Cognex Corporation (CGNX) Earnings Call Transcript & Summary

All right. Thanks, everybody, for making it. My name is Troy Jensen. I'm Managing Director at Cantor. I follow kind of industrial tech. I recently picked up Cognex. We're lucky enough to have Dennis Fehr here, the CFO of Cognex. So Dennis, why don't you just quick start with a background of yourself, and then we'll kind of open it up and kind of get into a presentation you delivered here.

Yes. No, happy to. Good morning, everyone. My name is Dennis Fehr, Chief Financial Officer of Cognex. I've been with the firm for about a year. Worked very long at Siemens, originally from Germany, came to the U.S. in 2018 when we spin out a company called Fluence Energy, scaled that, took it public and then had a shorter time at a private company in Boston time. So I'm here, I'll share a bit about Cognex, what we are doing before going into the details. Just a quick reminder, may do some forward-looking statements that you don't put undue reliance on such statements. Please actual results may differ, and please consult our latest 10-K and Q before investing. Who is Cognex? So we are a tech company based out of the Boston area, the town of Natick. While we ship, let's say, software embedded on devices, we really see ourselves as a software company as a true tech leader in the space, and it's also kind of the margins which we command with gross margins in the 70% range and then highly attractive operating margin and leverage. What we do is machine vision. So in other words, we sometimes say is we give machines eyes, but we also give it the brain, maybe that's even more important. So it means it's all about first taking an image, but then it's all about processing that image. And we do that in a factory and the warehouse environment. So typically, in-line manufacturing or in-line processing in warehouses, and that includes guiding like robotic arm guiding that includes identification that means reading of parts or letters, optical characters on parts or barcode reading. We do gauging measurements, quality type of things. And what we can help with this is basically that there are a lot of workers out there who do these tasks manually still today, and we basically can support giving them more meaningful tasks and focusing on other things. As a company, basically, we are riding on, I would say, 3 megatrends. I would say one of the oldest megatrends is probably since Henry Ford created the assembly line. And since then, engineers have been trying to drive more efficiency and more throughput and better, faster, higher. And that's what we are there for. But then certainly, over the last maybe decade or so, there has been also another topic coming into the equation, which is demographics that basically there are not so many labor availability out there who want to do such tasks, right? Because at the end, it's like a task, maybe you look at the same thing for 8 hours, and it's probably, especially in a developed world, people don't really want to do that. And then I would say, probably since the last maybe 5 to 8 years or maybe since the first Trump administration and certainly much more pronounced also recently, nearshoring, reshoring has become also a trend, which drives basically Cognex demand side. We serve an $8 billion market. There's like what we define as like a $6.5 billion core vision market, which is like areas like in logistics. So think about like e-commerce, distributions, warehouses. Certainly, then the second largest market is in the automotive space. There's electronics. So think about manufacturing of all kind of electronics devices and then other areas for medical, fast-moving consumer goods as well. We have entered some adjacencies over the last couple of years like vision sensors that was like organic and then also the optics, where we did an acquisition about 2 years ago. In our -- on our more than 40 years history, we have a strong growth pattern. So especially from 2011 to 2021, about 14% CAGR, have then seen, especially during the pandemic, quite heavy investment and a bit of a post-pandemic slowdown. I've seen the top line going down from '21 to '23, but we have returned to growth in 2024, largely driven by inorganic, while basically on an organic basis, more like started to bottom out and being rather flattish in that year. We work with the most sophisticated customers and let's say, solving the most sophisticated problems in machine vision. That's where we really -- our core DNA is as a company. So we deliver the highest performance and solving the most complex problems. That's kind of where we originate. But over the last couple of years, we have been also moving into what we call an easy-to-use and easy-to-deploy space. That has been enabled by some of advancements on the AI side, which I will talk about in a minute or so, just to broaden our customer base and basically serving also a broader part of the market. As just mentioned, AI is a key theme for Cognex. So we have been embracing AI very early on when a lot of people have not even been talking about AI. So we made an acquisition in 2018. We bought a Swiss company called VIDI, who had kind of an inherit or started to do AI models for machine vision specifically. And then we have launched our first AI-based product in 2022. And I would say, over the last 18 months or so, every new product, which we have been launching has an AI component it. So there is no machine vision product anymore, which we're launching without AI. And then within the AI, we will have 2 areas. We call it deep learning. So think about vision software, which sits on the cloud, high compute power, addressing the most difficult problems to be solved, but we also have what we call edge learning. That means it's really a pretrained model, which sits on an individual device is not cloud connected and uses basically the power of that chip in that device and basically has very high speed, high latency and can address some of the, let's say, kind of medium-range problems kind of in terms of difficulty to solve issues. We'll probably talk a little bit more about the AI space in a bit. To bring some of these products, especially the ease-of-use products to the market, we have worked over the last 2 years to transform and expand our sales force. So that means we have traditionally have been -- had a very highly technical and highly consultative selling approach, but we brought on an additional type of sales engineers, more like graduate students who are transactional focused and kind of selling these easy-to-use products. And then lastly or before coming here to the end, we have a very unique culture as Cognex. So I think within my close to 1 year in Cognex, I've been pressing up several times with different characters of different stories and kind of had a lot of fun with our Cognoids as we call them. And if we don't work hard or play hard, you may see a lot of our Cognoids running around on a turf field in front of our office and playing Ultimate Frisbee. So in that regard, having a nice culture here. And then last but not least, while we certainly will talk more about Cognex in the next 25 minutes or so, we have an Investor Day coming up on June 9 and 10. So if you would like to learn more about the company, please do visit us. Thank you.

