Cognex Corporation (CGNX) Earnings Call Transcript & Summary

Good afternoon. We'll go ahead and get started, and thank you for joining us for Cognex, President of Machine Vision and ID Reading Solutions in the industrial sector. Very glad to have Matt Moschner, who's President and CEO of Cognex with us today. We're going to be able to take your questions. So feel free to send those up and probably just raise your hand as well. I think, Matt, we'll just kind of dive right into Q&A, but maybe if you have any opening remarks, just to level set us.

Yes, sure. Maybe just a brief introduction. So Matt Moschner, I assumed the role of President and CEO, July 1, succeeding Robert Willett, who led the business for 17 years. Prior to that, our founder, Dr. Bob Shillman. We're a 44-year-old technology leader in the field of industrial machine vision. And we are going through, along with the rest of our industry, sort of a generational shift in technology. And I think now is -- it couldn't be a more exciting time to be involved with the business. We have ambitious growth plans. We have a balance sheet and a P&L to support those growth plans. And so it's very happy to be with you all today.

Excellent. Maybe just rewind just a little bit. We just wrapped up the third quarter, 13 kind of percent apples-to-apples growth, little noise around some of the adjustments. It was all good, but 13%, strong number, lot of momentum in your logistics business. ,consumer electronics also growing, packaging growing.

Absolutely, yes.

Automotive, maybe not so much. But the fourth quarter outlook seemed to suggest kind of normal seasonality as well. So some normalcy maybe in the business that we're calling out. I guess -- and that does assume that sales will be down high single digit kind of...

That's right.

Quarter-to-quarter on a sequential basis. But I guess maybe just as you go through those end markets, kind of give us a flavor for what you're seeing in demand at the current point.

