, great resignation. Companies are realizing that they need the -- that they need to drive efficiencies RAIN RFID by automating the process of identifying manufacturing, transport and sale of those items and giving visibility or traceability fundamentally altered the calculus for these end customers. Not only in retail apparel, but retail general merchandise, supply chain and logistics, automotive parts and automotive supply chain, aviation and in many other verticals. So I would say for us, at least, that coming out the other side of COVID-19, we have seen an acceleration in adoption in 2 dimensions. One, just overall retail apparel and the initial verticals adopting at least as rapidly as they were previously and an expansion in the number of verticals opportunities.

Cary, on the growth side?

Yes. Chris highlighted the unit CAGR growth that we've seen over the last 5 years. That has predominantly been retail apparel, probably 70% of the chips that the RAIN RFID market sends into the market each year, it goes into the retail apparel vertical. Retail apparel is probably 25% penetrated at this moment, yet most retailers have made the decision to move forward with retail as Chris cited the reasons why. So we think there's opportunity to continue that unit CAGR as retail becomes more penetrated -- retail apparel becomes more penetrated. We're looking at a couple of opportunities to layer on top of that in the next few years. First, within retail, the definition of retail is expanding. It's moving from apparel into home goods and this expansion is led by Walmart. In the last several months, Walmart has announced their plans to move from apparel into home goods, into electronics, into toys and into auto parts. So the definition of retail is expanding and we think that provides an opportunity to layer growth on top of the unit CAGR that we've seen in the past. And then Chris also mentioned supply chain and logistics. The big shippers of the world are putting the infrastructure into their distribution centers today to begin reading tags tomorrow. We think as early as 2023, we'll begin to see large tag volumes from the North American supply chain and logistics companies that we have relationships with. So there's opportunity to build upon the unit CAGR that we've seen over the last 5 to 10 years by that retail expansion and by also supply chain and logistics.

Great. Based on both of your comments, it sounds like what used to be retail apparel-centric market is now broadening. And you mentioned a couple of verticals we were seeing customer pull. Of all of those, are there 1 or 2 or 3 that really excite you? Or are they kind of all equal in magnitude from an opportunity perspective? Where are you seeing the most customer pull or customer traction?

Sure. Thanks. There are -- there's 2 that we're very excited about and are focused on, and there's a necessity focus in general because there are so many opportunities out there in the market that we, as a company, need to need to focus. So our focus areas are retail, from retail apparel to general merchandise and supply chain and logistics. Now saying those words does not mean we're completely abandoning automotive, manufacturing, aviation or the so many other markets. But we have a very large partner base. We sell to the market through partners. We have many hundreds of partners. We support our partners as they work to deliver into those other vertical opportunities where we focus our know-how, our innovation engine and our engagements directly with the end customers, that retail, general merchandise and supply chain and logistics space basis.

Got it. Got it. Based on the positives of RFID technology that you just discussed and some of the dynamics coming out of the pandemic, adopting RFID seems like a no-brainer in many cases. To the extent you get pushback from customers and customers say no, thank you, what is the pushback? Is it technology? Is it cost? Are there any competing technologies to RFID outside of traditional barcodes that we should be cognizant of?