Awesome. If you guys could actually keep that up, a couple of my questions are actually addressed in some of your presentations. So I'd love to start with that, the $8 million (sic) [ $8 billion ] market served, if you want to go back to that slide. All right. So 2024 was a good year for Cognex despite the market conditions. You guys grew 9% year-over-year, about 1% organically if you exclude Moritex. 70% gross margins, 17% adjusted EBITDA margins. So I'd just love to hear your thoughts about the market growth rates for machine learning and how penetrated this market. And conveniently, you got that all displayed right here for us.

Right. So yes, no, I think as I mentioned initially, I think it's a highly attractive market in that sense from a traditional growth rate, right? So on the one side -- or maybe let's maybe start historically, right? Historically, we have seen the market growing somewhere in like a low teens, 10% to 11%. And if you look at external analysis, there's like interact analysis or others are out there, they're basically also projecting that forward. Now you may think like, okay, where is that 10% to 11% market growth coming from? So first of all, there's typically like an underlying secular market growth. Think about maybe start with logistics, right? Logistics market may grow whatever, 6%, 5% to 6%, somewhere in the mid-single digits. But then basically, there's kind of that additional automation and AI-driven vision penetration in these markets. As we bring in more AI into machine vision, we basically can solve problems which haven't been able to be solved in the past without these AI algorithms. And that basically enables additional use cases and therefore, drives adoption. And then that, for example, on a logistics market, we would say like, while the logistics market itself maybe grows is mid-single digits, that can add really up to 10 percentage points to somewhere in the mid-teens we would expect like a logistics market to grow. So logistics is overall the market we think will grow the fast, which today is one of the least penetrated from an automation and machine vision area. And then we have other markets which may not grow as fast, for example, automotive, right? We are not very, let's say, bullish in terms of underlying automotive growth, but you will also have some level of additional automation and machine vision penetration there. So all in, we would see like a 10% to 11% market growth. And back to the beginning, we think we're in a very attractive market.

So Dennis, I think of you guys as the market leader in this space, right? And the fact that you guys are only kind of like 15% to 20% share, can you just talk about competition? Who are the other big players that you see the most?

Right. So we -- when we talk about the competitive landscape, we basically separate between logistics and warehouse automation and factory automation. So factory automation includes automotive, electronics, medical and all the others, right? So in that area, our largest competitor is a company called Keyence. That's a Japanese company, running more a -- not a very Japanese business model. So they're very sales coverage and, let's say, sales focused, which is, I would say, not so typical Japanese style. And they basically -- while we traditionally serve the most sophisticated customers with the most sophisticated problems, they are more in the broader market, serving basically a large breadth of the market with a very good market or sales coverage. And then on the logistics and warehouse automation, we have a company called SICK, S-I-C-K, that's a German privately owned company. They are also doing like safety sensors, and they're coming more from the vision sensor area, where they have moved into the, let's say, machine vision space over time, and they are basically the leader in the logistics and the warehouse and logistics space, and they are more the incumbent there, and we are basically have been entering the logistics market probably like about 10 years ago.

Okay. All right. So -- and you hit on this a little bit, too, but 70% gross margin selling cameras, barcode readers implies to me that you have a ton of software in your products. As you said, you're a software company, but you're launching all these AI-based products. So can you just talk -- what's the difference between AI-based products, software versus kind of legacy products with all the software?

