Are we all set? All right. Thank you, everyone, for joining us. My name is Joe Giordano. I cover industrials and industrial tech here at TD. As you can see from every sign outside, we're kicking off II season. If you think we've earned a vote, we really appreciate it. It's important to the company, so therefore important to me. Thank you very much for that. We can get to the real stuff now. We're really excited to have Cognex here with us. And Rob, CEO, he's going to kick it off with a couple of minutes of a brief overview, and then we'll jump into Q&A. Feel free to raise your hand anytime. I'll stop, and let's make it as interactive as we can.
I think it's more interesting that way, even though 0% of the chance that it ever actually happens ends up being just me anyway. You have the option if you want. Rob, thanks for being here. I'll turn it over to you, and then we'll get going.
Yeah, thanks, Joe. And thanks, Cowen, for welcoming us. And you've covered Cognex a long time, so I appreciate that. I'll make a few quick remarks, really aimed at those who don't know Cognex before we get into some specifics. Cognex is a technology growth company focused on advanced automation. We make technology called machine vision. Think of it as the eyes that help computers see. We apply that technology to the most sophisticated manufacturers in the world, manufacturers of discrete products. Some of the applications would include the manufacturing of smartphones, of medical devices, of automotive parts, of e-commerce packages as they move through a sorting center and make their way to your door. We have about 30,000 customers that we sell to. Generally, they're among the most sophisticated manufacturers in the world. Our revenue was over $900 million.
Over the last 10 years, we've had adjusted EBITDA of 28%. Our technology is primarily software. We spend a lot on R&D, around 15% of revenue. We're considered the technology leader in this industry that we serve. What is it that we do? We're in the business of everything we do is around machine vision. Just think of your own vision for a second. You have an eye, and you have a brain. Your eye is gathering information, and your brain is making sense of it. That's what we're doing in the world of computer hardware. We're acquiring an image using a lens and optics. That image goes onto an imager in the form of pixels. You've all heard of pixels. The light hits the pixels, and it creates a number, or a lot of numbers.
It sends that data, that's the eye part of it, to the brain where Cognex's machine vision tools or algorithms make sense of it. This is very sophisticated technology. It's been around a long time. The company's 43 years old. In some ways, it's technology that still has a long way to run. Just think of your own vision again for a second. Light is changing. Things are moving. There's color, there's vibration. Think about a manufacturing plant where we sell into. Products are moving quickly. Things are changing. It has to work at 99.99%. We think our products are used in the manufacture of about more than 2 billion products a day. Anything short of super high precision and how you make that work isn't acceptable to those customers overall. What are generally the things that we're doing?
We think of four things we're doing. We're guiding. Our technology might guide a robot to put a chip on a board or a windshield on a car. We're identifying. We're reading letters and numbers or barcodes on a product. We might do that on an e-commerce package or a silicon wafer in a fab. We gauge, which means we measure. We might look at an automotive part and say, is that groove in the brake pad the right width and depth? Finally, we inspect. We look at something and say, is that bottle cap correctly on that product? Is that pharmaceutical pill chipped or damaged? Is that smartphone case presenting to the customer the way it should? The final thing I want you to know about Cognex is we have a very strong culture.
We refer to our culture as work hard, play hard, move fast. We are kind of famous for our Halloween celebrations. We celebrate leap year by inviting Cognoids—we call ourselves Cognoids—to jump out of a plane with the founder with a parachute on. Don't worry. We have a lot of fun. Look at our annual report if you don't believe me. OK, with that, should we?
Yeah, let's do it. Like you said, machine vision has been around a long time. I've been covering a long time, even when you just started really ramping up with some of these big markets that you're in now. I feel like over the last decade or so, the rate of change has increased dramatically. Maybe how would you describe the landscape over the last kind of decade, how that's kind of played out?
Yeah, and you're talking about the technology. You're talking about the use case.
I think it's both, right?
Yeah, I think it is.
It's kind of both.
I think it is. There has been a big change that's gone on in the world of computers. That's from what I will call rules-based to AI. Cognex started out, and for up until about 2017, everything we did was based around very precise machine code and algorithms that we wrote to describe a problem. Then something happened, which is the concept of deep learning really got traction. It's been around a long time, but it really started to get traction. Instead of writing precise math, you created a model, and you trained the model. We saw that trend quite early. The founders of the company, who came from MIT, had seen many false dawns around deep learning, where the technology, we thought it would change the world, and it didn't. Suddenly, here it was, and it was really working.