So there I'm going to work backwards a little bit. I don't know of any other technology, including other types of RFID beyond RAIN RFID that allow reading 2,000 items per second at 10-meter range without line of sight and give you all the benefits of being able to scale additional capabilities with Moore's law. There may be something out there, but I don't know what it is. I mean we see what we're bringing to the market broadly being a significant differentiator in terms of supply chain visibility. And I'm using the term supply chain very, very broadly. And as you think about those, the opportunity there and the words that we've been using in terms of the transformation that businesses undergo, it would seem like obvious, why wouldn't everybody want to do this thing? Over the past 2 years, we've had a significant impediment which was the -- it was product supply, wafer supply. We sell billions and billions of endpoint ICs that go on to items, consume a reasonable number of wafers and we have been constrained by wafer supply. Over the past 5 quarters, we said that our demand has exceeded our supply by more than 50%, and we signaled on our last earnings call that we expected that same trend to happen in the third quarter. So we've been constrained by supply and that supply constraint has impeded the ability of end customers to adopt. We have essentially slowed down adoption because of lack of supply. In a broader sense, the bigger impediment -- maybe I shouldn't use the word impediment, the bigger challenge for end customer adoption and the things that end customers have to get over is they fundamentally need to rethink how they run their stores and run their operations. Previously, they kind of SKUs. I had a certain SKU and it was a red shirt, whatever of a certain size, and I may have had 29 of those red shirts and they're in some stores somewhere. We're now delivering visibility down to the individual items. And that visibility requires retooling the software backend, basically track individual items, retooling to operations to understand where those items are in the supply chain and the visibility, building out readers, either fixed readers in the supply chain and total readers in employees' hands in the stores or both and then even thinking beyond that, doing, for example, self-checkout. So it is not the RAIN RFID element that actually is the impediment to adoption or really the cost. We don't really have much in the way of cost discussions now on the cost of the endpoint IC as measured in pennies, the value, the ROI is very high. It is the need for businesses to transform their operations. The flip side of that coin is once they transform their operations and go and become reliant on RAIN RFID for running their stores or their operations or the supply chain, companies don't flip back. We have yet to see a company flip back that is truly deployed and say, no, we're not going to be doing RAIN RFID anymore. Once they transform how they run their operations, they go forward. And for those of you who fly, for example, Delta Airlines, you use the track my bags app -- the track my bags app is based on an endpoint IC, one of our ICs that's in the bag-tag. Think of the whole build-out deploying more than 10,000 airport -- repoints at airports and facilities worldwide. That's a transformation that Delta has gone through and gets incredible value reducing the incidence of lost bags by more than half, increasing flier satisfaction. Once companies take that step, they don't go back, but it takes time to take that step because they retool how they run their business.

I was actually [indiscernible] by that at the summer. So thank you.

Good. Okay.

We were in Italy, but our bags were in Norway.

Thanks for buying one of our tags. More than one, I hope you shipped a lot of bags.

That's great. I definitely want to come back to the supply side of the equation. But before we go there, the competitive mode, and I think you sort of touched on it a little bit in some of your remarks, but the question that I get is, if Impinj is severely constrained in ICs and hypothetically, NXPI, your competitor has a ton of supply because they're bigger and so on and so forth. How easy or how hard is it for Avery Dennison to say, "Okay, I'm sorry, Impinj, we're just going to use more of NXPI"? Or more broadly, what is your competitive mode is the question?

Yes, I understand the question. So there's 2 parts to my answer to that question. Number one is that our industry's supply has, we believe, been broadly constrained. And if our competitors had hugely more supply than we do, I find it unlikely that we would be able to say for 5 consecutive quarters that demand exceeded supply by more than 50%. So you know the supply constraints extend across the entire -- multiple industries. So what we're going to be looking for going forward is how quickly we can ramp supply relative to our current competition to ramp supply. And of course, we have design wins that we're going to be doing our best to fulfill into and grow there. So there's -- it's not just a straight-up competitive dynamic. So obviously we're incredibly focused on supply. And our industry broadly, at least we believe, has not been able to meet the demand. From a competitive dynamic, at least on the endpoint IC and our primary competitor, NXP competes with us only on the endpoint IC, not on the reader ICs, not on the readers, not on the software or anything else, just the endpoint IC, we provide -- all of us use an industry standard protocol which defines the base level of functionality, how bits go over the air and how you identify an item. We bring differentiated features to the market, for example, that Protective Mode I mentioned previously, that allow us to win end customer opportunities uniquely with our ICs based on those features. So I want to think about it another way, we aspire to take a certain portion of the market opportunity off the table with our differentiated features and then compete in the competitive portion of the market that's left with the likes of an NXP. If we can't provide enough supply and NXP can, then the end customers will, if they desperately need supply, either find a way to use just the base features just to get the base functionality either -- or generally temporarily until we can supply back, but that won't always be the case. But basically, they'll find a way if we can't find enough supply and somebody else has it because they need the supply in order to run their operations. And so that is the dynamic. And going forward, our focus is on getting more supply and continuing to drive those differentiated features of the market so that we can take some portion of the end customer opportunities off the table with the differentiation. And we are uniquely able to bring those differentiations to market because feature we put in our endpoint IC, we support in our reader IC, we support in our reader, we support in the software features. We work with the end customers directly, use those features to solve end customer problems. And for the most part, features that we bring to market are based on requests, not for a feature, but to solve a problem directly from an end customer. So that Protective Mode feature where you turn a tag invisible actually came from 2 opportunities. One was driven by retailers who wanted to protect consumer privacy. And one was actually driven by the government who came to us and said we'd love to tag our [indiscernible] systems and theater operational, basically the weapon systems. We need to make the tag invisible when the ordinance is out in the field, otherwise a bad guy could put a reader in the field and whenever the reader read anything, load an IED. So can you find a way to make the tag invisible, but by which we could still read it when the weapon system gets back to the depot. So it's the end customer coming to us with a problem where we develop a feature, roll that feature out in the market, end customers adopt that feature, and that's a core focus for us going forward. But to answer your question, if we can't deliver and somebody else can, general end customers, if they have to lead and want to keep running their operations will find a way to use the base functionality until we can deliver again.