Yes, happy to. So we define 5 key end markets for the business. The largest and fastest growing at the moment is logistics, warehouse automation. And so maybe it's helpful just to know what is that? Well, we define 4 subsegments in logistics. retail, right? So this is retail distribution. This is working with brands like Walmart, Target, Kohl's, TJX to ship their products directly to consumers where -- and/or their stores where they would own and operate their own fulfillment centers, and they would use Cognex technology to do traceability and other vision inspection. So that's a big part. The second is e-commerce, right? We've been riding the wave of the transition to e-commerce fulfillment last many years. This is the Amazons of the world, the Coupang's in Korea, the Flipkart's, the TIMUS machines in the area of fast fashion. That's been a great grower for us over the last several years. Couple of new segments within logistics, airports automation, where we have great technology as a lot of the world's airports look to modernize and largely automate deploying Cognex vision inspection and barcode reading technology. That's been a good growth driver for us. And lastly, a market we call parcels. So this is the sort of the legacy parcel and postal players around USPS, UPS, FedEx, DHL, relatively new market where we have less share. So logistics being our largest end market, we have -- we're about 10 years into developing technology for those customers, and it really feels like we're hitting our stride. We have good share with the leading providers. The investments in R&D that we've been making are paying off. And I still see tremendous white space for growth. The biggest area, I would say, is really taking what is today mostly about reading barcodes, right? If you go into a modern warehouse and you see a lot of yellow cameras, that's Cognex. Those devices are still largely using the pictures to read barcodes, and we do that better than anyone else. There's a huge opportunity for us to do 2D and 3D visual inspection in addition to reading barcodes. We launched a product just before Q3 called the SLX. What is that? It's fundamentally about bringing our latest generation AI inspection tools to that market, and AI vision is very well suited for logistics applications. If anyone that's been in a warehouse or ordered anything online, what makes the logistics market such an interesting one for a machine vision company is the sheer variety, right? I think the statistic is a typical Amazon fulfillment center has 9 million items, 3 million SKUs, right? So think about just the sheer variety that a system has to perceive different shapes, sizes, surface textures. And so AI tools, which we're heavily invested in, has been a huge unlock for us. And so I think logistics has been a good grower for us, 7 consecutive quarters of double-digit growth in logistics, and I think it should continue, and we expect it to heading into next year, and we have great customer relations there. Automotive, second largest end market for Cognex. And so you can think of machine vision has been sold into the automotive vertical for over a decade, doing quality inspection to the assembly of the mechanical systems, the electrical systems of a car has, over the last couple of years, been probably the biggest drag on growth for the company on the heels of the EV cycle that kind of was and then wasn't, a lot of the geopolitical uncertainty surrounding where vehicles are made, and really what the platform of the future is. But I remain optimistic for the automotive sector heading into next year. As we said in our Q3, we're seeing -- we called sort of the bottom of automotive. We're seeing very much that stabilize and even return to growth in North America. Asia, I would say the one area where we're still a bit cautious on our automotive growth is in Europe, okay? And I can go into why. The next would be packaging. So we define packaging as a vertical, which is really fast-moving consumer goods, sort of consumer products and health care, okay? Together, we call it packaging. Package consumer goods, really kind of driven by consumer demand. And I think we're all sort of attuned to some of the macroeconomic uncertainty around consumer budgets and consumer buying behaviors, but a largely underpenetrated market when it comes to the technology that we produce. And we've made big investments in our sales channel as a way to drive penetration and drive growth in consumer products. In health care, which is the other side of packaging for us, really riding the wave of a lot of the manufacturing investments around some of the blockbuster drugs like the GLP, GLP-1s that we're seeing today as there are more of those products available. And on top of that, as the push to relocate and even domesticate the production of those next generation of pharmaceuticals, Cognex is very much involved with the mass production in that sector. So that's our third packaging, and we expect growth in both consumer goods and health care to carry into next year. Advanced semiconductor manufacturing. So that is the next market vertical that we participate in have for many, many decades. Actually, the origin of the company was in advanced semiconductor. We work with the leading machine builders and OEMs that service the large fabs and packaging companies. And I think we're all aware of the super cycle that we're in the middle of, largely driven by advanced AI chipsets, associated memory and a lot of the capacity being built out not just in Taiwan or China, but also in the U.S. and in other parts of Asia and Europe. And Cognex is reaping the benefits of those build-outs. And then there are several other markets that we're keeping a close eye on that are maybe more nascent for us today that I would call out. Aerospace and defense, not a huge market for us, but one that we see a lot of investment going into at the moment that we're really thinking hard about. And then the other would be data centers, the large infrastructure investments going into data centers and what is the role of automation, both in the assembly process of the servers, the server racks as well as the ongoing automation of maintenance for those data centers. So you put that together. And for sure, there's a lot of uncertainty as we look into 2026, and I think we've articulated that well, but also quite a bit of optimism across those end markets as we look into next year.

I want to dig into those markets a little further. Maybe just before we do so, again, the current administration has put a lot of plans in place to try to incentivize more capital investment, more manufacturing investment. How do you think that starts to manifest? Is it starting to manifest? Can we see it anywhere? Or can you see it anywhere in your order funnel yet? And if not, maybe what's the time line that your customers are talking about?

Yes. I mean I would rewind even to the previous administration as well that placed quite a bit of an emphasis on domesticating semiconductor manufacturing through things like the CHIPS Act. And then, of course, this administration has done a really nice job to create incentives in a business environment where investing in domestic manufacturing makes sense as well. We, of course, are participating in that. The thing I would say is a lot of these plans are multiyear, right, where a lot of the announcements you would have seen in new facilities and new capacity in the U.S. in manufacturing across those verticals could take multiple years, and we are working with those manufacturers in terms of what the role of automation and what the role of machine vision is in those plans. So yes, I mean, is it a tailwind? I think it is. Exactly what the time window is between policy to investment announcements to build out of facilities to go live in operations, it varies by sector.

Some -- I think of your packaging business, in particular, but broader factory automation may touch smaller enterprises. That's where you're also trying to penetrate with a new sales approach. Are you seeing any -- do the tax incentives near-term accelerated depreciation? Or is any of that starting to resonate with those customers, do you think, smaller ones?