We began applying it to machine vision. What that has meant is there are many more things that it can do. What it specifically can do are much more human-like tasks, things that we're really good at, looking at an area and seeing imperfections that aren't necessarily precisely described. That is one change that's gone on. The other thing is, like all technology companies, the technology has become smaller, more powerful, easier to use, less expensive. It has gone from being used only by the most sophisticated industries. I'm thinking of semiconductor and automotive. Think where there really aren't a lot of people in general. It is becoming more available to lots and lots of customers.
Particularly, that would apply to, more recently, areas like packaging companies that before, when we'd gone with our technology to look at that bottle cap running on a line, it was too expensive and too difficult for them to apply. As we've taken this new AI technology and we've made it run much more simply, less code-intensively at the edge on a processor on a smart camera, it can be sold and applied and be working in a matter of days.
Yeah. When I first started covering you guys, logistics was not a market that we talked about at all. Now it is one of your largest end markets. How do we think about the potential for something like that popping up again, like another huge market kind of coming out of nascent stage to something very developed? What are some of the potential possibilities for something like that?
Yeah, I think to answer that question, I'd be thinking about where do you see a lot of people? Where do you see a lot of people in manufacturing doing things that probably in the long run, computers and technology will do better? One area we see it in is human visual inspection. We estimate there are about 35 million people in the world who go to work every day. All they're doing is really looking at something and saying, is it good or bad? If you go to a big contract manufacturer in China, you see halls and halls of thousands of people looking at things like connectors and saying. Vision technology is getting to the point where it can cost-effectively replace those types of applications. That will play out broadly. It also might mean we'll get into markets that we haven't served before.
The founder of our business, Dr. Bob Shillman, used to like to say, we don't inspect anything made by God. It's like natural variations are quite hard for machine vision to work on. With AI, you can train it, and you can show it 100 coffee beans and say, these are good, and these are bad. You can set it up and running, and you can replace potentially people who are doing that kind of work. That's one area. I think another area where you see a lot of activity today, which I think in the years to come, I don't think in the next year, but I think in multiple years to come, will be replaced. That's more around picking of things that robots will start to do more and more. The concept of picking is very difficult. It's a difficult concept.
It involves all sorts of mechatronics and vision. It's going to get there at some point. I think we'll see a lot of replacement of those areas too. Maybe the final area I would say is I think I'm waiting for the next consumer electronics big thing to happen. Certainly, when we saw the smartphone take off, when we saw facial recognition and OLED screens that happened, it's been a little while. I'm looking forward to that happening. I think when that does, when we get mass-produced products in high volume, the world's going to look to Cognex to help manufacture those. I did read the story. You did probably too about OpenAI and Jony Ive and their quest to develop something that's perhaps a successor to the smartphone. I think these are all areas of interest.
Over the last two years, you've started to pursue this emerging customer initiative. Why make the move when you did? What's gone right? What hasn't? Maybe just describe that whole kind of strategy.
Yeah, yeah. Cognex got into deep learning technology around 2017. We hired this amazing engineering team. We bought this amazing engineering team in Switzerland who made these very powerful models. They were great, but really only the most sophisticated customers could use them. That team developed something called edge learning. What it did is it pre-trained these very powerful models to use much less code and have them work on a small piece of hardware that we can manufacture and sell very profitably for a few thousand dollars. It brought a lot of that capability to customers who we could not serve profitably before. With that, we said, great, but we need a sales channel to sell that. We need a sales force to go and reach a lot of customers. We think of our market like a pyramid.
We kind of have a very strong share on the top of the pyramid, the world's most sophisticated companies. Below them are, we estimate, perhaps 300,000 customers who can benefit from our technology. We created a sales force to go out and do that. We hired out of college young salespeople who we then trained and put in the field. In the first year, we spent about $30 million hiring, training, and deploying them. They made the first class 80,000 sales calls. They won 3,000 new customers. They sold at high gross margins. They referred a lot of business to the rest of our salespeople. Now we are on our second year, where the second class is entering the field. I should say that class last year was selling by the end of the year about $1 million a week.
That's kind of where they were. We're on a pace to keep going with that and drive and meet those new customers and sell to them. You asked what didn't go well. I think because we were probably a little over-indexed on automotive in terms of where we sent them to sell. We were very comfortable in those markets. The automotive industry was not good. I think we're realizing that there are other markets, such as those in packaging, that I think are very good for these customers, these salespeople to call on and sell more broadly.
What does that market look like? Is it people who aren't using anything? I know Keyence is a major presence in that area.