All right. Okay. And I guess, again, a follow-up on the competitive dynamic. Again, it's a high-growth market. In 5 years, 10 years' time, it's going to be a much, much bigger market. If another player outside of NXPI comes in and says, hey, it's a great opportunity, we've got scale, we want to participate in this market. Again, what's sort of the competitive mode? Is it that you having the reader ICs in the systems and the total solution? Or is it something else?

So we're building our competitive modes around our entire platform, around our engagements with end customers to drive preference for our products, around our intellectual property, know-how overall and just -- and our leadership position in the market and to the extent that we can going forward, driving supply. So those are what our focus areas are in terms of the competitive mode. We're -- we basically pioneered this industry and we want to continue to be the pioneers, the leader to the end customers. Not just our direct partners, but our end customers directly come to us with a problem, ask us to solve it, or ask to deliver a differentiated solution through our partners to that end customer and win those opportunities.

And then coming back to the supply constraints, to level set the audience, what have you seen from a supply perspective the past 12, 18 months? I feel like it's getting a little better, a little less bad. What are you hearing from your foundry supplier? And at what point do you feel like you can supply to demand as opposed to your comment earlier, more than 50%...

I'm not going to be able to answer the latter part of your question because I don't have -- I don't -- predicting the future in terms of when we're going to have sufficient supply overall, it's very difficult at this point in time, especially given that steepness of the live shortfall we saw and if it's going to come down steep on the other side, I believe. And so predicting the time when there will be sufficient supply is very difficult. What we're seeing now is early signs of supply beginning to free up. It's freeing up because there's other industries that are over inventory and are taking less product. The nodes that we're in, the older generation nodes are in general the slower ones to get additional capacity. And so even as other companies might be reporting that they have sufficient supply, we do not have sufficient supply yet simply because we're in the older generation process nodes where supply hasn't freed up that much. We are seeing supply beginning to free up a little more quickly in the older 200-millimeter wafers rather than the newer 300-millimeter wafers, but that's kind of just a broad statement. So I think the best answer I can give is early signs supply is freeing up, we've signaled that over the next 2 quarters, we're going to be gradually increasing the number of ICs we deliver into the market. That's based on that supply visibility. We have expectations of more supply for it based on what we're seeing in the market. But we still are in constrained nodes. Anything you'd add there, Cary?

No, I think you covered it, Chris.

And the internal constraints that you had talked about several quarters ago, a couple of quarters ago, those have been solved and you're comfortable at this point it's primarily external foundry supply?

It is primarily external foundry supply. So when we get a wafer from a foundry, that wafer goes through a set of post-processing steps. That we don't do. We actually outsource, but we significantly purchased the equipment and cited at the outsourced company. That includes bumping – testing, bumping, thinning and dicing the ICs. And so what we did over the past 18 months is build that post-processing capacity sufficient such that when wafers free up, we will be ready to catch. So we overbuilt and we -- and to continue overbuilding that post-processing capacity so that the only constraint is the wafer supply constraints that have or said another way, when the logjam breaks, we want to have a nice straight shoot that we can take those wafers and drive them into market as fast as possible.

