Hey, guys, we ready to start? Great. Fantastic. All right. Good morning, everyone. I'm Sean Keresey. I'm a managing director on our investment banking side. I'm joined today by Dennis Fehr, Chief Financial Officer of Cognex. On behalf of the Morgan Stanley folks, welcome to our 14th Annual Laguna Conference. Cognex is one of the leading industrial companies in machine vision, an early innovator in AI, and a key element of the factory floor. You're coming off Q2, which was a record for you guys.
Thank you.
You upgraded, what was it? Four of the five key end markets for yourself, which is great to hear. Again, thanks for joining us. Maybe just to kick us off, give us a little about what Cognex does and where you guys are adding the industrial supply chain.
Sure, absolutely. First of all, thanks a lot for having me. Good morning, everyone. Thanks for being here early morning. Yes, as already introduced by Sean, my name is Dennis Fehr, Chief Financial Officer at Cognex. At Cognex, we solve some of the most challenging machine vision tasks in the industry. That means we are helping our customers across different verticals, we'll probably talk a little more about these, to inspect their either finished products or some of their components they use to identify, think about parts in the production or also in the logistics, warehouse automation, e-commerce type of space. We also help to gauge and to measure. Think about, and guide, think about robotic arm guidance, for example.
In that regard, we're playing really in a diverse set of end markets from semi, packaging, electronics, and of course, logistics, our largest end market.
That's great. Thanks for that. As you talk to customers now, you've had a few new product launches. What are the kind of problems that they're trying to solve, and what is Cognex enabling them to unlock in sort of those problem sets?
Yeah, no. Let me talk through maybe a few examples, on an industry-wide basis, but to some extent, many of them, you could also apply it to some of the other verticals. Perhaps starting on the packaging side. First, what do we say with packaging? That means fast-moving consumer goods and kind of pharmaceuticals. It's called packaging because very often you kind of inspect not the product itself, but basically where the product is packaged. Here, a lot is about resolving for throughput. That means the speed the line can run on, and then for quality, right? I think one of the biggest issues, think about a shampoo manufacturer and so on.
One of the biggest issue for them is if they deliver a product, and it sits on the shelf, and it just doesn't look like there's scratches, there are wrong labels or only partially labeled, things are opening, just a bad quality experience. In that regard, we are helping them, especially with speed and quality. Then, in logistics, in warehouse automation, we're really helping to resolve topics like labor shortages and then especially cost also. That means optimizing how much does it cost to ship a single parcel and to run that through an e-commerce distribution network. Then maybe lastly, on the semiconductor side, here, we're really helping our customers to expand capacity. So right now, that's very much a capacity build-out type of play.
I think the exciting thing on top of all of these underlying customer needs, we are layering in new AI-based vision tools, which really drive additional penetration. That means we are able to unlock applications for our customers, which in the past they could not solve with machine vision, and they needed either human beings for it, or they couldn't solve it at all. In that regard, definitely, a lot of kind of secular growth drivers here.
Okay, on the secular growth drivers, when we think about the semiconductor and electronics end market, what is kind of the mix that you are seeing? In particular, is that broader capacity additions? Are we doing supply chain relocation, new devices, form factors? What's the mix that's driving your business?
Right. On semi, I would really point towards the build-out of the AI infrastructure, right? I mean, semi is the one market where we point towards much less of a penetration play, but it's a capacity expansion play. Here, obviously, you see the news, you see all the investment going into that space, and then you read also about memory cost increases and so on, and we are basically helping our customers to ramp up their capacity. On the electronic side, it's a very different story. It's a much more broader-based kind of growth driver. I think a lot of investors ask, especially those who maybe know us for many years, maybe they think and ask us about, like, "Hey, is your electronics this year driven mainly by new form factors like foldable phones?" My answer continuously is, it's a factor, but it's not the factor, right?
In that regard, the electronics growth which we are seeing this year is very broad-based, first by customers. So it's not tied to one single customer, and it's not driven by one single kind of item, like a foldable phone, but really from, in general, kind of ramping up device quantities. It's about more devices. Outside of phones, so think about AI gadgets, wearables, glasses, for example. Then also here in electronics, we have a portion, still small today, but growing at interesting growth rates, is helping the data center supply chain. Before I talked about the chips itself and semi, but here it's about a full server rack assembly verification. In that regard, electronics is a very broad-based play at the moment.