Yes. No, absolutely. See, I think maybe traditionally, I give you a bit of a journey of the company. So the company started really as a truly software-only company. So if you go 40 years back, it was software only. So we provided machine vision software. And then our customers would buy some dumb hardware connected to the software and make that work. I think as time progressed, more and more customers wanted to have an integrated. So basically, the next step for us was to embed the software into the devices and then so like a one-stop solution as a device. However, that was all what we call today rules-based algorithms. So that means some very smart PhDs and software engineers would sit there and think like what's the problem the customer want to solve and then write like if this, then that. in a very, let's say, sophisticated way and certainly building a lot of domain expertise over the years. So that means we have been the leader in this rules-based software approach. But then rules-based software certainly requires a lot of software engineers writing a lot of code, right? That's kind of what it is. Now when we move to AI, we're basically going away from this rules-based approach, but we are building models similar like maybe you think about it ChatGPT, right? They're building large models, which then with the models that they have created, they can address certain specific or universal type of requests and questions. And we are basically building very specific for the machine vision, industrial and factory automation and warehouse automation, specific machine vision models. And with these AI models, we can do 2 things: A, we can solve problems you could not solve in the past with rules-based. And I can give 1 or 2 examples in a minute. And then the other aspect of AI is that we can make it easier to use. So for example, in the past, when you bought a machine vision system, when you set it up, you may need to go through some tuning steps. And you needed to have an industrial automation engineer to help and basically do this tuning, this initial setup of that vision system. Now we recently launched a new generation of ID readers, and they have an AI auto tune. So that means you're basically removing additional engineering time to set up device. So that means the AI also helps in terms of an easy-to-use and an easy to deploy approach. And maybe to have one example for like how does AI make it easier. I give you an interesting example. I went a couple of months ago, I went to a big brand name for consumer goods. So they do a lot of different things, but I walked the line of a simple thing like a dust sheet. So think about like a piece of paper you used to clean. And in the past, basically, you could use machine vision to inspect to make sure are the edges cut in the right, is the shape right? So that means you basically could do some inspection tasks of these dust sheets, which are flying by like 1,500 pieces a minute, right? So that means extremely fast. But what you could not do is like to inspect the pattern of the sheet because at the end, it was white on white. It's like, can you inspect the structure? Is the sheet structurally intact? You could not do that with a rules-based algorithm because it was just too complex to address. But with an AI, you can really trained through pictures with good pictures, showing, hey, this is good and then a couple of pictures which are bad. So now with an AI-based system, you could really identify and also inspect such kind of structural issues. And that's kind of one of the examples where you can see that you create additional market adoption and additional use cases for machine vision if you use AI-based tools.

So do the AI-based products have better gross margins? So if that becomes a larger percentage of your sales, would you expect to see some gross margin expansion?

I mean we have typically already commanded very high gross margins, right? So we think about the AI more like as an market accelerator and part of creating market growth, right? So in that regard, when we think about our own kind of years forward, and we think AI will help to get us to this 10%, 11% market growth, but we would not necessarily think like that it's a lever for margin expansion as we're already commanding that level as a software company.

Okay. Just to stay competitive and kind of lead the competition, too. Yes. All right. So maybe go to the new sales initiative. To me, that was pretty important for you guys. I think you had 80,000 customer visits last year, 3,000 new customers. I think the OpEx grew by $23 million in '24 because of that. Can you just talk about the sales efforts here to accelerate growth?

So again, maybe first, quickly back to the strategic rationale of the initiative, right? Traditionally, you have been serving the most sophisticated customers with the most sophisticated problems. That has given us really being the technology leader in that space, and we have a very high market share with this group of customers. But they are really sitting at the top of the pyramid. And that means there's a large market beyond that, right? So they are -- traditionally, we have been serving about maybe 25,000 to 30,000 customers, but the total market is maybe 300,000 customers, right? So we serve maybe like top 10% of the customers. Then, however, when we see like how is the rest of this market, and we looked at companies like Keyence, for example, and they are actually commanding also attractive gross margins in that business with these other customers. So we thought like, hey, here's a fantastic opportunity that you can expand basically into customer areas, but you're not -- right, when you go from the top of the pyramid more down, you would think like you would lose margin, gross margin would come down, but we can actually see that this is not the case, right? So we saw that from competition. We saw that over the last year that actually these sales engineers going to these new customers that they are commanding actually accretive gross margins, right? So they're selling easy-to-use, easy-to-deploy products. They're usually smaller order sizes, and it's part of why we can ask for higher margins. These are not the largest customers, but it's a gross margin accretive business. So therefore, we really are focused on this initiative, and we have invested quite some money. So you mentioned $23 million and that was the incremental dollar value in 2024. So we already invested about $28 million in 2023. So on total like run rate basis, around $50 million run rate. So it's a significant investment for the company. And so what we're really trying to achieve is like expanding the strategic positioning of Cognex. And therefore, we are ready to make such investments even though we understand that it creates a quite a headwind in terms of the bottom line for the time being.