Yeah, yeah. I go out on sales calls with these guys, pour them. I just try to observe. A lot of companies I've been to have been in food and beverage type applications. There's nothing there today. They have a problem. Something's not working in their production. They're having quality issues. Boxes aren't being closed correctly, or they're in the wrong orientation, or the things that can be simply identified quickly by a salesperson who can put in a solution to make it work. They're pretty small companies, so without something.
It sounds--and we'll talk more about this at Investor Day coming up--but if M&A becomes more a part of the strategy on a more consistent basis, what are some logical kind of adjacencies that you think could make sense?
You mentioned the magic word, which was Investor Day. June 10th, please come to Cognex, where you can really see our technology, meet Cognoids, understand kind of our business more, see use cases, and hear from our new CEO, who's taking over from me at the end of June, about his vision for the business. Your question is about where can we grow in terms of adjacencies. I think I'd start by saying I don't think we need to move into adjacencies to grow strongly. I think in our own surf market, there are a lot of good targets. There are a lot of companies that I think make good acquisition targets for us. Or they're almost like near adjacencies, like we did with Optics, where we bought the company Moritex a couple of years ago, which you can't do machine vision without a lens.
They were the premier lens supplier into our industry. They bought a lot of penetration in Japan and semiconductor that we liked. I think those represent the real sweet spot. There are adjacencies, which are sort of markets that we've dipped our toe into, where we can see going deeper. I'm thinking particularly more in the sensor space. I don't expect us to make some kind of completely new thematic adjacency. We won't be getting into the eye care business.
Fair enough. To your point, you announced your retirement after a long time at Cognex. What do you think are the two or three most important things that your successor needs to get right?
Right, yes. Matt Moschner is taking over from me. I've been the CEO for 14 years and with Cognex for 17 years. I think the things that he is focused on, and you'll hear him talk a lot about at Investor Day, one is leading in the application of AI technology to factory automation and machine vision. That's something we're on a mission for. We love what we're doing. The world recognizes us. We have to keep going with that core competency. That would be the second thing, really around customer experience. Our number one company value is customer first. We've invested a lot, certainly, in systems and processes to do that more effectively. When you're looking after a customer when you have 10, it's different than when you have 50,000.
We need to make sure that we're making progress where every customer feels very special about the way we treat them and the satisfaction that they have with us. The third would be growing our number of customers. We've really got to turn the crank on that. We can't just be the supplier to the elite, most sophisticated manufacturers, who love us, by the way, and love our engineers and love working with us. We need to broaden that number down into the base of the pyramid and do that. Those are the three things I would say. If I had to have a fourth, it would be culture.
Yeah. I don't know him personally. I'm excited to meet him in a few weeks. What gives you the confidence that he's the right person to take over? Why was now the right time to make that transition?
I hired him eight years ago. He was reporting to me for a while over those eight years. We take succession planning very seriously at the board level at Cognex. We have had a number of really excellent candidates that we have brought along over that time period. Matt has risen to the top. He has been field tested. He led our logistics strategy back in, I think it was 2018, when we really started to inflect into some major growth. He managed our response to the chip shortages and a fire that we had at our major contract manufacturer. He led our entire engineering organization. Most recently, he has led the logistics business back to growth. I think his credentials are awesome. I think Cognoids see that. I think they are looking forward to his energy and drive.
Right. Maybe I'll pause for a second and see if anyone in the audience has anything before we move on. Seems like I was right in the beginning.
Is that always how it goes down? I travel the world. And in some markets, there's 100 questions. Like India, everyone would have a question. And then in other markets, Japan, everyone looks at their shoes. We're in Japan today.
Yeah. So we get asked a lot about the difference between computer vision, smart vision, where one is appropriate, where the other is, and different price points, different applications. How do you see those two markets evolving? Where do you think you're positioned long term? What's your moat there?
I think if I'm interpreting your question correctly, you're saying sort of like a generally available vision models, like Llama, what's the term?
Correct, or like writing your own code on an off-the-shelf optics.
Right, right. We have been in this industry for 43 years. I would say there is a dynamic that was true before and is still true today, which is what you have to do in our industry is have a lot of industry know-how about applying technology to work at 99.99%. A lot goes into that. It is the technology, but it is also the support you have with customers, the global rollout, your experience having seen applications before that they are doing. I think probably the question on everyone's mind is, are new machine AI models going to replace what we do? I think come to Investor Day, because you will really hear, I think, a very much better than I can give you a coherent response to that. The answer is the precision to do that is extremely difficult to do.
