Hi, everybody. Welcome. Good morning, good afternoon, depending on where you are. Welcome to the Rosenblatt Age of AI conference, to our June session. I am Mike Genovese, the Cloud and Communications Equipment Infrastructure Analyst. Super happy that you're all joining us today, and very pleased to have the management team from Fabrinet. We have Seamus, we have Csaba, and we have Garo, CEO, CFO, and Head of Investor Relations. Great to see you guys. Welcome.
Thanks, Mike.
For those of you in the audience who want to ask questions, the best way to do it is through the widget on your Zoom interface there. If you type in the questions, they'll come to me, and I will make sure that in the next 45 minutes, we ask the question to these gentlemen. Let's just start at a high level, Seamus, just kind of the standard, where you see the company right now, the strategy, the positioning, the key opportunities. For those on the call who may be a little bit newer to Fabrinet, just kind of frame the company at this point in time for us, please.
Sure, Mike. Hi, Mike. And thanks for having us. We appreciate it. Yeah, so Fabrinet is a precision optoelectronics contract manufacturer. We focus on the optoelectronics space, primarily optical communications, which accounts for about +70% of our business, a little bit more. That is split roughly between data com and telecom. The data com business has been growing quite well over the last couple of years, but our telecom business is also back to seeing some strength this past while. Our primary manufacturing base is in Thailand. We have two campuses, Pathum Thani and Chonburi. We also have a couple of on-ramp locations, our new product introduction facilities, one in Santa Clara here in California and another one in Israel. We see ourselves as the leading contract manufacturer for the optical communications space. About 50% of the outsourced optical communications business is with us.
The other 50% is split between a number of companies. We continue to focus on that industry and similar related industries. We may be a little bit different to some of our contract manufacturing peers in that we do not have any of our own products. We are a pure-play contract manufacturer, and we plan to stay that way. We do not have any of our own products, and we will not be. We are a service company. We provide a manufacturing service. We have been growing nicely over the last number of years. If you look at the last 10 years, our revenue has grown, compound annual growth rate of 16%, and we have compounded our earnings 21% over that time period. Growing faster than the industries that we serve, but also faster than the contract manufacturing industry.
We're in Q4 of our fiscal 2025, and at the midpoint of our guidance, if it were to come to pass, we'd be up about, I think, about 18% versus FY2024. And then we're positioned well for the year ahead with a number of growth factors in our favor as we enter into a new fiscal year.
Okay, perfect. Yeah, kind of keeping it at a high level and sort of what makes the company unique, it seems like you've always, compared to contract manufacturer peers, right? You've specialized in optical, right? There's this optical sort of subsystems and component specialty. And that's, I think, I'd like to talk more about this, but sort of allowed you to have higher margins, higher gross margins, higher operating margins than your peer group looking at contract manufacturers. Now we see maybe more advanced packaging wins, more systems wins. Kind of frame that for us, how your core competency in optical, how that plays into other sort of adjacencies, how maybe large of opportunities those are, and if those support sort of the exact same margin structure as the optical business.
Yeah, our margin has tended to be a little bit higher than the industry peers. Our gross margin has been in the kind of 12-13% range. Our operating, our OpEx is less than 2%. It's about 1.8%. Our operating margin tends to be in the 10-11% range, which is a bit higher than our competitors. Not because our prices are higher, it's because our costs are lower. We have a very compact footprint. We have a very low-cost footprint. We take a compelling combination of very high capabilities and unique capabilities coupled with low costs. We focus very much on making sure we keep our costs under control. We keep our costs low. Again, our footprint is quite compact. We have been expanding over the last few years, if you like, vertically.
We like to do more for our customers than just, if you look at essentially the start of the company, we were an optical component manufacturer. Over the last several years, we've expanded the offering down deeper into the stack by doing more packaging. We're able to do a lot of the precision packaging for our customers and also up the stack into the complete network systems or subsystems. We've been able to do that while maintaining and actually improving our margins because we really index with our customers on stickiness and doing more for them and making sure we enable them to get their leading-edge products to market faster than they would if they were using somebody else or maybe a number of suppliers. We're able to do that while maintaining and improving the margin.
