Cognex Corporation (CGNX) Earnings Call Transcript & Summary

Good morning, everyone. Thank you for joining us today. My name is Jay Huang. I'm the Bernstein Industrial Technology Analyst. Together with me today is the Chairman and CEO of Cognex, Mr. Rob Willett. We will have a Q&A section. Before that, Rob will give us a short presentation about CapEx and the recent development. Rob, thank you for joining us today.

All right. We'll put our slides up. I don't know, Jay, can you see our slides. Here we go?

Yes, I can.

Perfect. One second. Great. So anyway, hello, everyone, and it's a pleasure to be here. Sorry, can't see you in person. But thank you, Jay, and the Bernstein team for inviting us today, and thank you for joining. Okay. Little housekeeping. Keep in mind that any forward-looking statements we make are based upon information that we believe to be true as of today, and things can change, and you should refer to our SEC filings for risk factors around anything that we say. Let's start with a quick overview of Cognex. So Cognex is the leader in the market for industrial machine vision and exciting technology that gives automation equipment, the ability to see. We were founded by 3 MIT folks and have been in business for 40 years. We've sold more vision systems, and we believe have a better-recognized brand than any other company in our industry. Everything we do is machine vision related and we believe we're still in the early stages of the adoption of machine vision technology in automation. We're primarily a software company with high-growth operating model business. And in 2020, our revenue grew by 12% to $811 million. That's good, but lower than our 20% long-term target in what was a difficult year for us and many companies. Our gross margin was 75%, and we report software [ light ] gross margins. And we're highly profitable at the operating margin level with 27% last year, excluding restructuring and other charges and tax adjustments. Let's talk about machine vision technology for a moment. And in some ways, it's analogous to human vision. Your eye is like a camera, and it captures data, and it sends it to your brain, which like machine vision algorithms, they make sense of the data captured by the camera. And while that happens easily, and we don't think about it in humans, it can be very challenging in automation and manufacturing, getting the answer with high accuracy, speed, repeatability, reliability in what can be a very noisy and changing manufacturing environment, requires a lot of technology, optics knowledge, software capability and application experience. So what do we do with that technology? Well, we apply vision technology to difficult applications performed on high-speed production lines in manufacturing and logistics. And it's something we do better than anybody else. Our products perform critical applications that could be summarized in 4 broad categories with the acronym GIGI. So the first is, we do guidance. Cognex's machine vision may guide a robot putting a printed circuit board into a device or a windshield onto a car. We identify. So we read barcodes and alphanumeric characters to identify and authenticate items at multiple stages of production and at the point of sale. We gauge, which really means we measure. Cognex vision performs high-speed, in line, precise noncontact measurement of critical features such as -- parts, such as this automotive brake pad. We may be looking at how wide and deep is the groove in that brake pad in this picture here. And the fourth thing we do is, we inspect. We visually inspect items like this syringe package to ensure that there are no surface defects or irregularities before the item leaves the factory. Machine vision is a great market. It's growing quickly. It's difficult to do, and there is a lot of advancement currently going on. Millions of people go to work every day, where they use their eyes to perform basic tasks that can be done better and more cost effectively with vision technology. COVID 19 has been accelerating this trend. The potential to boost productivity by freeing workers from manual, dangerous repetitive task has become ever more important. And in turn, vision technology frees them up to use their brain to do something more creative or valuable for their organization. Cognex serves a $4.2 billion market. Let's look at this. It's -- we define this in a very narrow conservative view of, if we could sell everything that our applications apply to and that we would like to be selling into, we would be a $4.2 billion business. It excludes potential opportunities that may develop over time or those more broadly addressable ones. So it's a pretty narrow definition of the business that -- and the market that we serve. We calculate our own market share estimate because there's no reliable third-party data available and we introduced this estimate about 18 months ago. So it's a little bit out of date at this point. But based on 2020 revenue, we have about a 20% overall share in the market. On the left of the diagram that you're looking at are the core market, where our business began, 2D Vision. And you can see that's about a $1.4 billion market, where we have about a 30% share. And over time, our business has moved more to the right as we've entered new adjacencies and new applications for our technologies and ones where we see high potential growth. Let's talk about our 10-year revenue. So we're a growth company, and we invest a lot in technology. And we have a stretched goal of growing our business at 20% a year over the long term, and almost all of our growth is organic. We generally haven't nor do we intend to buy large bolt-on pieces of revenue. And up until 2018, we were seeing compound annual growth rates over a 10-year period approaching and sometimes in excess of 20%. But our history demonstrates that revenue growth isn't linear. There are distortions over time, and that results in some breakout years of huge growth. Years like 2014 and 2017. And so over 9 consecutive years of growth after, I should say, 9 consecutive years of growth, our revenue declined for the first time in 2019 due to weaker manufacturing confidence, weakness in automotive and a slower electronics capital spending, a lot of uncertainty and distractions around trade tensions. So coming into 2020, we expected a strong growth year for Cognex. However, after the disruption caused by the global COVID outbreak, we quickly recalibrated for more modest growth while keeping our product road maps and plans intact. From a business perspective, the second half of 2020 was much more positive than we had