It requires a lot of domain expertise. Our market is not so big that any of those models are being developed with it in mind. We take large models, and we benchmark against them, like Llama. We developed our own transformer model that we launched last year, which way outperforms those models in our own applications. We apply that in a way that customers really can use and make sense for their production. We have the best customers that we're working with on it. That is kind of how it plays out. The technology itself, how we then can pre-train it and put it into efficient embedded hardware at the edge, is something we do extremely well. How we sell that really makes a huge difference.
It's true today, and it was true 17 years ago when I joined, that we outperform pretty much everybody in most of the difficult applications for that industry.
A lot of when we get people more cautious on the story, a lot of it is like, how is this not some sort of inherent threat in the future that these models keep getting better? It is some sort of an inevitability that it becomes able to do that to the level that you can. How do you kind of protect against that? How do you respond to that?
Yeah, I think they will get better. And we will get better at how we use them and apply them. I think of the story of maybe every technology company in Cognex is we're inventing on the high end. We're doing difficult things. And over time, those become easier and easier. They become commoditized. We don't do them anymore. An example would be reading a paper barcode, a paper-printed barcode like you see at the airport. By the way, they don't do that very well. You're waiting to board. There it is. Those things. We don't really sell into that market anymore. We're inventing things to replace human visual inspection and applications. There's a lot of domain knowledge, a lot of IP and patents. We have over 1,000 patents in this area that we then manage.
We work with the most sophisticated customers as it migrates. Then over time, we sell all the products to customers who are more price-driven or whatever. We are inventing high-performance technology at the high end of the curve. I see that continuing. I think our ability to understand where the technology is at the cutting edge, as applied to our not huge industry, is kind of how I see that playing out and supporting customers globally.
I think cost leverage at the Optics line seems to be a pretty big opportunity. I'm just curious, what is your ability to take AI and kind of leverage that to maybe have less human talent needed to program or something like that?
Yeah. Now we're talking about something different. We were talking about how we apply it to our own hardware. Now we're talking about how do we apply it to our own engineering methods. Right. I think every company is working on that idea of how do they make their company more productive and capable. In our case, we have a lot of software programmers and tools like GitHub Copilot, which really allow a lot of code to be written automatically. It is your Copilot. A Cognex coder is writing the difficult pieces of code with the domain knowledge. The Copilot is taking care of a lot of the simpler parts that happen. That can be as much as 30% of the code that is being written today. It will be more. You can code like the best coder.
You can have the Copilot getting it done. We are either going to pivot that into fewer engineers or more applications and more technology advancement. A lot of that will depend on how the market is doing. We want to grow our Optics slower than our growth. That is what we see ourselves doing.
You mentioned culture in the beginning. It's definitely going to be on display in a few weeks. I know it draws people to the firm. As you move into things like emerging customer, does it require a different type of style? I know the established competitor in that specific world has a very, very different culture than yours. You tend to take people from them. They're also very successful in that world. Is there some element of maybe there is one way of managing to one outcome versus your historical kind of core?
Yeah. I mean, to kind of expand on your question, we have a large sales force. And lately, we've recruited a lot more people straight out of college. And the existing sales force were often industry veterans who had a lot of experience, some of whom we recruited from competitors. Like the one you're mentioning, they came to Cognex. I think a lot of them felt like they'd died and gone to heaven. It was like fabulous company, great culture, better technology. What was not to like when they were being managed more in a very aggressive culture? That's been great and great at retaining them. This new profile we have, I think they enjoy a lot about our culture. We've trained them up and made it where you'll see they're playing Ultimate Frisbee with us.
They have great sense of camaraderie and a class among them. They enjoy our culture in a lot of ways. It is different. I'd say our philosophy is more we value people's individual contributions. We encourage them to be creative within a certain prescribed things. Like some cultures don't allow salespeople to go out and make a call on a Friday. For instance, that's not our culture. We think they should decide on how they manage their schedules. We will manage them more by their performance and the metrics that we see.
I have no doubt that they like it. I don't think I've ever seen someone go the other way.
Yeah, yeah. I think it's going well. We implemented Salesforce about five years ago. We were really on too many iterations of that, where now we're really able to understand the work that goes on and help them improve it. I think it's true that in any company, recent graduates have a higher attrition rate. Certainly, that's something we expected and are seeing.
So you don't feel like success is not dependent on some sort of grind at that, sort of, it doesn't have to be that maniacal.
Joe, I think all success is dependent on grind.
I guess the wrong time.
The question is, are you enjoying it?