We haven't heard those new tax programs necessarily tipping the scale on whether they would or would not choose machine vision or automation in their production facilities. I wouldn't call that out specifically. But for sure, packaging and specifically consumer products has benefited from really 2 things. One, with the advent of AI in our products, and that's not a recent phenomenon. We've been working on that for almost 10 years. It has definitely reduced the complexity and upfront cost to deploy machine vision. And what the customer segment in the market that has likely benefited the most from that are our packaging customers, specifically consumer goods players who tend to be smaller, more regional manufacturers of whatever that is, a food stuff, a consumable product like a razor blade or a shampoo bottle. And so as the technology has gotten easier to use, the ROI for them has improved, and we're seeing penetration in consumer goods accelerate. And then couple that with, as you mentioned, investments we've made in our own sales force. So we largely sell our products through a direct sales force. And we have hundreds of Cognex salespeople promoting the benefits of machine vision around the world. We've increased that organization substantially in recent years, really to reach this cohort of smaller regional manufacturing players that I think are disproportionately in the consumer goods arena. So you put those 2 things together, a product that is easier to use and faster to deploy with a sales organization that is now more capable of reaching them. I think we're seeing a lot of benefit in that part of our business.

Yes. Let's go back to logistics real quick. Again, you talked about 7 quarters of double-digit growth there. That's in the context of maybe when greenfield investment has been lower as well. So obviously, a lot of brownfield investment happening. What's -- to kind of sustain that double-digit growth, what are you seeing in the brownfield arena? Is it the uptake of more machine vision applications? Or do we need the greenfield activity to start to elevate?

Yes, I don't think you need the greenfield activity. So the way I'd characterize it is most -- a big portion of our growth in logistics from, say, 2017 to 2021 was mostly funded through new capacity. These are new facilities. You all have seen them spring up next to the side of the road, big boxes that are either 3PLs, e-com or something else. That was maybe the first wave of growth. You're right in saying as the capacity build-out trend kind of subsided post-COVID, a lot of the focus was on how do we make the facilities that we have much more productive. And that's where we've gone to work and been very successful driving penetration of vision and automation into those existing facilities. I would say today, it is a lot of traceability and putting barcode reading scan points in more places to get better fidelity on where that item is in the building. But for sure, the second piece of that is adoption of 2D and 3D vision for the purposes of inspecting for damage, for missortations, for hazardous labels, for human readable marks. There's a variety of visual inspections our customers and warehouse automation would like to do. We now have the technology to do it, and you're seeing good penetration there. And then on the 3D vision side, which is really about measurement, right, if 2D is about inspection, 3D is really more about measurement, there's been increased demand there really on the heels of tariffs, right? A tariff really drives demand for knowing more precisely the dimensions and weight of the thing that's either coming into your facility or leaving your facility. And so we're seeing increased interest and demand from customers for weights and measurements as well, which we're suited to deliver.

Would you make any kind of regional distinction between the logistics growth? Has it been North America-centric? Or where are the other regions of the world in their...

Yes, I really wouldn't. Our logistics business is relatively more concentrated in North America. That's where I would say we see more maturity, particularly in e-commerce, right, a large player out in Seattle that has been driving that, and maybe in retail distribution as well. But we have a great growing business in Europe, very sophisticated players across those 4 sectors, and we're very present in Asia as well. We've been servicing logistics in China for over a decade, Japan, Korea, in the ASEAN region, I was just in India. They have big ambitions to grow domestically their own fulfillment and warehousing network, both for retail distribution and e-commerce. So maybe on a relative basis, the growth rates would be higher in Asia, but I think we're seeing good activity across regions.

Yes. I kind of want to go back to just -- you called out aerospace, defense and data centers, maybe some nascent markets that you're reviewing. What would be the insertion point for machine vision or ID in those applications that you're thinking about?