Got it. When you guys talk about this big gap between supply and demand and lead times, the natural question that we get from investors is risk of double ordering and what kind of visibility does Impinj have? And I guess what controls do you have internally to gauge what is true demand versus what's excess ordering or fake demand or what have you? Can you kind of talk about that?

Yes, I can take that one, Toshiya. So as Chris mentioned, for the last 5 quarters, we have cited demand in the period has exceeded our ability to supply by greater than 50%. And we think that was likely to occur again in Q3. There's a lot of rigor that goes into that number. We first start with demand that's requested in the period. So we ignore any unfulfilled demand for the prior period, just focused on the current period in which we're trying to supply. We then back out any suspected double orders and we back out any channel inventory refill. We have good visibility into both because of our relationships with the end customer, we know their demand. We also have good relationships with our partners, so we know their channel inventory levels. We back all of that out of the equation. And then we judge the number down even further to provide what we think is a sustainable run rate demand in a supply unconstrained environment.

Got it. Shifting gears a little bit on the financial side of things. Maybe a couple for you, Cary.

Okay.

Gross margins, it oftentimes can be a little bit bumpy on a quarter-to-quarter basis. Generally speaking, I think you guys have done a really good job in expanding margins. I know mix is a big dynamic. But as we think about the next couple of quarters, what are some of the puts and takes that we should be aware of? And any levers that you can pull to improve margins longer term?

Yes, of course. So today, I think of our gross margin in the 53% to 54% range. And if you go backwards in time 2 years ago, our corporate gross margin was 50%. The accretion of 300 to 400 basis points that we've been able to achieve has been through the introduction of our M700 chip. It's our first chips on a more advanced process node. And we're leveraging the cost efficiencies that we have gained with those chips. Now as I look forward, there is more innovation to come on the M700 family of chips. We will introduce chips that can solve complex use cases and therefore drive a price premium in the market. We will also continue to innovate on our volume running chips to provide more cost efficiencies there. Both of those factors should provide opportunity for us to deliver even more gross margin accretion than we already have in the first 2 chips.

And then in terms of the pricing environment for your chips as well as your input costs, I think you've managed really well, the increases in wafer costs over the past 12, 18 months. I guess what are your expectations going into next year and your ability to offset that?

So historically, we have seen an environment where we've had price declines in the low to mid-single digits that go into effect every Q1. That hasn't been the case in this inflationary environment for the last 18 -- 12 to 18 months. We have received price increases from our foundry partner. We've been able to margin those costs increases up and pass them on to our end customers. Not been easy conversations by any stretch, but we've been able to do that in this environment. As to what happens next year, it's hard to say we don't have the visibility just yet, but we hope in the next several weeks, we will have visibility. But you should assume that we will take the same approach and try to maintain the integrity of our margin model if we were to receive cost increases.

Makes sense. And then I guess in terms of your investments and how to think about EBITDA profitability, until the most recent quarter, I believe, your message was expect us to be EBITDA breakeven or better. The tone definitely or the message has changed.

Yes.

I guess what's changed? Why now? There's still in investment mode to all your statements, it's an amazing growth market. What's driving the shift in your messaging?

So we continue to invest in the business, as you noted. What has changed is we received an incremental amount of supply. As you noted earlier, it's less bad than it was. And that incremental supply allowed us to grow the top line a little bit more and highlight some of the leverage in this business model even while we're continuing to invest. We haven't put a long-term target in place. But the way to think about our investment profile going forward is assumed we will continue investing in our R&D line at the same rate as revenue growth. This innovation is what's helped us build the technological lead and what will help us expand our technological lead going forward. But as you move down the P&L, sales and marketing line, we don't need to invest at the same rate as revenue growth. We leverage a 2-step distribution model to go to market, both on our endpoint ICs and our systems. We have a very strong partner network. So there will be investment in sales and marketing, but nowhere near the level of investment as engineering. And then obviously, there will be a lot of leverage in the G&A line item. So that's how I think about our cost structure and then going back to the previous question, our gross margin going forward. So it gave us the confidence to say to change the financial posture of the company to invest in breakeven to invest and be profitable. And the next step will obviously be free cash flow generation.

