Good morning. Thank you everyone for coming. I'm Samik Chatterjee. I cover the hardware and networking companies at JPMorgan. I have the pleasure of hosting the next fireside chat here with Fabrinet. Seamus Grady, CEO
Good morning, everyone.
John Swierhun, CFO. Thank you both for coming to the conference.
Thank you, Samik.
Let's kick it off. Seamus, for you, I think just overall, given the amount of interest we're seeing from investors, maybe I'll ask you to outline a bit more about Fabrinet's role in the supply chain, and I know it's a more 101 question to start with, but let's do that, and then we'll move into the more deeper questions here. How would you outline to investors Fabrinet's role in the supply chain and why the role expands as you look at the future in terms of what Fabrinet's role is in the supply chain? How does that expand? Why does it expand as well?
Thanks, Samik. I think our roles of Fabrinet, we're a contract manufacturer. We make other people's products. We don't have any products of our own, I would call us a pure play contract manufacturer. Again, we don't have our own products. We help our customers in a number of ways. First of all, the background of the company was the company was established originally in 2,000 and then went public in 2010, and was established as a contract manufacturer specializing in the then fledgling optical communications industry. 26 years ago, the company was founded by a gentleman called Tom Mitchell. Really, we specialized in optical communications contract manufacturing from the very start and built up expertise that came originally from the disk drive industry, actually.
A lot of the expertise came from the disk drive industry. We're headquartered in Thailand, and that's where most of our manufacturing footprint is. We've really built up a lot of expertise in, you know, putting these products together right down from the wafer level all the way up through finished systems. We do a lot of packaging work at the wafer level. We make components and sub-assemblies and subsystems, and we also make, you know, finished products for most of our customers. Our customers range from companies like Nvidia, Cisco, Coherent, Nokia, Ciena, several of the leading companies in the world.
Our role really is to provide leading-edge manufacturing capabilities for our customers, you know, to be a trusted partner for them, and to do it in a way that allows them to realize some hopefully significant savings if we do our job well. You know, we're a, I would call us a very high capability, high technology manufacturer, but we keep our costs under very tight control because at the end of the day, if we can't produce the product at a better cost than the customer can produce it themselves or than our competitors can, then, you know, there is much reason to go with us. Our role has evolved over the years.
We've gone from being more of a specialist component, contract manufacturer to we still are that, but we also do a lot of complete system assembly as well. Full suite of products and services. A lot of the work we do is at the packaging level. Approximately 70% of our manufacturing space is clean room space, which is quite unusual for a contract manufacturer. We're also expanding quite significantly over the last 10 or so years. Up to 2025, we grew our compound annual growth rate was 16%. In FY 2025, we grew 19%. We're in FY 2026. We're in Q4 of FY 2026. At the midpoint of our Q4 guidance, that would put us at a growth rate of 34% for this year. Our growth rate is actually expanding.
It's increasing, actually, over the last couple of years.
All right. Okay. We'll get to the growth rates in a bit. From our perspective, Fabrinet was very focused on the optical contract manufacturing market. Now you broadened that out to participate in the broader HPC market as well. When you reflect on that shift, what drove you to take the company beyond optics, which is where you are well known for, you're specialized in? How do we think about, are there more adjacent markets that you see opportunities in beyond even HPC for Fabrinet to go into?
Yeah. We're a specialist, and we want to always remain a specialist. We think the contract manufacturing industry has been, I would say, somewhat obsessed with diversification for a long time. You know, if you get too diversified, you end up in 10 different industry segments, and you end up being an expert at none of them. We're very much focused on being, you know, a real expert in the contract manufacturing space. High performance compute makes sense for us for a number of reasons. One is a practical one. We signed a warrant with Amazon, with AWS, a couple of years ago, and the first foray into that business for us was they asked us to produce a range of their high performance compute PCBAs.
It's a good fit for us. They're very complex PCBAs, very difficult to produce, and we've done a good job for them. Secondly, we believe high performance compute is a good fit generally because over time, photonics and optics will become more pervasive in high performance compute. A lot of the high performance compute bottlenecks, you know, are no longer the compute power themselves. It's more the ability to move the huge workloads around, which is where you need optics and photonics. It's a good fit for us. We've also started to, it's early days, but we've started to produce transceiver for AWS as well.