Okay. Perfect. Just a competitive question. RFID, right, another company that follows Impinj, and they've had a lot of success recently with logistic customers like FedEx and UPS. Were they previously using barcode scanners? Do you view RFID as a competitive risk to some of your barcode scanning?

Yes. I would say RFID, right, if you think about it, it's probably a technology which has been around for more than a decade. So in that regard, we have seen that customers in warehouse spaces have been working with RFIDs and very often, they have been working it alongside with machine vision systems. So we have not really seen that as a competitive threat in that sense that would kind of disrupt of what we're doing. And we have also not seen that it has created like a very large market share in the overall identification space in the warehouse and logistics automation space. I would say maybe beside RFID, there are other technologies maybe to mention is like, for example, a LiDAR technology, right? So that means some of what we do with measurement, for example, you could also achieve with the LiDAR technology. So there's some adjacency over there as well. And in that regard, probably a LiDAR technology is much more adjacent to what we do. And we actually do have in the logistics space, some areas where we use some LiDAR technology ourselves. So there are these type of adjacent kind of technologies. But overall, they are more helping with the overall solution than they're kind of competing with each other for market share or tech share.

Okay. If you could go back to that $8 billion market share to you, I guess I want to kind of go through the verticals a little bit. So start out with how about automotive, right? I mean that was a struggle for you guys and now we are in the tariffs. And can you talk about your expectations for the automotive vertical?

Yes. I know automotive has been probably like the most difficult markets to be in, in the last 12 months, maybe 18 months, right? So traditionally or maybe not traditionally, maybe over the last 3 to 4 years, automotive has actually been our largest market and our -- in terms of revenue, was no surprise last year by logistics and warehouse automation. We had had a very nice success there, especially in 2023, also with the a lot of investments into EV battery and kind of capacity expansions into this new kind of technology. But I think the market has seen a lot of uncertainty and a lot of challenges itself, right? So it means the transition from ICE to EV has not gone as fast as maybe many expected. It's not stalling, right? I mean it's still happening. I think I read some recent numbers, maybe BEV sold whatever, 8% up year-over-year, but people may have thought it would be 30% or 40%, right? So it means there has been a lot of overcapacity created in that area. And therefore, not a lot of investment is happening in EV space. And then at the same time, if you look on the ICE side, also there are not a lot of investment happening. And I think talking about specifically North America, at the moment, a lot of uncertainty out there in terms of the tariffs, and that doesn't really help with good investment climate. So in that regard, automotive was our weakest market in 2024. Actually, in terms of revenue, it declined by 14%, and we still have very muted expectations for 2025 as well.

All right. How about the more exciting markets, logistics, and I'm assuming that includes e-commerce. And in semiconductors, you think the growth you've seen there is sustainable?

Yes. No. We're very positive on both logistics and the semiconductor space. So logistics, maybe first talking about what do we mean with logistics. So there's -- e-commerce is a big portion in that. That can be big names like Amazon and then certainly also players like Walmart and Target. But then also think a bit more globally, companies like Shopify, Instacart, coupons or some of the Asian names out there. So that's kind of an area. Then there's really like warehouse automation, like almost each industrial company does have a warehouse. And very often today, they use maybe handheld barcode scanners, something we don't do, that's really like we would consider it maybe no tech. But -- they may also want to further automize and use what we call machine vision tunnel. So that means really much more automated type of either incoming or outgoing identification. And then the last subsegment of logistics is what we call Parcel and Post. These are the guys like the UPS, the FedEx and the DHLs of the world. So that's a market we have traditionally not served very much. We have been starting to entering that maybe over the last 2 years, but that's more a market share gaining market, right? So that means e-commerce is a strongly growing market, whereas automation is a strongly growing market where we have been growing with the market and gaining some market share, whereas Parcel and Post is a very stagnant market where we rather have been focusing on winning share. And then SEMI. SEMI has been actually our, in terms of percentage base, strongest growth market in 2024, certainly in part driven that 2023 was a very down year, right? So growing off a low base. But we have seen a lot of investments into high-bandwidth memory chips, and that has been driven a lot of demand for capacity expansion. And typically, that's what we see in SEMI that when capacity is expanding, and our business is going very well there. And what we have seen so far, certainly, there was the DeepSeek announcement in January and that created a lot of questions around like, hey, is that sustainable. But I think we can clearly say, yes, we see that the market is continuing. And when we looked at some of the, let's say, hyperscalers on their announcement that they're talking about their CapEx, we're not seeing any stop or any slowdown in terms of investments in the SEMI market.