Got it. I wanted to pause here. We have about 10 minutes. Question in the front, please.

Earlier, you talked about supply chain and logistics. UPS has been on their conference calls, very bullish in terms of cost savings [indiscernible]. Can you talk -- give us a little bit more details in terms of your thoughts on the ROI for them [indiscernible]?

It's Chris. Thanks for the question. I can't speak directly to any particular customer opportunity, but I can talk in general about the supply chain and logistics opportunity. And I'll try and do so in a couple of ways. First, as you think of large supply chain or logistics companies, unlike retail apparel or retail general merchandise, where the reading is done with -- primarily with a handheld reader in a store to get visibility, maybe they do a scan, some do a scan twice a day. It takes just a couple of minutes to scan a fairly good-sized store. Supply chain or logistics is moving items through a supply chain, conveyors on truck. So there is a lot of what we call fixed infrastructure, fixed readers. So in most -- or most, I would say, actually, to date, all of the supply chain and logistics opportunities, the first step in the opportunity is to deploy that fixed infrastructure. If you want to go with this way and say, do the plumbing first. Once those readers are in place, there's some quantities that the supply chain and logistics company can benefit from visibility into the tagged items. Then they will start tagging the items and rolling out the broad tagging opportunity. We have said and several touched that at least 1 or 2 of the large supply chain and logistics companies that their volumes that they will run through will start to accelerate and we see a fairly large tagging opportunity starting next [indiscernible]. In terms of the ROI, I think you'd have to look to them. The way we think about the ROI is in reduction in missed shipments generally, understanding the cost benefit of reducing those miss shipments, reduced labor costs, especially when they can't get the labor and overall efficiency that supply chain and logistics company can drive to improve their operations and improve their customer satisfaction. In the particular case of UPS, CEO Carol Tome has said that they believe they can eliminate 20 million manual scanning per day. So that's a gigantic number obviously. It points to speed up efficiencies, labor -- essentially getting more efficiencies out of the labor they've got. I'm not going to try and work through the actual dollars and cents of the ROI. But I don't think you would see a CEO making those kinds of statements if the ROI wasn't fairly well demonstrated and positive.

Any other questions?

I looked into your -- it seems like TSM is your only fab. Why do you only use TSM? Why not broaden it? And what nodes in particular are using spread over 200 and 300 millimeter?

So thank you for the question. So we have historically had a very close relationship with TSM. We've worked with them for the past 20 years. Our endpoint ICs really hearken back to kind of the old days of IC design. We use physical design and that we need to use this physical design, really deep in the guts of the IC design, working with the foundry to build the ICs because we're building this incredible radio. It's 300 by 400 microns on a side. It consumes in the range of a micro watt of power and it's a full radio on a chip. To me it's the smallest and lowest power complete radio that anyone has ever built. And it's just amazing. And so we work closely with the foundry in developing core technologies that allow us to build that radio in logic. I've said before that we would be remiss in not looking at alternatives and I would be a CEO in being remiss and not looking at alternatives, but our close relationship with TSMC and the support they've given us over the years and the deep technical abilities they have and we have combined have allowed us to market. So expect us to continue having a good relationship with TSMC going forward. Oh, on nodes, we have -- in our 200-millimeter, we use an optical shrink of 180 or it's 152 nanometers. In our 300-millimeter...

We haven't cited the node in 300-millimeter. We have said that 28-nanometer would have been a step too far for us to make when we moved away from 152.

But we took a fairly significant step down when we went to our 300-millimeter wafers.

And the reason for the ambiguity is competitive reasons.

Yes, competitive reasons.

I apologize, this feels like a stupid question before I ask it. But the retail apparel sector, inventory has been a car crash. And I'm wondering, from your perspective -- well, let me ask the question this way. For your customers, for customers that adopted your solution, how, if at all, has their experience been different? And what are some of the learnings and sort of use cases used sort of evangelizing [indiscernible], if it's been different, perhaps, yes.

Want to start there, Cary? Or me?

Go ahead, Chris.