initially expected. And 2021 has started off very well. We reported strong revenue growth and operating margin expansion in Q1 of this year on both a year-over-year and a sequential basis. For Q2, we expect to report the fourth quarter in a row with revenue growth in excess of 30%. And the consensus of analyst estimates for the full year is reflected in the 2021 number there. Let's talk about technology. So our industry leadership and our ability to deliver strong top line growth over the long-term comes from a long period of investment and ongoing investment in engineering and research and development. We have product road maps that go out 3 to 5 years, and we're very focused on executing against those. Each year, we spend about 15% of revenue on RD&E. It was 16% last year or more than $130 million. We are a high intellectual property business with more than 1,000 patents issued and pending. And the sophistication of our software and our application experience creates, we believe, significant barriers to entry into -- to competitors. I want to say a moment because it's really important about Cognex culture. Our success is underpinned by a very strong corporate culture. Our motto is work hard, play hard, move fast, and we strive to create an environment where people love what they do and do their best work. And we believe it's really important to have fun together as a team and not taking ourselves too seriously. And as the company grows, we want to maintain our entrepreneurial routes. And we encourage Cognoids, which is what we call ourselves, to move fast, take risks, to reduce waste, to be their best selves. A good example of that is what we call our C-Cubed Program or it stands for Cognoids cutting c***. And in business, I think we all know, particularly anybody who works in a large organization, that a lot of cut the c*** builds up. You're writing reports that no one's reading or you're performing tasks that serve no purpose. Some companies, and we do a bit -- address this with Kaizen processes, but we actually just pay Cognoids to identify wasted effort and then we stop doing it. We find it's highly motivational. It's part of our move fast culture. There's so much to tell you about Cognex culture, but that's just one example. Let's talk now about where the business is growing quickly today and where we see great opportunities. So first, I'll talk about 4 areas. But the first is logistics, which represents a $1 billion market opportunity for Cognex, and we believe that estimate has grown significantly since we first introduced it about 18 months ago. Logistics is a great market globally that's still in the early innings of adopting machine vision and advanced automation. Most of our business in this market today is barcode reading and e-commerce fulfillment, where they need to ship items to individual customers faster than ever, is driving massive changes in technology and innovation in automation. Our logistics revenue grew by 40% last year, largely due to growth in e-commerce, and it was the fastest-growing end market for us in the first quarter and became our largest end market for the first time in the first quarter, this year. We're benefiting from major e-commerce and omnichannel retailers investing in automation to enable higher throughput and cost reductions in their business as they look to serve customers more and more online. Other sectors of logistics, such as brick-and-mortar retail and airport baggage handling, who are very struggled last year through the pandemic, but they're starting to pick up now after struggling for a while. Second area I want to talk about is deep learning. Cognex is the leader in applying deep learning technology to industrial machine vision, which -- and we believe it's the most exciting technological innovation to be introduced to our industry in the past 20 years. Deep learning is expanding the served market by enabling us to address applications where traditional rule-based vision cannot be applied or isn't practical to apply. Our most important introduction from a product point of view in the last year, one of our most successful products ever launched was the In-Sight D900 smart camera. The D900 leverages are widely used In-Sight platform. It's the most widely used smart camera program platform in the world to enable more customers to apply our deep learning tools to problems. And they're applying them to problems such as the detection of scratches and chip surfaces that were previously too difficult to solve using traditional rule-based vision. Revenue from applications utilizing our deep learning technology more than doubled year-on-year in 2020. And as we look at the opportunities ahead, we believe we're just scratching the surface, excuse the pun, of what we can accomplish in this space. Next area I'll talk about is 3D vision. There's a high potential in a technically challenging market that's growing quickly. We recently launched a breakthrough smart camera platform that's highly competitive with PC-based laser displacement systems that are more common in this market today. Our new 3D-L4000 product combines true 3D vision tools in an easy-to-use interface and breakthrough optics with powerful image acquisition that are programmed, again, using our best-selling In-Sight platform. And the fourth area I'll talk about relates really more to Industry 4.0. And it's Cognex products, and we have a very, very large installed base of products. They generate more digital data from the factory floor than any other company I know. And our new Edge Intelligence platform that we launched recently can extract value from the billions of digital images generated daily by Cognex products. And this can help customers understand the performance of their devices deployed across facilities, quickly identify issues in their manufacturing process and take corrective action. Cognex Edge Intelligence is primarily a software data-driven business being used by our current customers and current markets. And it easily integrates with Industry 4.0 solutions and larger automation companies, providing access to valuable factory floor data for predictive analytics. In summary, Cognex is an industry leader in the exciting field of industrial vision. We have an experienced management team that understands the industry, a great brand, built from years of technological leadership and a strong partnership with our customers and a great business model that generates high-quality revenue growth with tremendous pull-through to the bottom line on incremental revenue. And with that, I'll say thank you and join Jay by the virtual fireside.