Yes, sure. So let's talk about each maybe separately. So aerospace and defense. So these are highly engineered products where precision matters. Maybe the throughputs are relatively lower, but the cost of poor quality is extremely high. And these are also highly regulated areas where traceability at a part level matters, right? And so those are the sorts of things, the ingredients we look at when we look at a new market. Do they care about quality? Is the cost of quality high? And do they care to serialize? Serialize really understand every single component that goes into that thing. And I think you have those ingredients in aerospace and defense. And so machine vision systems are very well suited to solve both of those problems. Data centers is a bit interesting, where if anyone's been to a modern data center, there aren't a lot of people around. And often, what we do is we're replacing some kind of a human operator with a machine vision system, right, trading these eyes for a machine's eyes, if you will. And data centers is really what we're targeting is given the sheer scale and pace of the build-out of these facilities, how and quality matters, right, to have a server fail, a server rack assembled incorrectly, really, that's a big deal. So how do we participate in the entire supply chain of the electronic equipment that goes into a data center from inspecting the raw components, the boards, the servers, the server racks, the connectivity that underpins the server racks, all of that can be automated in terms of its quality assurance, if you will. You take a picture, 2D or 3D and assure that was assembled correctly. We're seeing novel uses of robotics in the automated assembly of these things, again, as a way to keep up with the demand. So that's more on the construction side. On the operations side of a data center, there's a lot of ongoing maintenance around drive -- failed drive removal and destruction. And today, that's a very manual task where a person has to go to the rack, pull out the drive, detect what it is and destroy it in a way that's fully traceable and documented. We're working with customers on a way to automate and drive efficiency in that process as well.

Okay. Implementation of robotic in that?

Absolutely. Yes. And vision as a means of enabling those automated robotic drive removal and destruction activities.

I see. I see. Consumer electronics, we didn't really dive in there. That was a business that had returned to growth after a couple of depressed periods, still the consumer electronics market in general, smartphone market in general, not great. So kind of what's working there for Cognex right now? And you talked about visual inspection. Have we gotten any more traction around those type of applications in consumer?

No, you're absolutely right. I think I missed that in my vertical market sweep initially. But I think what we said in Q3 was we expect consumer electronics to return to growth for the first time this year for us since 2022. So that's very encouraging. We said that growth is broad-based, which is really beyond just any one customer, which is also encouraging, and there's a lot happening in that industry. So consumer electronics from the phone in your pocket to your ear buds, to your tablets, to your laptops, all of those things as they get manufactured and assembled, you would find a Cognex vision in those assembly processes, and there's a number of things that are driving growth in that sector. One, obviously, we're seeing new form factors and new device types being introduced, and those device types like glasses, for example, are very precisely manufactured where cost matters, quality matters and produced in quantities of tens or even hundreds of millions, right? So that's -- those are the ingredients that are really ripe for machine vision to play a key role in the manufacturing process, any substantial changes to form factors in phones, for example, would require a more substantial retooling of the machines that assemble that device, and as the large providers of those smartphones and other devices consider new form factors, you'd see us participating in that transition. Technology, right? AI, we are doing things now around cosmetic defect inspection. I think that's maybe what you were alluding to with advanced AI methods that we couldn't do before. That's driving penetration of vision on those machines. That's very exciting. And then the last thing I would say is, historically, the majority of consumer devices were really made almost exclusively in Mainland China and we're seeing some of those manufacturing footprints shift to other geographies for a variety of reasons, cost, geopolitics, tariffs, whatever it is, to ASEAN, to India, to other parts of Asia. And so as I would say, the supply chain of consumer electronics reshuffles, that creates opportunities for Cognex for us to take share, for us to introduce new technology, for us to work with new partners, machine builders. And so you kind of put all 3 of those or really 4 of those ingredients together, I think that's really underpinning the growth we're seeing in consumer electronics at the moment.

Yes. You made a mention earlier just around deep learning capabilities, AI, Cognex has been on that road for quite some time. I think externally, it's difficult to assess who's winning in AI or with AI, of course. Maybe a few points around why you think you have a leadership position there. How do we -- how do you demonstrate that? How does it compare -- who are your competitors in that chase, do you think relative to Cognex at this point?