Maybe this is a good spot to talk about some of the applications you're seeing in the data center. Could you give everyone a preview or an overview of what you're seeing there?
Right. I think what we really saw over the last, I would almost say six months or so, since we launched a new product called the In-Sight 3900 together with OneVision, which is our cloud-based AI training tool, that we are now able to solve tasks for these, let's call it contract manufacturers. At the end, these are the contract manufacturers for the hyperscalers who build the server racks for them. If you build a server rack, while it's itself, I would say it's not a very complicated production process. Obviously, there are a lot of things which need to be at the right places, a lot of connections at the right places, and certainly you do not want to find them when you start shipping these or have shipped these server racks into the data centers. You want to find that at the factory.
In that regard, we're really helping our customers to do a full server rack verification. That means we have one of our machine vision systems mounted on a robotic arm, going around of the server, taking pictures, and then verifying between 100- 200 inspection points. That's really enabled by what I said before, by new AI technology, what we call OneVision. This wouldn't have been possible six months ago. In that regard, we are very excited that we can now provide such kind of solutions to our customers. Therefore, we also feel very positive that we can drive additional growth from penetration in the data center supply chain.
Maybe just to round out the discussion of end markets, packaging, automotive, and other, how do you see those contributing to the mix going forward?
Yeah, I would really want to highlight the packaging. I already talked a bit about it. It's really a market where we have very purposefully diversified in over the last couple of years. If you would go back, let's go back 10 years, maybe 2017 or so, you would not find packaging even as mentioned by Cognex as a major market. In 2025, it was already our second-largest end market. Here it's really all about driving penetration. In that regard, it's about finding new customers. We did a lot of push over the last three years or so in growing our customer base, and we were very successful here on the packaging side. Then it's also bringing some of the latest AI applications to our customers to help them solve tasks at high speed.
I mentioned before, packaging is a high-speed kind of requirement market. In that regard, very excited about this market because by itself, it's not a very cyclical market. Electronics typically is a rather cyclical market, but packaging is not. In that regard, we have been really looking forward to find ways to diversify our growth from highly cyclical or volatile markets into other markets which provide more stable growth, and packaging is one of these success examples we can point to.
We hit on OneVision, and I recall having the opportunity to use OneVision--
Right.
--at the Investor Day a couple of years ago. Why is OneVision such a game changer for you, and how does it open up different applications for your customers and keep you differentiated relative to your peers?
Right. Before I talk about OneVision, let me say a few words generally about our AI journey and bringing AI into our products. We are almost, you could say, almost 10 years into our AI journey. We have been really pioneering this technology with an acquisition late 2017, early 2018, where we bought a company called ViDi in Switzerland, and that really formed the nucleus of our today's AI vision team and brought our first AI-enabled product in 2022, which essentially is what we call training on-device. That means our customers really want to solve vision tasks on the device for two reasons. One is the latency, the speed topic, and then the other one is cybersecurity concerns. You could almost argue nowadays, maybe a third item comes onto it is token cost, right?
If you solve things on the device, you are not having token cost. Whereas if you do it in the cloud, you also get a variable cost into your processing, versus if you solve it on the device, you pay upfront for the device and you do not have variable cost. In that regard, that is really where we started, where customers really wanted it the most. What does it really solve for the customer? It solves two things. A, it is much easier for them to configure the system, and two, it enables to solve tasks which were much harder to do it without AI, let us say, or just impossible to do. Now, where does OneVision come into play is that if you do things on the device, what is the preferred aspect for our customers, you are limited by the compute power of the device.
Now, of course, compute power gets more and more powerful on devices, but still it is limited. In that regard, if you want to do the hardest-to-solve tasks, you need to be in the cloud. But somehow, again, customers do not really want to be in the cloud, especially not when they run it live on their production line. That is why we created OneVision, which essentially is a training in the cloud. That means we only let our customers or help them to train their machine vision models in the cloud based on our proprietary machine vision models, which we provide as a pre-trained starting point. Once they have completed the training, they just bring back the fully trained model onto the device.
They are not in the cloud anymore, and they can solve very hard applications like the full rack server verification, or if you think about inspecting transparent surfaces, glass bottles, or transparent plastic wraps in the packaging industry, especially an important factor. In that regard, we are really providing this bridge for the customers to use the power of the cloud, but to avoid the pitfalls of the cloud. That is kind of very unique in the industry. I think we can really say that nobody else is able to provide such an offering today, and that is why we are so excited about OneVision.