Thank you, Rob. Thank you for the presentation. Just a reminder to others participating in this meeting, if you want to ask a question, you can use the live Q&A function. That's at the right-hand side of your viewing screen. We'll see those new questions and also if you vote for a question that you care, we'll also see that. With that, Rob, maybe we start with some of the priority areas you described in the presentation. Let's start with logistics. Based on the reporting, it was over 20% of Cognex revenue in 2020. I remember, it was just about 8% to 9% in 2017. So it keeps growing. What percent do you think the segment will contribute in the long term?

Yes, it's a difficult question to answer with the number, Jay, but we see just huge growth potential in that market where we're addressing it today, which is primarily barcode reading. But particularly in future, where I think we're going to see more and more vision being applied to logistics space. Yes, I mean, to give you some idea, e-commerce fulfillment and online fulfillment of customer retail in America is still low, right? It's certainly less -- I think, less than 15% overall. And as retail companies build out their back ends, as they grow, we can see a lot of growth drivers having a long runway. For Cognex overall, our business is still primarily U.S.-focused, but we see lots of opportunity and are executing well and taking share in Europe, where there are a couple of more entrenched competitors there and really serving the Asia market, which is still highly manual in terms of how logistics is being fulfilled, and we see a lot of opportunity there, too. So I -- over the long run, we set a stretch goal for ourselves of growing that business at 50% a year. We like stretch goals at Cognex. We achieved 40% last year. We achieved significantly more than that in the first quarter of this year. So we see a lot and a long run runway of growth in that industry, but I can't give you a specific endpoint.

Understood. And you mentioned there's a 50% growth, that's a target for the segment. In one of the slides you showed, I think, you're showing the industry growth to be 15%. Do you think that number was too low because that was 18 months ago? Or do you really think that's the difference in Cognex's growth versus the industry growth?

Yes. You're right. The number we gave you was, I know, 18 months ago. And I think the pandemic accelerated some spend, particularly among a narrow group of real industry leaders in e-commerce who have big, big ability to invest and really stepped up their investments. So I think when we come, it will be interesting to see when we revise that number and share it with you whether it's significantly more than 15%. There's no doubt that Cognex is growing much faster than our main competitors in that space. I think everyone would acknowledge that because we are the technology leader and we're seeing our technology being recognized and applied to on a quite challenging problems and where vision has a very good payback for customers. So I like to give the example sometimes that a lot of what we're doing in e-commerce is reading barcodes. Take a large distribution center. It might ship a million packages a day to customers. And if we can read a barcode 1% better through that process than our competitors, it's 10,000 packages that don't have to be reprocessed or relabeled or end up going late to customers. So the payback is phenomenally good. And as we see other companies start to scale up and they see the value of our technology, too. So I -- yes, I have -- and I think the main competitors or the incumbents in that space are companies that have been in the space a long time, and they're not really so much technical innovators. They are more a company that had a good sales channel and okay, kind of, probably, in some cases, good enough barcode readers for an older world, which is more about conveyor belts and customers taking packages and shipping them to retail establishments where they sat in inventory and store shelves for us to come and pick them up, right? Those days are dwindling. I don't think more and more of us, it's going to be -- we're going to be buying from a sophisticated supply chain that's reverse-engineered with a lot of automation to deliver the package to us. But another reason I would be very optimistic about the long-term growth in that space is this is an industry that's still adding a lot of heads in manual labor, right, which is extremely costly, not the greatest work environments for people either, not that safe, et cetera. So I think just the will to automate, particularly in more developed markets, is great. And we haven't even seen the labor content diminish yet. It's still -- people are still trying to throw human labor at the problem where automation is already better.