We're pretty excited about the opportunities as well that that brings us, that there's not too many contract manufacturers who can do what we do. Roughly 70% of our manufacturing space is clean room space. That'll give you an indication of how much of the precision packaging we do. We are quite unique in terms of the capabilities that we have, which, like I say, has allowed us to expand deeper into the stack, but also vertically to do more system build, which has allowed us to grow nicely while maintaining the margins.
Yeah. I mean, I don't think I've ever really heard you talk about sort of how big, I mean, if we look at the optical communications piece, right, then we'll leave the auto and the other non-optical stuff separate. How much can systems or an advanced packaging, sort of what percentage of the business of the communications piece that could become? Have you thought about that?
We have thought about it. We haven't talked about it publicly. We don't break it out that way. We break it out by kind of industry segment, if you like, or by end markets that we serve. We haven't broken it out that way. A lot of the packaging that we do, we don't really offer packaging as a standalone service. You won't see, we don't have a third-party revenue line in our reported numbers for packaging. We do the packaging for our own use and our own consumption. A lot of those services that we provide, we're providing them for our own use and our own consumption. Like I said, there is no third-party sales. It is a significant driver of growth, and we call it stickiness with the customers when we're doing more for them. They are significant drivers. I would say the packaging is a driver of stickiness, and then the system business is a driver of growth. The combination of the two is what allows us to maintain our margins.
Great. I kind of asked this question, this next question, just to sort of get it out of the way, unless there's something we really need to kind of dive into and keep to frame the discussion. Just the tariff and the macro concerns, right? I mean, it seems like that hasn't been such a big issue on last quarter's earnings or the guide for this quarter or what we think about the quarter after, or certainly any macro impact there. Just tell us what you're seeing from a tariff and a macro perspective, and if that's something we can kind of just move past or something we really have to think about in the story here.
I think probably like a lot of companies in our position, we're just kind of waiting to see what happens in the future. The future might be tomorrow. It might be a month from now. Who knows? I think the unpredictability is certainly not helping anybody. The macro, because the vast bulk of the products we make for our customers, if not all of the products we make for our customers, are for infrastructure-type products. They're not for consumer-facing products. We tend to be somewhat insulated from consumer sentiment and things like that. A lot of the products we're making, they're for capital projects that are rolling out over several years, and they tend to be quite stable and maybe less volatile and less susceptible to consumer sentiment or overall nervousness around tariffs.
The products we make for our customers, the terms we have with our customers are FOB, our factory. The customer takes possession of the product at our back dock, and then the customer is responsible for the tariffs. We have not seen any particular impact other than maybe a little bit of, I think, from everybody, maybe confusion because nobody really knows what is going on. The tariff situation seems to be quite, we call it, changeable. I think until things settle down and we figure out what the long-term tariff position is, I think nobody really knows what the impact will be.
Okay. Great. I'm getting questions coming in from the audience. We already have kind of a handful of questions here, and I want to incorporate them into my own questions. One topic is that if we kind of move into the data com side of the business now, I think you've been talking about converting capacity from 800GG to 1.6T for your main customer and kind of implying that if we're going to see a re-acceleration of growth sequentially here, it's going to be, I think, more tied to 1.6T . Can you talk about the timing of that? I guess if I fold in this question from the audience, it's sort of saying, "Hey, 800GG was that cycle kind of lasted longer. It was stronger. 1.6T got pushed out, but now we're shifting capacity to 1.6T . What are the data com expectations for the second half of this calendar year? Should we see that re-acceleration tied to 1.6T ? What else can you tell us about that transition?
Yeah, I mean, the 1.6T product that we're making for our customer, it's 200 per lane. Maybe more precisely, we should be saying we've been converting from 100 per lane to 200 per lane. We've been converting our capacity. At 200 per lane, you can have 1.6T , but you can also have 800G . We've been converting capacity as we talked about. The driver for 1.6T certainly would seem to be the new product that the customer is launching. The timing of that, it's not really our place to say. We'll follow the customer's lead, and we would expect to start shipping 1.6T probably one quarter ahead of when the customer is launching the product. There's still plenty of demand, I would say, at 800G . We could see an acceleration. I think for us, the real growth in 1.6T would be tied to the new product that the customer is launching.