Excellent. That is great. On to China, which is a fast-growing region for machine vision broadly, and I think a lot of that answer probably plays into the competitive advantage there. How do you assess Cognex's positioning overall within that market and how do you kind of measure yourself versus some of the alternatives that are coming out of that region?
Great. I would say we really have been able to shift from defense into offense over the last 18 months. If you go back to 2024, we clearly saw that Chinese local competitors starting to trying to catch up, and at that time we basically countered that and we lowered prices on our, what we call N-2 type of products. So older generation products where kind of they play in in terms of technology level. I would say that was a defensive move at that time to defend market share and that was successful to some extent at that time. But since then, we have really kind of made some key investments into the China market from the way how we structured our sales organization. We invested into China for China products. We invested into strengthening our local distribution network.
Now we can really say, at least our perception is that this year we are growing faster than the market. We have more than 40% growth so far in China this year. That points to that we are playing the offense, that we are able to take share there. In general, we feel like it is an exciting market and from that regard, feel very positive, kind of how it has developed. Of course, the market in general has become easier. That certainly helps. But at the same time, we are now able to do both. We are able to grow and we are also able to take price. That I think, really points to quite some strength.
Very good. Maybe I will just pause for a second. Any questions from the room? I will circle back if there is none at this point, but I just want to make sure everyone gets a chance to ask any questions. Acknowledging it is a little bit early. Well, speaking of offense, you have added 9,000 customers, in 2025 and another 4,500 in 2026. Maybe just talk a little bit about the customer acquisition, the model, how you are changing going to market, especially with your new tools like OneVision.
Sure. I think one of our top three strategic objectives, which we announced last year at our Investor Day, was to say we want to double our customer base. Maybe I first talk a bit about the rationale behind it, then I talk a bit more about where are we on this journey and where we will go next.
Great.
I think I already kind of alluded a bit to it. If you look back into Cognex's history, in general, I think we have seen on the one side, over 10- 15 years, all the way up to 2021, kind of 15% CAGR, but not anywhere close to a straight line growth. Actually, very volatile growth. That means very strong growth in peak years, but then also contractions in down years. Some of that was clearly also driven by high customer concentration and then by high exposure to more cyclical or volatile end markets. In that regard, on the one side, obviously we love to be strong with what we call our tier one accounts, and we certainly still love to be in the electronic space.
But very clearly, from a management perspective, it is also to say, is there a path to diversify the growth engine to drive more consistent growth and less volatile growth? That is kind of behind the thinking of that expanding the customer base. That means there is a diversification of customer play into here, and it is not the only angle which we are driving to drive diversification, but that was the first one which we really tackled the strongest. Here, I think, we feel like we have been quite successful in what we would call the land and expand strategy. That means adding 9,000 customers last year, that was 3x of what we have been able to add in 2024, adding 4,500 customers in the first half of the year.
That means this first step, I think we feel like we have built that engine, and core of that was to reorganize our sales organization. That means first bringing all sales organization into one team, but then have specific profiles. That means we have one seller profile which is really looking for this new type of customers, and then we have a different seller profile which is really focused on existing established end users, and the third one is for machine builders. That means the sales force structure as it is in place has been really playing into that, and we are very happy with that. But now really the next step is about the expand. That is kind of where our mind more and more shifts to.
So in that regard, would we think about like, is the next strategic objective to double the customer base again? Probably not. It is really now about expanding on these customers which we have been winning already. Here now other things come into mind. First of all, serving them with the right technology. Our vision is an important piece to that, but it is going beyond that. It is also, what do customers want? They certainly also want to have the right level of service, but that is where our other strategic objective of being the number one in customer experience plays a role. Then they also want to have, at points, full solutions, and they do not want to just buy products. We see ourself really as a product technology company, and we are not looking to provide integration services to our customers.
But it is also clearly there are many other companies in the market, kind of system integrators, who can provide that service. That is an area where we see still an opportunity for us where we can refine our go-to- market. That means we are right now building out a new, call it, entitlement/loyalty/support program for system integrators and also machine builders, so that we provide the best possible support for our customers there. By that, basically helping to go from land to expand. That is where I would say the go-to- market and maybe the overall strategic objective is evolving into, step by step, as we are making progress.
The profile of your customers within, let us just maybe pick a few out of these 4,500. They are smaller, but the trajectory and their scope of end market applications or applications in general for what they do is pretty broad down the road. Is that kind of a fair way to think about the strategy?