So the experience of the retailers that have deployed RAIN RFID has been significantly different. As we entered the pandemic, they were able to pivot from store sales to omnichannel fulfillment and to fulfill out of stores. They were, in general, on the positive side of the ledger of the retailers that did well during the pandemic. We've actually seen accelerating adoption across the entire retail opportunity as a consequence of how those companies performed relative to their peers. Up to date, most of the retail opportunity, the vast majority has been in retail apparel. We're only seeing the push into retail general merchandise now. If you look at some of the companies that have made public statements about their deployment of RAIN RFID and you look at how they fared even recently, I think you can see that they have, in general, done fairly well. For example, if you look at Inditex, which has -- was cited RAIN RFID an early adopter, they made a lot of positive statements about RAIN RFID. They just had a fairly strong report come out from them in terms of how they're doing in their overall business. So difficult for me to cite numbers from the vantage point up here. But in general the flexibility to adapt their operations for retailers that have deployed RAIN RFID and have visibility has, in general, reduced overstocks, under-stocks and allowed them to more efficiently sell the product that they have. That's a broad statement because obviously counterexamples all across the board. But that's the experience that we've seen out of the market, which has been an impetus to drive further adoption across all retail apparel companies.

Chris, I guess, in the last 2 minutes that we have, you talked about very strong demand, broadening demand across more verticals. You talked about your competitive mode. You talked about supply improving and you talked about better OpEx leverage going forward. So I guess the question is what is not to like or what keeps you up at night as the CEO of the company? What should we be worried about macro aside because macro you can't control. But in terms of the stuff that you can control, what are you focused on? What are you worried about?

So what keeps me up at night is on the difficult side, it continues to be the supply equation. On the competitive side, it is also the supply equation. I want to get supply first and I want to be able to deliver them on the market, and I want to delight our end customers. On the positive side, what keeps me up at night is I love to invent things. I love to come up with new ideas. I love to hear end customer pain points and find a way to solve them and come up with clever solutions and then roll it through our platform and roll it out to an end customer, into the light and end customer. And so our opportunity to deliver those delightful experiences is one of the things that keeps me up at night because I'm focusing on it because I think of all the things we can do going forward, we have just scratched the surface of this opportunity. If you think about it this way, right now we are -- we as a company are focused on giving enterprises visibility to every item they manufacture, transport and sell, I used those words before. We are still early days in terms of the total market opportunity for that -- for all of those items, we're still way up 1% penetrated. And here we're focusing on enterprises as the user. Beyond that vision is [indiscernible]. I want to bring value directly to people, you and me from having connected items. I want to see an IC in every item that lives for the entire life of the item, not just manufacturing, transport and sale, but used by a person all the way through to recycling for sustainability. We have an IC that's in a garment, it lives for the life of the garment, it will greatly facilitate recycling at the end. It will allow you to bring your items to a drycleaner. And instead of them using a Sharpie and writing in the hem some code that represents you, there's an IC in it and it's immediately readable. I want to see the endpoint ICs on food items, reduce waste, but also to give me visibility in my house where are the items, what food items I have, what's expired, what's not expired, what I need to buy. I see incredible opportunities in automotive, understand items that are in my car and the age of parts and things like that. I'd say we've even had medical companies come to us and say is there a way you could put one of your endpoint ICs in a medical implant like a knee or a hip such that at some point later on in life, we can bring a reader right up against the skin because it's hard to reach with the body and actually get the identity of that medical implant so that if there's ever an issue, we know what it is and basically rather than maintaining the information in a database, maintain the information in the item. And we don't know how to do that. I'm going to say straight up, I don't know how to put a tag on an item that's visible through the skin because we're basically saltwater and the readability in saltwater is terrible. So we don't know how to do it. But we actually already have companies coming to us and saying can you help us improve people's lives, drive sustainability, reduce waste at the individual consumer or person level? And that's what keeps me up at night because we're going to deliver into that opportunity. It's in the future, but there is a future where every one of us is going to be able to get the benefit of the tagged items.

Amazing stuff. Thank you so much. Really appreciate the time.

Thank you, Toshiya.

Thank you.

Thank you.