Okay. Just remind me, for the June quarter guide, did you say data? I can't remember if you broke out telecom and data com in the guide and said it would be sort of sequentially flat, sequentially up. What did you say about data com for June?
We said that all of our segments will be up sequentially. Data com will be up as well at midpoint of the guidance along with other segments.
I assume we'll see there'll be some 1.6T in that number. I mean, even if it's very small, there'd be something in that number for 1.6T . Is that true?
Yeah, that's correct.
Yeah. Okay. And when you talk about 100 to 200 transition, is that—I do not know if this is 100% overlap here, but I mean, is that kind of synonymous with sort of going from a VCSEL-based transceivers to EML-based transceivers? Is that what we are seeing in this 100 to 200 transition?
Yes. Yes.
Okay.
Going from VCSEL-based to EML-based.
Right. And then, as you said in the past, the product that's—so I think if we talk about Blackwell, right, 200, that's an 800GG product still, but 300 Blackwell Ultra, that's where we get to 1.6T ?
There'll be 1.6T and 800G for Blackwell Ultra. The switch level is 800G , whereas it was 400G for the previous product.
Yeah. Now, I mean, if we think the last several quarters for your 800GG sales, I think we've been somewhere roughly between sort of $250 million and $270 million in quarterly revenues for 800GG product. Do we kind of think that that's—okay, so billion-dollar run rate, that's kind of the business? Or does the 1.6T , I mean, even if you're kind of lead on the competitors, not as long maybe as it was in the 800GG, and the pricing, right, is not—it's not going to be double the price or even 1.5 the price, but it'll be something probably below 1.5. But with all that factored in, I mean, should we get to new peaks for the business with 1.6T , or are we on our way to a $2 billion data com business, or are we kind of going to be a $1 billion data com business?
We guide one quarter at a time. I think the number you mentioned, we're mixing apples and oranges a little bit because the 270 number is our overall—and Csaba can correct me if I'm wrong here, but that's our overall 800G and above business, which is a combination of the products we make for our main AI customer. But also, there's a lot of other products in there, and there's some telecom products in there as well. So it's not quite a clear-cut. That's not our shipment number, let's say, to NVIDIA. I think we could get to new highs when 1.6T launches. But at the same token, in the early days of 800G , at the very start, the NVIDIA-designed transceiver was the only approved transceiver for the whole network. Eventually, that changed when other merchant transceiver manufacturers got approved.
We were the only—let's say the transceiver that we made for our customer was the only one that was approved for a while. I don't think we'll have that luxury when 1.6T launches. I would assume that the merchant manufacturers will get approved quite quickly. I think the length of time that we have exclusivity, we call it, will be much shorter, if at all.
On the pricing, I mean, is it like a 1.3, 1.4, 1.2? I mean, how do you think about in terms of the multiple on the 800GG price?
We haven't really disclosed that, but it will be closer to the 800G price than you would imagine. It also is a function of how much content we'll be producing in-house. Generally, we find if there are components that we're purchasing from third parties, if we can produce those components in-house, we can do it at a much lower cost and a higher yield than third-party suppliers and save the customer a lot of money. We will be working very hard to make sure we're ready to ramp when the customer is ready, but also that we have as much fabric content as possible. That's the best way to get the price down, but also to improve our margin and improve the stickiness. There is a lot to be worked on still.
Yeah. Early, you made the distinction between yourself and some of the optical component players in the transceiver space, the Lumentums, Coherent, Innolights, which do design products for customers. Just from a decision-making process for the hyperscalers—so NVIDIA, we've got NVIDIA, and I guess there are other GPU players out there as well that could decide to design their own transceivers. There are hyperscalers that could decide to design their own transceivers. As somebody who does what you do and you're watching kind of the—I do not know if it's buy versus build or design versus outsource decision for transceivers for these large, large customers. Do you think in the future more are going to design their own? I mean, are things moving in that direction, or do we know yet? What are you looking at?