Yeah, you're unhappy.
Work hard, play hard, move fast. I tell my children, you're not going to get anywhere unless you really work hard. It's like it's 90% perspiration, 10% inspiration. We have a lot of inspiration, too. We play hard. We have a lot of fun. We move fast. We do empower people to do stuff, Cognoids to do stuff. We have an award we have called Don't Do What You're Told, Do What's Right, where people innovate. Recently, we gave it to an engineer who his boss said, don't try to do that with machine vision. It's too difficult. He went off. He spent weekends working on it. He came back. He made it work. We gave him an award. That's the kind of thing we do. We really respect at Cognex and celebrate.
That applies no matter what your level is in the company.
Perfect. Before we wrap, maybe we touch on the end markets a little bit, where we are. I mean, you have three large ones. Where are we in those relative cycles? What's your most up-to-date kind of view?
Yes. So actually, even we think of now four markets with packaging being the other one, which I think would be our third largest market now. OK, so I'll tick through the markets. So not one of the big ones, but semiconductor doing very well. Our fastest growing market last year. There's some sort of caution around some of the trade stuff that's going on. Still, I think we're in a very nice growth position with that market overall. Logistics, our biggest market, grew 20% last year. We see a lot of momentum. The kind of post-COVID kind of tightness in spend and overcapacity is behind us. We're now seeing a lot of great growth. That market is going very well. Consumer electronics, we see modest growth this year. I think we tend to have visibility around this time of year into the whole industry overall.
I think there's lots of opportunities for growth coming. At the moment, that's how we size this year. The market that really continues to be very difficult for us is automotive. So much pain in that industry. We shrank 14% in automotive last year. I think it won't be that bad this year. I don't expect. It is still pretty grim, I would say, in terms of what those companies are going through. I do think there's only a certain amount of time you can hold your breath and not spend money on automation. I think some of that will come back our way. It will be interesting to see where is the automotive industry going? How is automation going to play into that? Is it EV? Is it hybrid? What models are going to work?
On logistics specifically, I mean, I agree. Everything that we heard before all the tariff stuff started was that, yes, there's definitely momentum off of a trough there. Was there any sort of paralysis in decision-making as some of these tariffs got crazy? Have you seen any of that dissipate?
You're talking really about investment decisions around the world based on huge, huge.
I'm not in that one, too.
In EV is what you're talking about?
I was talking EV battery.
Oh, logistics.
That was a market that I think was absolutely recovering.
Yes.
I was just concerned that maybe that got put on pause briefly as some of these things.
Not that I see. Not in logistics.
OK. Yeah, in auto for sure.
Yeah. I think the markets that have been really impacted by these trade issues are sort of strategically important markets. Semiconductor have been impacted. Automotive and EV battery, particularly, yes. Not logistics.
OK. We touched on a bunch. This will be my last one. You have an Investor Day coming up. Without spoiling it, what are some of the things that you expect to focus on for everybody?
Yeah, yeah. I think you'll see our new CEO laying out his strategy, his vision for the company. I think it will be based around the three things I talked about: the AI leadership, customer experience, and increasing the number of customers. You'll see how we're doing that. You'll hear a lot about our technology. We're a technology company. You'll hear from our technical leaders. Something I love about good technical leaders is how they can explain something very complicated in a very simple way that we can all understand. I think you'll come. I think it will answer a lot of the questions you have about disruptive technology, what we do, how it will play out, how we can do more, what the competitive situation is, those types of things.
I think you'll hear a lot about, you'll see a lot about our culture, as you mentioned. Please come and do that. We do have an event the night before, which I hope some of you will. There's a customer panel, I think, where you can hear about how customers work with us.
We have 25 seconds if anyone has something they want to throw out. Otherwise, we'll wrap up.
Great.
There we go.
Let's do it.
Bring it.
The automotive industry has been difficult. I appreciate the fact that you're leaving. If you were to stay on, would you de-emphasize this business relative to the other opportunities?
Cognex has some amazing technology for EV battery manufacturing. I've got to know this process very well. It's a very difficult, capital-intensive process where our technology can make a huge difference. I think if we can see that market come back and continue to grow, I think it's huge for us. I did say, I think, recently that our expectations for EV were to do about $50 million more last year than we did. We really saw that market. There was a huge potential at the start of last year. It really just didn't happen. I think it sort of has to happen at some point, which would cause us to lean back in.
Great. I think we've got to leave it there. Thank you, everyone, for joining.
Thanks, Joe.
Thank you very much.