There's a good fit there between High Performance Compute and our photonics and optics capabilities that we think will serve us well in the future. We think there's other High Performance Compute customers who could use our services as well, so we'll be working on that.
Got it. Got it. Maybe going back to the growth rates. In the past, you've referenced long-term growth rate for Fabrinet at 2X the optical industry and 3X the contract manufacturing industry. Now your growth rates are accelerating, as you just pointed out. Is that accelerating because the underlying market's accelerating, or are these more a function of your outperformance not staying at that 2X, 3X anymore because of these moves to the adjacent markets that you are now referencing?
I think it's both. I think the services we provide, our customers appreciate what we can do. We're a little bit different to most contract manufacturers in the sense that we try to do as much as possible down the stack and at the wafer and packaging level. We don't make the wafer, but everything after the wafer is in scope for us, including packaging and making the components that go into the product, and then assembling the next level up in the system, all the way up to complete systems. I think we've always had, you know, very good kind of depth in the services we provide. We've also The breadth has increased over the years. I think that has been a factor.
Of course, the industries we serve have just been expanding very rapidly over the last few years. You know, the best strategy to increase your revenue is execution, and we have a very good reputation with our customers for being a very reliable partner for them and for doing a very good job for them. I think it's a combination of our, the breadth of what we're doing for our customers is expanding. The execution with our customers is very good. We do a good job for our customers, and they appreciate us. Of course, we're very fortunate to have really the cream of crop in terms of the customers we have. Their business is exploding right now, and we're right there with them to help them.
It's a combination of all of those factors.
Okay. Okay. recently you announced that you're acquiring additional land in Thailand at, I think, Nava Nakorn.
how you pronounce it.
That's right.
You're continuing to expand the Chonburi real estate as well. Can you now discuss the strategic rationale to acquire in Nava? Beyond the cost, the low-cost location that you have in Thailand, what otherwise makes Thailand the best suited to, for you to continue to expand in that geography?
Yeah. We have two main campuses in Thailand. The original campus is in Pinehurst, which is closer to Bangkok. We have a significant footprint there, but we're essentially landlocked in Pinehurst. We've converted over all of the non-manufacturing space that we could. We've converted over to manufacturing space. We're in the process of converting about 150,000 sq ft right now in Pinehurst into manufacturing space. That's it. We're kind of maxed out in terms of manufacturing capacity on that campus. The campus in question that you mentioned, or the location in question, it's in Nava Nakorn, also known as Nava. It's about 15 minutes from our Pinehurst campus, so it's a very good campus.
It has a factory on the campus and room to build another one that if we were to build out everything on that piece of land, it would give us additional capacity for about another $500 million of revenue. Right now our revenue, let's say our revenue capacity, before that campus is about $5.5 billion. That would take us to about $6 billion. We're also building at our second campus in Chonburi. We have a building under construction right now, Building 10, which, when it's finished, will be about 2 million square feet with additional capacity for about another $3 billion. That would take our capacity to $9 billion. Then we have enough land again in Chonburi to build another 2 factories, each of which would add about $1.5 billion of revenue capacity.
You know, right now we have land capacity, let's say, for about $12 billion of revenue. We are growing fairly quickly. We You know, the Building 10 that I talked about, it will be fully completed by the end of the year, but we'll actually be occupying two floors of that building, one in July and another in October, to make sure we can accommodate our customers' needs. Lots of room to grow. Thailand as a location, it's a really excellent location we have found to do business. You know, you have availability of really excellent people, who we're able to train and retain. The cost is predictable in Thailand. It's a friendly place to do business, a good location to do business.
The government there are business-friendly. We found it really to be an excellent place to do business, and we plan to continue to grow and expand there for many years to come.
Got it. You mentioned you'll have capacity up to $12 billion.