Yeah. Maybe I go back to the example of the packaging industry.
Okay.
For many, many years, we have been playing very successful with some of the largest names in the space. Think about companies like P&G, Unilever, and so on. Just examples. But this market is very regionally diversified, in the sense that you have a lot of smaller businesses all the way down to mom-and-pop shops, from companies who are producing soups for local fast food chains or convenience stores, and they maybe have 20- 30 employees. That's kind of what we have not reached in the past. That's really where we can drive additional growth and where we can drive additional customer penetration into. From that regard, that's really where we have been able to penetrate.
I pointed to that growth and this kind of much more stable growth in that area, and that's kind of how the strategy unfolds in that area.
That's helpful. I'll pause there on end markets and model. Any questions from audience? Okay. A lot happening on the financial model as well. You've evolved meaningfully in generating stronger margins. Very different Cognex from a year or even a year or two ago. Cash generation and operating leverage is also kicking in. As you look beyond the current margin, and the current cost-out program, how should you have investors think about the margin story evolving?
Yeah. First of all, I would say we are very pleased with what we have achieved over the last 18- 24 months. If you think back 2024, 17% Adjusted EBITDA margin, I would say, really very far away from what Cognex has done historically and could do. Certainly, it was very clearly that when Matt and I came into the roles, that that's really the area where we have to work on. The biggest issue which we saw was really that our OpEx efficiency was pretty low at that time.
I think the success of bringing up the bottom line margins from that 17% to our guidance midpoint for this year, 30%. That means a significant expansion of the margin, the bottom line, in a comparatively short time, w as first of all, was really going hard after OpEx efficiency and very clearly was about running cost-out programs. But I think at the same time, having in mind that we did not want to sacrifice on growth. So in that regard, it was really all about thinking about what's the right productivity level throughout the organization.
As we started to see accelerating top-line growth this year, combined with taking out cost, we really see very strong leverage. We have in Q2, we had 100% revenue fall through to the bottom line. I would say really quite beyond our expectations, especially in time of the speed how we got there. But now, are we at the point where we think we cannot go further? No, absolutely not.
I think if we look at our operations of how we run the business, I think we have now been 18 months into transforming our operating model, and we made quite some progress. But clearly, there's another level we can go to, right? That means we are not at the end of the journey in terms of driving efficiency into our processes. But it will show differently in the P&L, right? That means, now especially in the second half of this year, we're really showing OpEx in absolute terms reducing, and that's not what I would expect for 2027 going forward. But it's really all about being able to drive further growth without meaningfully expand the OpEx line. When I say meaningfully, I would say maybe in line with inflation.
That means think about like 2027 inflationary growth on the OpEx side while still outgrowing on the top line and by that, still driving fall-through into the bottom line. Then certainly our eyes was also on the gross margin line, where we have seen throughout the year quite some positive effects from the mix, as really some of the highly attractive end markets for us in terms of margin performance have been leading the growth. Pricing is still a lever where we can work on. I think an area where we haven't been too active in the past was Cognex, and we think as many other industrial companies, we can use pricing as a compounding lever over time.
So it's not like big numbers every year, but over three or four years, I think we can really turn pricing also in a meaningful kind of compounding effect into the gross margin. So all in all, I would say we are clearly not at the end of our margin expansion story, but also very clearly is that we see this as profitable growth. We're not seeing it profit or growth. We're really seeing it as both together and as management team, we are very clearly focused to make sure that we are getting both, and that we are not sacrificing one for the other.
With the shift, or kind of I guess would be better said as an expansion of your go-to-market model, how do you guys evaluate sales productivity in that mix?
Yeah, it's a great question, and that's where we really spend a lot of time on it, and we created a lot of data analytics, and we keep on challenging ourself. I mentioned before about three different seller profiles. So the team, which is looking for, or the seller profiles, which is looking to win new customers versus those serving end users versus those serving machine builders. And they have very different kind of objectives in terms of the productivity. We look at it very simple when we say sales productivity, it's at the end, it's like, okay, what is bookings versus dollar expense in a dollar number? But very clearly, we have a different expectation from somebody who's hunting for new customers than somebody who's serving established customers or is trying to penetrate existing customers deeper.