Rob, help us understand what are some of the other vision functions in addition to barcode reading that can be used more and more in logistics and may help to address this labor issue?

Yes, sure. I think, well, I think we see a lot of vision tasks now being performed on packages going through logistics system. So examples might be looking for boxes that are damaged, right? Is it the right size of boxes? So dimensioning, high speed course dimensioning of products is key because there's a lot of wrong sizes of boxes going through and being delivered to the wrong places or not the right amount of packages going into containers or trucks, so they're not optimally filled. So that's kind of an interesting application. Then we also see problems where there are labels and things that need to be read on boxes, so hazardous labels. For instance, a big problem for a lot of e-commerce fulfillment companies is shipping dangerous goods or not shipping them through the right channels, can cause big fines from governments, et cetera. So that would be another one. I think as we look further out, some of the really big high-potential applications, which really are, I still believe, some way off from being served in a big way is replacing human picking activities with robots, right? And in logistics, those challenges, that's what we refer to as structured robot picking, which is something we do very well today in all sorts of markets. And then there's more random bin picking. If you take a big e-commerce company and, I think, Amazon, for instance, that they have 50,000 SKUs a week to their product line, right? And if you think of the complexity, just from vision point of view, looking like a barbie-doll in a clear package, see-through package, the complexity of seeing that and picking it up reliably with a -- with robot-guided vision is very difficult, right? And then think there are thousands and thousands of new SKUs coming into the distribution center. It's going to happen, in my opinion. And probably when it does, it's going to be great for Cognex, but I think some of those applications are a little bit further out.

How do you see the customer base within the logistics segment evolving? Last year, I think the disclosure says there is a customer that's 14% of revenue for the whole company. And that's probably the majority of the logistics segment. Is that change -- is that situation changing in 2021? And maybe in the next 2, 3 years, where do you see new customers coming?

Yes. We don't give specific customer guidance or revenue guidance, but -- so I'll talk more in generalities. I think you're right. I think the pandemic exacerbated a few things. Like, as I mentioned, I think some of the big, powerful, very technically-savvy companies such as the one we might be referring to there, they invested a lot, and they had great engineering capacity and strong balance sheet. And they saw it as an advantage to -- possibility to take share and extend their competitive lead against other companies. And then we saw other companies perhaps, and I referred to a large new logistics customer of Cognex in the first quarter conference call. But that would be an example of a company that perhaps was well-established in retail, but is now looking to be much more capable in terms of the back-end e-customer fulfillment of orders, right? And that's a company that's been on that road for a while. They have been strengthening their engineering team and developing it and they recognize our technology, and we're continuing to grow with those. And we're going to see a lot more customers. I believe, there are already some of the multimillion-dollar customers of Cognex who are on that same road, so we're going to see that happen. So I think we're going to see more traditional brick-and-mortar retailers who have aspirations to be more e-commerce fulfillment kind of companies coming up. We're also seeing OEM-type companies that have developed a capability in the space to do customer fulfillment technology for other companies, right? And so there are more OEMs. And so they're then selling their technology to other players in the fields, like I'm not saying, whether they're a customer or not of Cognex, but if one reads about the British company, Ocado, they're selling fulfillment equipment to Kroger and other retailers, right? And they have a very sophisticated technology. It's a type of technology that benefits from the capabilities that we have. So I see that kind of moving us from e-commerce more into retail -- broader retail. And then we have general retail companies as well, smaller operations that are even subcontracting or trying to develop their own capabilities, and they're starting to come back in a stronger way post-pandemic, and they realized they have to have this technology to compete. Then we have just some other vectors that are going to drive our growth in logistics. And one is there are other markets such as parcel and post market where we haven't necessarily had the technology or capability to compete, and we're becoming more and more competitive. So the big parcel delivery companies are more and more interested to work with us and starting to learn about the capability of our technology. And there's massive investment they're planning to roll out in automation over the next few years. And then a smaller market for us, but one that kind of died last year and is coming back to life is airport baggage handling, right? So that's a market where we're starting to see a lot more investment in vision to help us all get our luggage when we get back on those planes to start traveling again. Hopefully, Cognex vision will hopefully get your bag in the right place on the right time and tell you where it is.