Yeah, we think so. We think it's inevitable. If you look at the margin elimination, if you like, that these very large customers are able to achieve, it's a pretty compelling—the economics of it are pretty compelling. If they can develop their own transceiver and have Fabrinet make it, our margins are much more modest than product company margins. We don't have the same expenses or cost structure that the product companies have, so we don't need the same gross margins. We are able to produce at a much lower cost. If the hyperscalers are able to get the product designed at a cost that's affordable for them, and we are able to produce it, they can save a lot of money. It is really as the volumes become very significant, which they are now for 400G , for 800G , these short-reach, low-power, low-latency transceivers, the volumes are staggering. If they're able to develop their own transceiver and have Fabrinet make it, they can save themselves a lot of money. I think it's pretty inevitable.
Yeah. Okay. As investors, we're still kind of not super clear on what you're doing with Amazon where you announced working with them that would be tied to warrants. I think you said on the last conference call that kind of going forward, your Amazon revenues would hit a number of your different revenue categories. I think there's some in data com, some in telecom, some in non-optical. It seems like you're engaged with Amazon on multiple sort of fronts. In terms of tying it into what we just talked about, sort of designing their own transceivers, is there any of that going on? What can you tell us about this Amazon win?
Mike, as I'm sure you can appreciate, it's not really our place to disclose what Amazon is doing and what products they're working on or designing. We're very happy with the relationship and with the range of services we'll be providing to Amazon. We believe we will be shipping revenue in a number of categories. We're currently shipping revenue that's in our telecom category. We will have high-performance compute as a category. We'll be introducing that next fiscal year. In our Q1, we'll be showing that as a separate category. Of course, data com, we'll be producing hopefully in the future, we'll be producing products in the data com space as well. All three, and they're just an amazing company, really. When they put their minds to something, they really know how to go about achieving things that are seemingly impossible. We're very much looking forward to that relationship expanding and becoming a significant relationship for Fabrinet.
Now, so it sounds like, I mean, that's a decision to report segmentation with the high-performance computing category in Q1. I mean, does that imply that the level of revenues for this new category are inflecting in Q1 and are higher, or that's just a new year? We're going to break this out.
It's a practical matter, really. First of all, for the first quarter or two, we're really just shipping qualification volumes. So it's included in our other category. As it starts to ramp, as the product starts to ramp, it doesn't belong in any of the other categories. It's obviously not automotive. It's not laser. It's not telecom, and it's not data com. It kind of requires us to create a new category. The logical point to start disclosing that category is when we report our Q1. We'll be reporting that in our Q1 results. If it was going to stay at the low level that would allow us to keep it under other, we would probably do that. It's hopefully going to be too big to remain in the other category. We decided to report it separately.
Yeah. I mean, I'm used to all of your product. I mean, it seems like unless there's a 4x movements or something like that, we've kind of very consistently have these near 13% gross margins, 11% operating margins. I'm not used to seeing sort of business mix making a big difference on margin. Since you've got this whole new revenue category, I think I'll ask on the Amazon if there's any kind of difference in the margin profile of that business versus your other businesses.
We work very hard to maintain our margins. It's not easy. It's also a function of what we're doing and how much value add we're providing for the customer. We'll be working hard to maintain our margins as we continue to deliver outsized growth. That's our job. We have to make sure we do that.
Okay. And you're very good at it. All right. It is a kind of roadmap for the rest of the conversation. Certainly, just for the people in the audience who are asking questions that we are going to get to, we are certainly going to get to the telecom side of the business and talking through some of these systems wins with Cisco and Ciena and Nokia. I just want to finish up the data center conversation maybe by touching on Co-packaged Optics because since OFC, you have been announced by NVIDIA as being part of the Co-packaged Optics solution that they are doing. I am wondering if there are any other customers that you are working with. Can you kind of tell us more about what you are doing with them and o f course, the question for Co-packaged Optics, is this sort of all additive and incremental, or is there any kind of cannibalistic aspect to what you're already doing? Maybe if we can touch on the margins for Co-packaged Optics as well. There's a lot in there. Let's start on touch on CPO.
Yeah. So we're working with three customers on CPO products, of which NVIDIA is one. NVIDIA is the one that the customer disclosed at GTC. I guess we're okay to talk about that one because the customer already disclosed it. The other two that we're working on, the customers have not disclosed those projects. We're not in a position to name the customers or talk about the specifics. Other than to say, our main focus for Co-packaged Optics is helping the customers figure out, okay, how do we produce these products in volume at an acceptable yield and an acceptable cost? That's really where we bring value to the customer. We have a couple of projects. The margin profile, probably a little bit higher than our regular business because the material content would be lower.