You also referenced 8.5 on the earnings call that you'll have revenue capacity for. Help us think about timelines. Like, what does it typically take from the time that you start putting sort of brick and mortar in place to filling up that capacity? When you're outlining 8.5 and then 12 of capacity, what should investors think about, okay, this is probably a good way to think about how you've filled up capacity that you've rolled out in the past?
Yeah, I mean, if you go back to, you know, the first factory we built on the Chonburi campus, which was all the way back in 2017. I remember at the time it was 550,000 sq ft. I remember thinking, "That's gonna take us a long time to fill that factory." We filled it very quickly. We built the next factory, which was called Building 9. That was 1 million sq ft. Again, we filled it within a relatively short period of time, you know, 2 or 3 years to get that filled. Building 10, again, 2 million sq ft, $3 billion of capacity. It will take us a couple of years to get that filled up. You know, we guide 1 quarter at a time.
We don't guide. We don't give long-term guidance. It's very difficult to give long, long-term guidance in our business. You know, I think the fact that we're building Building 10 and we're taking over, if you like, two-fifths, two of the five floors ahead of the completion of the full building kind of gives it an indication that we're, you know, we're quite optimistic about our ability to fill up that footprint.
Plus we have the plans already ready to go for Building 11 and Building 12. You know, we think we have a good runway for the next several years, but we are looking for more land as well, in the Chonburi area, to continue to expand because once we finish, like I say, building 10, 11, and 12, then we don't have any more land on that campus. I think at the time when we started there in 2017, we thought building 12 would last us for a very, very long time, but here we are coming up on 10 years later, and we need more land. We'll be looking to continue to expand.
Okay. Okay. Maybe let's talk to the flip side of this. As you ramp capacity, how do we think about gross margin headwinds? You've been seeing some already from a ramp cost perspective, but clearly you're laying out here a multi-year roadmap of continuing to add capacity. How does it play out on your gross margin line?
Our, we have some, you know, gross margin headwinds as we ramp new products. We have many new products ramping at the moment. We have a number of new products with Ciena and Cisco. We also have the high-performance compute business ramping for AWS. We have a number of new transceiver programs that are, all of which are ramping. It's all good news. They're all products that are ramping that, you know, once they're ramped, we benefit from that. They do create a little bit of gross margin headwind as they're ramping. You know, notwithstanding that and maybe also notwithstanding any FX variations, our gross margin has been in the, I would say, 12.5%-13% range.
Our, our, maybe more, more importantly, our OpEx is tiny. You know, as a contract manufacturer, you have to keep your OpEx under very tight control. Our OpEx right now is about 1.4% on a percent of revenue basis. Really, as we grow the company, we don't need to add a lot of additional OpEx. There's probably a little bit more leverage on operating margin than gross margin. We, you know, we do plan to continue to nudge the gross margin up over time. We're, again, we're a service company. We're a contract manufacturer, we're not a product company. We can't just increase prices just because there's a lot of demand out there, and there is a lot of demand out there. I've certainly never seen anything like it. We'll be careful.
You know, we can't just increase prices just because we can. We won't do that. We value the relationships with the customers very much, if anything, you know, we'll be using this growth to make sure we stay very competitive for our customers. Where we can, we'll actually bring prices down as we're able to bring the cost down. We'll be using it to actually grow the business. Little bit of margin expansion, but it'll be gradual and over time rather than any big step-up in margin.
Csaba, just checking, did you have anything on the gross margin?
I think, one thing that is important to mention as we are building out this capacity, our fixed cost structure remains very, very small. Our fixed cost is about 5% of our revenue. If anything, where the growth would pause, we have very, very intact gross margin structure and profile. In that sense, there is not a lot of leverage when the top line grows at the gross margin level. On the other hand, we are protected from the downside as well. I think we are in a fortunate position in terms of, on that.
One of the questions I get often is, clearly Fabrinet is known as the leader in optical contract manufacturing today, and you've talked often when we've discussed about clean room being a key differentiator. What does it take for new competitors to come in into the space and do sort of play catch up to you in terms of what capabilities you offer? Are you seeing any competitive threats emerge, particularly as demand becomes more significant? Are customers looking for additional suppliers and are you seeing sort of new suppliers being helped on that front by the customers itself?