In that regard, we have very clearly articulated productivity targets for each of these seller profiles. For this year, we have a clear productivity improvement objective for each of these, and we can say we're actually ahead of these productivity targets. And we think that next year there's another productivity improvement and the year after. In that regard, again, to this point, do we need to drive OpEx growth as we drive top-line growth? Not very much. Of course, again, inflationary adjustments may be needed, but not beyond that because we really have a lot of kind of productivity growth. And there's clearly things, as I mentioned before, like how we work with the system integrators. That's an area where we support marketing automation is a big piece for us, where I would say we probably still have more work to do.
I think we run a couple of pilots where we said, "Hey, here's how we think this could shape out." We had some very interesting learnings from this, and now basically going back to the drawing board and putting these learnings into an updated concept. And I believe marketing automation is an area where over the next maybe two years or so, we probably can also drive further efficiency gains. And the great thing is it's an efficiency gain on OpEx and also drives growth. So that's really coming back to the profitable growth story.
You kind of mentioned the inflationary expectation within, it was something kind of like bill of materials.
Yes.
Are there any constraints to kind of the growth or, I guess said differently, the more durable growth that you expect for the next several years?
I think the biggest area where we are looking at right now is memory components. To some extent, we're also starting to look into the broader, call it maybe more passive type of electronics components. But memory is really the area where we see the most cost inflation happening. I think it's probably not a new theme for anyone, but really very high inflationary pressure there. While memory is not a big component of our bill of material, we see it as a headwind. We said a 75 basis points headwind to gross margin in Q3. At the same time, we are also able to drive really pricing on the other side and to pass on this increased cost to our customers. We think these effects as transitionary, I would call it.
That means like, yeah, there will be lasting perhaps in the bill of material cost, but we'll be able to offset it through pricing. So far, I would say we're not seeing kind of a shortage, but we are seeing tight supply. Memory is tight on supply and some other areas as well. So far we have been able to manage quite successfully. We would not see that this would kind of restrict our growth in the future.
Nothing that impacts kind of the sales timeline that you generally expect?
No.
Maybe on that topic, what's changed with your sales timeline over, call it the last several years as you kind of shift how you go to market a little bit, you're targeting more of the mom and pops as opposed to kind of the larger customers that required a lot of engineering support. Has that compressed meaningfully and how do you kind of see that evolving?
I would say it has not changed so much. At the end, I would say the sales cycle, whether you go to a smaller customer or larger customer, it's more driven by the application than what customer it is. If you go into a more complex application, sales cycle tend to go longer, versus easier application, a bit shorter. The smaller companies, they more tend towards the more easier-to-use applications, but not always. So in that regard, clearly it's not limit to. Now I think I'll leave it there.
Fair enough. Any questions from the room? All right. Capital allocation. You guys spent a lot of time with this at your Investor Day. You continue to work towards really strong margins, strong free cash flow generation. How do you guys think about, call it return of capital, balancing that with M&A, and obviously organic investments?
I would say over the last 18 months, we have been very strongly focused on share buybacks. We returned close to $250 million or even a bit ahead of that, just through share buybacks without the dividends over the last 18 months. I think we bought at attractive share prices in the high 30s and low 40s. In that regard, I felt like we got a very good return on that capital allocation. Certainly share buybacks plays also a role into the future, but opportunistically. That means when we see a pullback in the market, that's the moment in time for us to deploy capital through share buybacks.
Then also very clearly, we said at Investor Day that want to basically have more or less an equal split between share buybacks and M&A, and we want to make M&A as part of the growth story, as part of the diversification of the growth. We haven't been active in M&A over the last 18 months or so. I would say right now, we probably feel like we are probably in a very good position and good time to do M&A. I think we have been able to return to growth. We have been able to drive significant margin expansion. We are in a much better shape there. We have been making a lot of progress on transforming our operating model. I would say probably timing for M&As couldn't be better right now than during the last 18 months.
Then, of course, it always comes back to do you find the right target for the right price? But very clearly, I would say if we find the right target for the right price, and especially to help us to diversify our growth engine and to go into adjacent markets, then we would definitely deploy capital there and very much looking forward to that day.
Yeah. It's a strong market out there, but again, finding the right adjacency is probably the key question, right?
Right.
Look, any questions from the audience? We probably have time for one. Okay, well, that brings us just about to the allotted time. Dennis, thank you so much for joining us at the Laguna Conference on behalf of Morgan Stanley. We appreciate everyone attending this morning, and enjoy the rest of the conference. Thank you.
Thanks a lot, Sean. Thanks a lot, everyone, for attending.