A lot of the work that we're doing is precision packaging where we don't own the wafer. So the material content is lower. Therefore, the margin should be higher. Is it additive? So far for us, what we've seen is it's additive. We're not at the point yet where Co-packaged Optics is just about to replace transceivers or anything like that. The opportunities for us are more at the switching level, which would be new for us. It is additive for us. We think it presents a number of new opportunities that we'll be focusing on.
When does it become kind of meaningful, or is it already in the revenues?
I think it's probably a little bit away before it becomes meaningful. The work that's going on right now is kind of developmental in nature with a couple of our customers. It's probably a little bit further out before it becomes meaningful.
Okay. That's great. Very helpful. Like I said, I mean, now we'll get to the telecom. Obviously, it's been very notable when the telecom business took a downturn a couple of years ago. The data com business really took off. When data com has taken a bit of a sequential pause in the last couple of few quarters, telecom's really driven a lot of growth. I think part of that, right? I mean, some of it seems to be a return of kind of the bread-and-butter component business, but then on top of that, there's systems wins. I believe with, at a minimum, Cisco, Nokia, Infinera, and Ciena. Not sure if there's any more. Yeah, talk about kind of help us dimensionalize telecom in terms of the base business coming back and then adding in these wins. Any wins that have not come into revenue yet, but that are ahead of us, kind of call that out. If you could talk about the customers individually as much as you are allowed to, that would be helpful.
Sure. So yeah, we do have this countercyclicality in our business when, like you say, when one part is down, another part seems to be up. Some of it is good luck. Some of it is targeted. We try to target customer wins and products so that we will have this countercyclicality effect because it is inevitable that one area of the business is going to be softer than the rest. We have benefited from that over the last few years. Like you say, when telecom was soft, data com was very strong and vice versa at the moment. Our telecom revenue, the high point for us was in about seven Quarters ago, $405 million of telecom revenue. That declined because of inventory digestion. We did not lose any customers.
We did not lose any business, but because of inventory digestion, that dropped all the way down to, I think, 285- 290, something like that at one point. Last quarter, we were back up at $406 million. On the face, you could say we were at $405 million seven Quarters ago. We are back to $406 million. Therefore, we are back to where we were, but not really. The tail of the tape underneath the headline number is the baseline business is back to growth, but not back to anywhere near the same level as it was. It is back to growth, but at a lower level. What has brought us, what has brought our overall revenue back up above the high watermark from the past is really two things. One is new system wins. You have mentioned a few of them. Secondly, DCI, more specifically, 400ZR.
Very nice growth in 400ZR, which now represents about 10% of our overall revenue. Actually, it was at one point 10% of our optical revenue. Now it is about 10% of our total revenue and growing nicely. As these super clusters get rolled out around the world, they need DCI to connect them together. We manufacture for a number of companies, several of whom are seeing nice steady growth in ZR, 400ZR, ZR Plus, and more recently, 800ZR is beginning to take hold as well. The system wins, you have mentioned a few of them. The Cisco wins that we have had, we have had a number of successes, I would say, with Cisco with taking share away from competitors. That is going well for us. We are very happy with that relationship. The customers are very happy with our performance.
We continue to do a good job. The best way to grow the business is to do an excellent job on the business that you have. That is really our focus. It is just on executing very well for our customer. Ciena as well. We have won some new business there recently with a newer product. That is starting to ramp and probably will ramp as we go throughout FY2026. That business will ramp. Nokia, Infinera, it is really early days with Nokia yet. Obviously, Infinera was a customer for us and continues to be a customer. We just have an excellent relationship with Infinera. We will be looking to capitalize on that with the Nokia team. It is early days yet in that Nokia relationship. There are a number of others as well that we are working on. Either too early to talk about or sometimes the customer does not like us to talk about these things. We are working very hard on all of those.
Yeah. I mean, with your expectation, this quarter for telecom will be up and kind of the trends in the market and the expansions. I mean, it sounds like we're certainly headed to new peaks on the telecom revenue, right? I mean, we don't know how big it can get, but it's already past peak and we're growing. So we're going to find out.