Got it. And let's move on to customer electronics that's another very important end market for Cognex. In the past, we think a lot of the vision adoption is associated with the increased complexity of smartphones. What do you think will happen in the future? There is the feeling that for some of the leading players in the market, the innovation is slowing down. Does that mean the vision demand growth will also slow down in that sector?

Yes. I think -- so I think the history of automation and kind of innovation in electronics is very long, and I think it's going to go on for a long time. I think -- and -- but is it going to be smartphones or are there going to be other devices where most of the innovation is going to come from? That's an interesting and longer-term question. I think for Cognex, we see growth in components, high-performance components. A number of years ago, we saw OLED screens being kind of key driver of growth in our business, cameras, sensors, 3D sensors, et cetera, starting to come into phones and into other devices. So I think that's certainly one area. I think we also see growth where companies want to get labor, take labor costs out of producing electronics. So that's certainly helping to drive growth. And then I feel very confident that there's a lot of new interesting technology coming to market over the next few years that I think can be very promising just as way back flat screens were big, OLED screens were big. be these other things help to drive growth. So I think probably as the smartphone market matures and we all see and hope smartphone production units probably isn't over the long term, likely to grow very quickly, what is going to drive it is taking out labor costs, right, applying new technologies into that space, moving production from current locations to other locations. And then I think more broadly, as we've all seen, smartphones have spun off other things, whether it's the one you're wearing there, Jay, or whether it's a lot of the other devices that we see for which smartphones are hub, so that's helping to drive growth also certainly. Then I think what's more is Cognex basically has been selling very much a lot of our electronics revenue into final assembly and test of products where there are literally millions of people involved in manufacturing. There are also adjacent spaces where we're relatively underpenetrated and where particularly our deep learning technology can become very useful, right? So we see other areas of electronics where we see -- where we're relatively underpenetrated and can move more into. So I think electronics is kind of always an area for massive investment, massive development. And we are working with all the lead players in the industry, where we're allowed to. Because there's a few Chinese suppliers, we're not allowed to work within the industry, but pretty much everybody else we work with, very closely. And I'm optimistic about their long-term product development plans and how vision can help benefit and help them reach their goals.

That's helpful. Speaking of 3D Vision, we've noticed Cognex introducing some very interesting 3D Vision product. And actually, we hear some good feedback from market as well. When you look at the 3D Vision adoption, where is it happening in terms of the end industry? And what kind of application really requires 3D and 2D is probably not enough?

Yes. Well, Jay, I know I remember when I first met you. You're a PhD physicist from Stanford, right? So you probably appreciate the laser despeckle optics capabilities that we patented and launched. I think that's kind of breakthrough technology in terms of acquiring really high-quality images to perform -- to run our world-leading machine vision tools on. So yes, so you're right. We've been -- we see 3D as a great market. One we were late in getting into. Our main Japanese competitor has a very strong position in that market. So we see it. To answer your question, right, where is it adopted? It's very broadly adopted, very similarly, in factory automation to our overall revenue makeup. So a lot in electronics, there's a lot of demand. Again, that can be countercyclical and volatile depending on product road maps and things that are going on. So -- but that's the big market for the industry, I think, and for Cognex. Automotive is also a very major market. But we also see it in all kinds of applications, packaging, machinery. For instance, there would be a market, a customer I was talking with recently, to inspect bottle cap closures, things like that. We can see a lot of it. Then that's kind of the laser line displacement approach to 3D. And then there's a whole different approach to 3D, where we've really focused on investing a lot with acquisitions and elsewhere over the last 5 years or so, and that's area scan. So that relies on projecting a patent basically of light. It can be high-end products, more than 1 million points of light that are projected and moved on our product. And that's very applicable, particularly to some of our growth markets and logistics, where it allows us to do -- we launched the product 3D-A1000 a little over a year ago. Very good for high-speed dimensioning of moving boxes online, right? And it's not millimeters accurate, but it can certainly tell the difference between general box sizes and allow customers to make sure that boxes are full of the right things that there -- for instance, if you think about 6 bottles in a case, it can look down and say, yes, all 6 bottles are present or not as that's going by very fast. Is it the right size of box to double check with your order? You're expecting a shoebox, you're not getting a large box, right? So these are kind of applications where our technology there is selling well at the moment.