I think what we're seeing with the customers is the customers are asking us to do more. More at the packaging level and at the component level and at the sub-assembly level and of course, at the complete system level. You know, what would it take for a competitor to, you know, get into the business and successfully compete with us? It's easy. You just need about 20 years and, you know, a lot of experience along the way. You know, we've been doing this for a very long time and because it's all we do, we've become quite good at this. There's really nothing to stop.
You know, I think what makes it difficult maybe for a competitor to compete with us is the sheer number of process steps that we do for the customer. We don't just assemble. If you take transceivers or if you take the DCI business as a good example, we don't just assemble pluggable transceivers. We make, in many cases, a lot of the components that go into those products. In some cases, we make up to 60%-70% of the bill of material in-house. You know, that allows us to be very competitive for the customer, but also it creates a lot of stickiness with the customer. The customers trust us. They know we'll never compete with them. We'll never have our own products. They know we won't let them down, and they know we keep our costs under very tight control.
It's kind of a winning formula. Some of the things we have seen some of our competitors do is to maybe get into the product business and become more ODM-like than contract manufacturers. You know, that's their choice. That's something that some contract manufacturers have started to do. That's their choice. But certainly we've tended to benefit from that because when the contract manufacturer becomes a product company, typically a customer somewhere gets upset. We've tended to benefit from that. Really, you know, there's always competition. I mean, everybody wants our business. Everybody's after our business.
We're accustomed to that. You know, we're confident we can continue to stay ahead of the competition and really make sure we continue to kind of invest in the technologies of the future and the customers of the future.
Got it. Given that we're on that topic about investing for the future, how is the investment in Raytec extending your capabilities in packaging? Just help us understand the opportunity you're seeing there, and maybe any other areas that you feel like you should focus your investments in as you prepare for next-generation technologies.
We have not been very acquisitive. If you look over the history of the company, we've grown the company, like I said, we've grown compound annual growth rate of 16%, last year 19%, this year to be about 34%. We've grown very nicely without any acquisitions to speak of. There was a small acquisition in, I think, 2016. It predates me, but it was very small. There's been no acquisitions in the last 9 years. It's all been organic growth. We're not against acquisitions, but we typically try to focus on, you know, growing the company organically we think is a good approach. It works very well for us.
We will do acquisitions or investments if there's a particular capability or technology that we don't have that we feel we need to have. That's really the category that this Raytec investment falls into. It gives us access to for co-packaged optics and also for silicon photonics. It gives us access to some wafer-level packaging processes that we don't have ourselves. We don't have them in-house. We don't have that capability in-house. We've been using Raytec as a supplier. They're a very good company, a very capable company, and we really like how they operate. It seemed to make sense for us when the opportunity came up to invest in them. It made sense for us. We invested, I think, $32 million for about 14% of the company.
It gives us access, like I say, to a number of wafer-level packaging processes, including copper pillar bumping and a few other processes that we want to have access to. With that investment, we also, we haven't announced it previously, but Raytec will be establishing a manufacturing footprint on our campus in Thailand in the coming quarters. We're pretty excited about that. We'll have access to that technology and capability, and it'll be on our campus. Raytec then, of course, get access to our customer base. It's a kind of a symbiotic winning relationship that we're pretty excited about. As for other technologies, again, as we go through all of the new products that we're working on, if something comes up that usually, we develop the capability in-house.
If we find something that we don't either want to develop in-house or we feel we're not able to develop it in-house, we would be open to making further investments. We're not too keen on investing, let's say, and acquiring just to grow the top line. We don't feel the need to do that. We think we can continue to grow at a nice pace just by organic growth and continuing to expand the relationships with our customers.
Got it. I'll ask you the next question, but just before that, a heads-up to the audience. If you have a question, you can raise your hand, and we'll get a mic over to you. While we're waiting for that, you announced the transceiver, new transceiver programs, including you said AWS. Can you help us understand the customer, what drove the customer's decision to outsource the program or engage Fabrinet on the manufacturing side?