Let me just pause and see if there's any questions from the audience, Matt?

Can you talk about what exactly you're doing at size of logistics was last year and what your run rate growth flows in the fourth quarter? And then can you speak specifically to what extent you're doing inspection and to what extent you're doing 3D? And talking about the whole 3D taking place in [indiscernible] and even last week, they closed, and I don't even know how big [ Photoneo ] is and Intel shutting down the 3D camera business, it apparently was a market leader. What's happening -- [indiscernible] and what's happening in it?

Right. So logistics, maybe -- let's start there. So logistics, as mentioned, largest market for us, both in market size as well as in revenue size. It was about 23% of our revenue in 2024. So that means about $200 million or so, has been growing very strongly in 2024. It's about a 20% year-over-year growth. And what we have said on the most recent earnings call was that we expect a similar strong growth to continue in 2025. And in general, we believe that if you think now further out, like maybe 3, 4, 5 years, the logistics will continue to be the most strongest market in terms of growth percentage, driven by -- it's the market which has the lowest automation degree from all the markets which we have been talking about. And therefore, we are very bullish about the logistics market. Now to your second question, 3D. So 3D applications, you find, for example, very strongly in automotive. So you will not find a lot of 3D applications in the logistics space, for example. So it's really very strongly automotive. I would say we do have a strong 3D offering. We last year launched the first 3D machine vision system, which is AI-enabled, call it the L38. And certainly, on 3D, there have been other players as in any of these markets, right? You can see it's a fairly fragmented market with Cognex having a 15% market share. There's a lot of gray bar in each of these markets. And then there are companies they try to enter and sometimes there are companies they try to exit, right, because they have felt that they are not very competitive in the space, right? So I mentioned Intel. There may be other companies who we feel like may get started to wash out some of the smaller Japanese players, for example. And then Zebra has been in general, trying to move from their, let's say, warehouse logistics, maybe a little bit less tech kind of into a more machine vision player and did a couple of acquisitions there like a Photoneo, for example. And we certainly understand what they're trying to do. But I would say it overall doesn't change the competitive dynamics, which we see in the space.

[indiscernible].

We haven't disclosed like how much we do in 3D, but you could imagine that it's a good share. Yes, we haven't disclosed, so we typically don't talk about product lines.

Curious on where the intersection or with some of the wearables that are out there that are used in a more commercial setting smart glasses or things of that nature. Is that an area that is feeling it somewhere where you can be a provider either the software or the hardware [indiscernible] that market?

Right. Yes, I would say -- see there are quite some adjacencies if you think about like using kind of image acquisition and processing. I think you mentioned wearables, for examples, but you could also go into cars, autonomous vehicles, right? So you have some camera tech there. And the latest discussions we get is about humanoids, robots, like would there be -- so there are a lot of areas where you have some level of vision tech. I think as a company, I have been very much focusing on discrete manufacturing and maybe discrete logistics. So I think as we do strategic investments like in sales forces, we are right now really much focused on staying in this broader factory and logistics automation space. Would we have like the base competencies to be active in these areas? Probably yes, but it's, at the moment, not part of the strategy.

Any other questions from the audience?

[indiscernible]?

Right. I mean, see, we have this very high level 4 use cases, right? Certainly, they are becoming much more specific when you talk on [indiscernible], right? We talk about guide, identify, gauge and inspect. When you think about innovation and going to the next level, it really happens across all of these 4, right? So take something which a lot of people think is a very simple task, which actually it is not. But let's start on identify, which is either barcode reading or it is optical character recognitions in part. So when you introduce AI into that, you could, for example, think about it in a distribution center, you want to run parcels and you want to run them as close as you can, right, and as fast as you can. So that means it's sometimes very hard to read a barcode, which is down here and the next parcel is just side by side. So with AI, you can do like enhancing, you can do accelerated finding, you can enhance the picture which you get and then you can do additional filtering. So in that regard, almost in each of these kind of 4 main categories, with AI and with general innovation, you can drive additional use cases and in general, come back to that objective that you can do things faster and better and with better quality and basically drive efficiencies throughout a factory process or through a logistics distribution center.

And with that, we do need to pause. We're at the end of our time. So Dennis, thank you so much, and good luck with everything.

Thanks a lot.