It's interesting, you mentioned this laser technology for the laser profiler for 3D measurement Cognex introduced, that's a great product. Is it the Cognex intention to invest even more in the laser technology, introducing more products there? Or do you still focus primarily in the camera technology, the area scan technology?

So generally, we don't discuss kind of our product road maps. It's kind of we're spending $130 million a year on R&D, and it's current proprietary, and I don't -- none of our competitors discuss it either. But I would say this, we did -- we launched a very innovative product platform, the L4000 with -- it's very easy to use in our In-Sight environment. It's a smart camera. It doesn't require all other kind of boxes or connections for it to be programmed or operate, and it has like world-leading optics in terms of high-quality image acquisition. We've really launched a narrow range, so I think you can expect us to do more with that platform over time. And your question seems to get at 3D versus 2D. I'm still of the opinion that 3D requires a lot of data to process and a lot of cost to acquire the image. So I don't think 2D is going away, and I would certainly expect us to go on investing. 2D machine vision has a long way to run, particularly in the area of deep learning, I think, where -- to manage data quickly and effectively to make it applicable, means that 2D is likely to be a place where we'll see most penetration of that technology.

Got it. And speaking of deep learning, that's another very exciting technology area Cognex invested into. Why did Cognex need 2 different acquisitions to build up the deep learning technology? How do ViDi and Sualab contribute differently to Cognex?

Yes. Okay. So yes, thanks. It's a great question. So I think I'm going to give more of a historic perspective, if you like. I think the folks who founded the company 40 years ago, they saw kind of deep learning come and go. And I think they had a lot of old [ stones, ] right? I think people thought deep learning is going to change the world, showing examples of good and bad products, and computers learning to do that rather than being programmed specifically to do something. I think we had a lot of questions whether that technology was really ever going to be ready for market and then a lot changed, right? So what we saw, we saw a company in the market, ViDi. It's name was in Switzerland, phenomenal engineering team, just excellent. And we got that product, and we tested it, and we saw, right, really had amazing functionality, way beyond our expectations. And so we love the team, and we acquired the product. And then we just had a lot of success with that. It really hit a sweet spot in the market. And we saw it was succeeding, but we didn't have the number of engineers we really needed to fulfill on that. And then also -- and I think it gets to your question is ViDi had a real core competence in imaging small volumes of images, right? So you can program our ViDi technology with as few as 10 images, right? And to be able to run on lower cost hardware, right? So very suited to embedded machine vision, right? And that technology, we took and we've launched it in the In-Sight platform with the 3D -- I'm sorry, with the D900 In-Sight camera. So this is something I think probably 5 years ago, no one would have envisioned we would have been able to do, which is to take a lot of very powerful kind of the vision technology and put it into a small smart camera. And maybe for those of you, kind of, less versed in machine vision, I think, the chips that perhaps your kids have on their gaming systems from companies like NVIDIA, those things are big and they're hot and they need a fan to cool them, right? They're very, very processor-intensive, right? And we like the technology, but it just isn't going to work in a factory, too expensive or -- so what has been amazing about ViDi and it aligned very much with what's great about Cognex is, they had some really fabulous vision tools that ran very well on embedded platforms. Now so that helped address a particular segment of the market, but we saw other segments of the market that did require more high end, more processor-intensive vision tools, and that is particularly applicable to the electronics market and the market in Asia, right, which is more generally built around PC-based automation infrastructure, right, not embedded. And as we were working in that market, we came across the company called Sualab in Korea. And again, we were really impressed with what we saw that engineering team doing, and we saw a great opportunity to apply it to our electronics customers. So we've been focusing. And as a result, we were able to focus the Swiss ViDi team more on embedded applications and the Sualab team more on PC-based and high processing power applications. And those applications really are very powerful for replacing human inspection type activities, and that's where that team has been focusing. So it's been a very nice. ViDi got it going. We were so excited by what we saw there in Sualab. We were able to bolt-on that. And now we have a great unified team of Cognoids working in both places in the market.

Got it. And so AI, our deep learning is, for sure, a very exciting technology that brings up new functions to the machine vision products. On the other hand, I think a lot of investors are also concerned that with AI, there are new entrants into the space. A lot of start-up companies are doing AI. Amazon is doing AI. And we were actually seeing Amazon introducing some of the vision services. I think it's called Lookout for Vision. You can upload the image to Amazon and they can analyze it using AI algorithm. Does that make -- does that introduce more competition to you? And does that become a risk from a competitive side?