I think we have a number of wins that we've announced. In most cases, what drives the customer to come to us is our ability, again, to make, to do a lot of the subcomponents that go into these products and then to produce. You know, we've produced, I would say, north of 50 million transceivers over the years. Well north of 50 million transceivers. We have a lot of expertise, a lot of capability. We're very good at producing transceivers in volume, at scale, with very high yields and very predictable costs. We're very good at this. In the case of these recent wins, you know, they really fall into kind of two categories. One is for a hyperscaler.
You mentioned AWS, where we'll be producing a product, a couple of products for them. They're not our designs. We don't have our own designs, we're a contract manufacturer. We'll be producing for AWS. Secondly, some merchant transceiver business. One is an existing customer. Sorry, they're both existing customers. One is a transceiver customer of ours. The other is more of a traditional telecom customer who's now getting into the datacom world. We're excited about both of those. We haven't sized the opportunities, they're both significant. Certainly, the combination of the two could be as big as our existing business with our main customer, Nvidia. In fact, either one of them could be as big as that. They're very significant. You know, it's early days.
We'll ramp those over the next 12 to 18 months. They're both significant wins for us. We're very excited about them.
Great. Let me just pause you and see if there are any questions in the audience. Any questions?
thanks for the presentation. When you look at your customer base, you mentioned Nvidia being a core. You look at the stack of different technologies, whether at the data center or the individual manufacturer or the component guys, Lumentum, Coherent.
or their subs going down to Dexerials. Where do you see your space? You mentioned Nvidia's core, but do you move all the way up and down as contract through the entire spectrum? If you could give us a little bit more color on that.
Yes, we do. I mean, historically, Originally, we were a component manufacturer, we've been moving both down the stack and up. Down, I think. We're not gonna get into the wafer. We're not gonna be competing with TSMC or anything like that. Everything after the wafer has been produced really is in scope for us. We're able to do a huge amount of packaging. Interestingly, we don't do packaging as a standalone service. We do it as part of a broader service for our customers, where we're taking the wafer, singulating the dies, and packaging the component as long as we're then using that component. We don't have actually any third-party revenue for packaging.
It's we do it for our customers where we're then using the component in the next level up, if you like, in the assembly. Really everything right down to the wafer level packaging all the way up through components and subsystems, pluggable devices, all the way up through complete network systems and everything in between. It's all in scope for us. We like to develop the capabilities in a really kind of a deep way. We, you know, getting into any of these businesses, we don't want to get into those businesses unless we can be really excellent at it and provide a great service for the customer. We've managed to expand it over the last several years.
We're certainly doing a lot more now than we were you know, 7, 8 years ago, that really has contributed to our growth over the years. What it also does is it allows us to be very competitive for the customers. You know, we're able to provide a breadth of services that really no other contract manufacturer can. You know, our margin is, you know, very significantly higher than other contract manufacturers. One of the reasons for that is we're able to offer very compelling, a very compelling business case for the customer. By having us do more, we can save them more. You know, in many cases, by eliminating 3 or 4 other margin stacks that may go on, we can help the customer be much more cost competitive.
It improves the stickiness of the business, it increases the revenue and the, and the margin opportunity, and it also saves the customer a lot of money. It's a very compelling proposition.
Just maybe, I understand, you know, you can't give not a lot of visibility short term, you also do long-range planning, right? Capital outlays for your construction of business. There's a huge gap between no visibility short term and some long-term visibility. Can you maybe just give us a little more detail on what goes into your LRP for capital allocation that you don't have short term and just the sort of the push-pull? Can you sort of transform your business given what's going on more to LTAs or anything else to have a little bit more. You mentioned stickiness, so that's a component of visibility. Maybe just close that gap for us a little bit.
Yeah, sure. As you say, we guide one quarter at a time, but of course, we have to plan much longer than that. I think one of the interesting phenomena that we see going on, historically, you know, any contract manufacturer will tell you one of the struggles they always have is to get long-term forecasts from customers. Customers generally don't like to give very long-term forecasts. Usually where they can, they'll limit it to whatever is in the contract. Maybe a little bit longer for the purpose of planning long lead time components. These are not normal times. What we're seeing right now is most of our large customers are happy to give us, you know, two years or maybe even up to three years of visibility.