Yes. It's a great question. So we've been heavily invested in this new paradigm of AI-driven visual inspection for about almost 10 years, 8 to 10 years. We made a foundational -- 2 foundational acquisitions in this area. One was VD Systems, a Swiss company that was really at the forefront of industrial AI and then Sualab in 2020, which was a Korean company, also doing excellent work here. We've also built out our own team. And so it really is -- it's my top strategic priority. So we are funding this initiative very well. And it's mostly in the area of creating new tools that can do something that we've never done before, finding very hard to describe defects on very difficult surfaces. It's also doing things we've been able to do before much, much simpler, and so we're pursuing both vectors, and at the moment, we're actually trying to tackle both, kind of push the boundary of what's possible while at the same time, making a dramatically simpler experience for the user. If we can do that, that will be a huge accelerator for us, a huge accelerator for our industry. How do we measure if we're ahead or behind? That's a great question. Well, we benchmark ourselves and we work with customers in some cases to do that in terms of like-for-like performance of a visual inspection, and more often than not, we show that we're ahead. We measure it on the percent of our portfolio where an AI-based technology or method is the primary feature or function of that product. We think we're ahead on that. And then we're really trying to take AI from where it is today, which is really still focused on the visual analysis task to how can it transform the rest of the product experience, right? As all of us have become more savvy using prompt-based tools like GPT, how would you bring a prompt-based workflow to the setup of a machine vision system where there is still a big barrier for our customers today in deploying vision. They have to think like a vision engineer. So how do you solve for that? How do you have them do their job well, which is think about the problem they're trying to solve, not have to be a vision expert. And we think AI can bridge that gap for our customers, and that can act as an accelerator of adoption for machine vision across industries. So that's how we think about it. We're heavily invested. We think we're ahead. We have ways to measure that by benchmarking our product performance versus our peers, but that's something we have to continue to stay focused on.

Yes. A lot of what you just discussed relates to -- that's done in software -- and then that's viewed as one of your key competitive advantage, just the software algorithms, I think domain expertise is probably underappreciated to a degree as well. What role does the hardware engineering play in this, particularly when you talk about AI and edge?

Yes, it's really -- it plays a really key role. So we design, in many cases, our own hardware. And so we design -- we -- I think when Cognex is at its best, we're deploying the world's best vision analysis software and applications within the world's best industrial hardware, embedded computing, smart vision systems. And so we do have our own in-house embedded systems design capabilities. We design our hardware very purposefully for the software that it will run. And our computing vendors tell us that we get more performance out of their chipsets than anyone else in the world. So we know how to optimize silicon for high-performance embedded computing, and that's a huge advantage for us, right, particularly in the world of AI where the best model is -- far surpasses the computing of the best available embedded chipsets, right? And so a lot of where our IP is today is taking the best frontier transformer AI model, distilling it, adding context, adding our secret sauce so that it can run on a relatively low performing, relatively inexpensive piece of embedded hardware. And we do that better than anyone else in the world. And that's really important because it means we can design systems that are smaller, more cost effective, can be put in more places, bridging that gap from large data center scale AI to near edge or far edge, I think, is what it's called embedded systems that can run those tools is fundamentally where a lot of our IP is created today.

Yes. Just bridging off the AI question that was all externally customer-facing. Maybe in the limited time we have left to kind of blend this in with margin objectives. How are you applying AI internally? And just speak to maybe the path you're on from a margin improvement point to the extent that can help you.

Yes, absolutely. So we're very focused on getting our margins -- our operating margins back into the financial framework that we put during our Investor Day in June, where we defined an operating margin range of 20% to 30% adjusted EBITDA. The last couple of years, we have been underperforming versus those metrics, and we're very committed through a mixture of cost and growth initiatives to get back into that target operating range. And so I'm very optimistic that we will in the near term to get back to the 20% or greater level. And then obviously, we're going to continue to manage our costs, drive some cost efficiency programs internally, get back to growth and get operating leverage, the mixture of those 2 things, bringing us to the mid-20s, high 20s and potentially beyond. Cognex over the last 10 years has an average adjusted EBITDA of 28%. The last couple of years has been significantly below that, but we feel very optimistic that we can get back not just to that target range of profitability, but to the high end of it through a mixture of focused cost management and cost reductions as well as getting back to the growth story of the company.

Yes. Very good. Perfect. We're at time, so we'll stop there. There is a breakout session in the Salon A room. If you go out to your left, we can take further questions there.

That' great, thank you.

Thank you, Matt.