Yeah. We think it's going to continue to grow. We have some, like I said, the baseline business, the traditional, a lot of it is non-speed rated component business. That's back to very nice growth, very real strength there. The system wins. We continue to work very hard on those. When they happen, they're not gradual. They happen as a big chunk usually. DCI, especially ZR, is growing very nicely. Some really nice growth factors there for us in the telecom space.
Yeah. Okay. Great. I want to work through some of the questions that have come in from the audience. Some are very strategic and high level. Some are very focused. I'm just going to run through them and kind of bounce back and forth between high level and very focused questions. Starting with a focused one, how much do you expect your tax rate to increase in light of OECD pillar two in fiscal 2026, which I guess I do not really know about? What is that?
We have been operating at a mid-single digit tax rate. Our major footprint is in Thailand. We had pretty nice tax benefits over the last several years. As we continue to invest in Thailand, we continue to enjoy these benefits. We are optimistic that we will be able to maintain our tax rate going forward. We do not anticipate a significant increase at this stage.
Okay. Great. All right. Next question. I think, I mean, we went through this. I'm just going to ask one more time because we just had a long answer going through this. Just all the systems wins, Cisco, Ciena, Amazon, any others disclosed or not disclosed? I'm not sure what else there is to say about Cisco, Ciena, Amazon, but if you think of anything to add, we also touched on Nokia briefly. Anything else in there? Or does that pretty much cover the systems wins?
They cover the ones we've talked about. There's a number of others that, again, we're working very hard to win, but too early to talk about yet. They're the main ones that we've talked about. Did we miss anything, Csaba? I think.
We covered all of them. That's what we were able to announce.
Yeah.
Okay. How about inventory levels across the space? Inventory days are higher than they were in pre-COVID. Is this a structural thing? Does that create any risk to the story?
Do you mean Fabrinet's inventory days or the industry?
I think we're talking about your customers, right? I mean.
Yeah. So we track that very closely and have been for some time. If you go back maybe a year ago, let's say before the telecom industry recovered, we were able to predict because our CFO tracks our customers and even the companies that are big players in the industry, but not necessarily our customers, we've been tracking everyone's inventory very closely. We were able to see whose inventory was getting depleted. From that, we were able to kind of predict when we felt the inventory would recover. Maybe I'll let Csaba talk about the state of the inventory in the industry. It is improving. It may not be back to pre-COVID levels, but I think it is improving.
Yeah. Certainly, from our customers' perspective, we have been seeing a downward trend in the last three quarters. I think generally, the inventory levels have stabilized last quarter compared to the prior quarter. We'd like to think that this environment will stabilize without any bubble or anything that we have seen during the component crisis and pre-inventory correction crisis. In general, we see an improving environment compared to the last two years, but it's certainly not there before COVID. I'm not sure if it will ever be there.
Yeah. Yeah. I agree with that. I mean, it is good to hear that you are seeing the improvement. I think before COVID, right, we were kind of more in a just-in-time world and sort of like three months or less for most things. Maybe now the new normal is more like a six-month type of we do not want to be caught short again that short. I do think it is structural that we are just more of a six-month normal lead time than a three-month normal lead time is kind of, to me, the new normal is the way that I would read it.
The other thing that I think we see, because of course, with the uptick in the telecom business that we're seeing, one of the questions we're asked is, how do you know it's not inventory build again? Of course, the short answer is, we don't know for sure. However, we can tell by the customer's behavior. If the customer is building inventory, what they tend to do is load the orders and then leave us to produce it. In the case of the uptick that we're seeing in the telecom business, yeah, customers are loading us up. They're increasing forecasts in some cases, but then they're expediting us two, three times a week. This is for components that they really need to satisfy customer orders. We don't believe there's a whole lot of inventory build going on right now. It seems to be the demand is what's driving the upturn.
Yeah. All right. Because I'm going to get to all the audience questions, I got a very high-level one. We kind of tried to start with this with the questions we were asking in the beginning, but it's going to give you another chance to come back to this question, Seamus, which is, what makes Fabrinet unique versus others? As you grow, how do you maintain that kind of specialness and uniqueness to add so much value for your customers?
I think we're a specialist. That's the first thing. We're very focused on being really excellent at a handful of things. We're not obsessed with being in 20 different industry segments. We like to stay very focused on being the leader in a handful of industries. That's the first thing. Secondly, I think we're trusted by the customers. They know us very well. Our reputation is very good. We execute very well. I think our consistent track record of just excellent execution for the customers is what gives them confidence to reward us with more business. I think the fact that we're not a product company and we never will be. The customers know that they can rely on us, that we will never bring out a Fabrinet-badged product. Therefore, we will never compete with them.
That's something that some of our customers are very concerned about when they see contract manufacturers becoming ODMs. They get very nervous about that. They don't like it. We can look the customer in the eye and tell them, "We will never do that. We will never have our own products. We will always be a pure-play contract manufacturing service company." In a sense, what we do is it's the origins of the contract manufacturing industry. It's where the industry came from. It's quite traditional in a sense that that's what we do. We do it really well. We stay focused on being excellent at a number of things. I think the final is our costs are very low. We keep our costs under very tight control. A good measure of a contract manufacturing company, people look at gross margin. Gross margin is important.
You have to look at gross margin, OPEX, and therefore operating margin. If you, for example, if you get into the product space as a contract manufacturer, you can improve your gross margins. If you're adding four or five points of OPEX to get there to fund all the R&D, then you'll consume more than you generate. I just think we're very clear on what we're all about. We're very clear on what we do for our customers, but we're also clear on what we won't do and the areas of the business that we're not going to get into. I think we're quite unique in that sense. We're very happy with being the leader in our industry. It's what the customers need from us. We're going to keep doing that.
Great. Perfect. That would be a good place to end, but I think we have five more minutes and we're going to use it. So, Csaba, just remind me, do you guys, do you do any hedging on the Forex or?
Yeah. We do hedging on Thai Baht. Most of our spending is in US dollars, but we have our value-added labor and overhead cost in Thailand is in Thai Baht. We do hedge three quarters ahead. We hedge 100% to the next quarter, 50% in the following, and 25% on. We have a layered hedge structure to eliminate the volatility. That has been working well for us. That's why we continue to do that. Nevertheless, if the exchange rate environment in the recently, we have seen some strength from the Baht and also some weakness in dollars. This may present some headwind in the, I would say, in the Q2 2026 period, depending on the hedge. In Q3, we are seeing a bit of headwinds compared to the current levels of exchange rate.
Okay. Okay. Again, outside of that, we really should not expect the margin, I mean, the margin structure of the company is not changing, right? It is very stable, it seems. Is that fair?
No, structure is okay. We are very focused to make sure that we maintain our margin structure. We did call out a little bit of short-term headwinds in this quarter and in Q1 because of the number of product ramps that we have. We have some extra costs incurred before the revenues start ramping. That should be temporary and should be behind us in the second part of the fiscal year.
Okay. Great. So I mean, so this is a company who, I mean, over time, this sort of very consistent 15%-20% top line, has it been even better than that? Has it been 20%? Or has it been 15%-20%?
It has been. Yeah, it has been hardened up from time to time. Over the 10-year period, 16% compound annual growth rate in the top line, 21% in the earnings. In an industry that, if you look at the EMS industry, it is kind of low single-digit growth rates. The other thing that is important to point out, we should have said this, that is without acquisitions. We tend to not be very acquisitive. We have done one acquisition in the last 10-plus years, a very small acquisition. That growth has come through organic growth without acquisition.
Right. If we just think about kind of the catalysts, kind of new projects in fiscal 2026 that we kind of know should be coming in, I mean, we've got 1.6T , we've got Amazon, we've got Ciena. Is that fair? Am I missing anything?
They're the main ones. You hit it on the head there. Like you say, yeah, 1.6T , the ramp on 1.6T is in front of us. The Ciena new business win will ramp in fiscal year 2026. And the Amazon business will ramp in fiscal year 2026. Yeah, some nice growth drivers as we enter the fiscal year.
Fantastic. That sounds like a great story for me to be covering. And I'm looking forward to following the progress.
All we have to do is execute.
Yeah. Thanks so much for joining us today. I hope your meetings and the conference go well. I really enjoyed the conversation today. It's great to see Seamus and Csaba and Garo. Thanks, everyone in the audience, for your questions and for your attention.
Thanks, everyone.