That doesn't mean that they're making a purchase commitment, but they're at least telling us, okay, for the products we're making, here's the roadmap that they see in terms of the products that we're making, the new products coming down the track, and also the volume, and therefore the likely manufacturing capacity that they need from us. The, the customers are willingly sharing, like I say, two or three years of visibility, which allows us to make these large capital allocation decisions. You know, in our case, we have to build out capacity to support the customer. You know, armed with that knowledge, it's a relatively straightforward capital allocation decision for us.
What I, what I mean by that is if you take Building Ten as an example, it's 2 million square feet, it takes about 18 months to construct. The CapEx is about $132 million-$133 million, depending on the FX. At full run rate, you know, the building will generate about 40% ROIC, so it's a really good use of the company's capital. At full run rate, about 5 months worth of operating profit pays for the entire building. It's a really excellent use of the company's cash. On the downside, if something were to happen that the world economy were to collapse and AI were to disappear, the gross margin headwind for Building Ten, even if it were to sit idle, is about 15 basis points.
A tiny downside risk versus a very significant upside opportunity. We work very hard to make sure we kind of set ourselves up for that kind of success, where we capitalize on the upside, but, you know, largely insulate ourselves from the downside. It's that customer visibility that we're seeing that's giving us the confidence to, you know, continue to grow and invest the business, again, in a way that's, that we capitalize on the upside. We are risk-averse. We don't like to take big risks with the company's capital, we'll be making sure we protect ourselves on the downside.
Maybe moving on, let's talk about supply constraints a bit. We can see that there's industry-wide supply constraints within the optical industry. Maybe talk about where are you seeing the bigger impact relative to your portfolio, and should we view these, given the demand curve we're seeing, should we think about supply constraints now being structural for the industry, or do you think it's more transient?
I think it's ultimately, I think it's more transient, but how, you know, more transient, but how transient remains to be seen. I think it's really a function of the, you know, some of the component supply pipelines take a very long time to build out. If you take the, you know, a foundry for a laser, it takes a long time to get that up and running and to get the yield to the right level and to get the output to the right level. Really what we're seeing is the demand has accelerated at a much faster pace than the capacity's been able to keep up. There are some pinch points, I would say, particularly around laser supply, specifically EMLs. I don't believe it will be a long-term problem.
I think it will get resolved. That would be probably the biggest one that we're seeing. There's always other, you know, more short-term supply constraints that pop up, especially when you're dealing with significant volume increases like we're seeing. That one in particular, I think, like I say, I think it's a transient one, but it will take a little bit of time for the industry to catch up.
Okay. Okay. Fair. It's almost been a year since you entered into the AWS partnership and had the issuance of warrants. Are you seeing any other hyperscalers or any other customers looking for similar agreements in terms of trying to get visibility into capacity on their own by engaging with you on that front?
We're, I mean, we're talking to several other companies, I would say. You know, I don't think they, hyperscalers necessarily look at each other to determine their strategy. They each kind of do their own thing. Our relationship with, I would say, all of the hyperscalers is very good. We know them quite well, interestingly, predominantly because of our, the strength in our DCI business. Most of our DCI data center interconnect customers, you know, they're shipping those products, 400ZR, 800ZR products, to the hyperscalers. The hyperscalers are in our factories all the time auditing the production lines that we're using for our customers.
They know us very well, they know what we can do, and they, you know, they have a good, I think, a good impression and a good feeling for the capabilities that we have. That has been a real kind of a catalyst for us to allow us to really start a more meaningful dialogue with some of the hyperscalers. Any of these relationships take a long time to develop. In our business, it's not unusual for when you engage with a willing customer until you're actually shipping something, it can be 18 months to 2 years. There's a long kind of gestation period from engagement until you're shipping revenue.
We are talking to a few of them, we'd be hopeful we can work with more hyperscalers directly. Again, where it's appropriate, because in many cases it's not appropriate. Like I say, we'll never produce our own products. We'll never have products that we'll be competing with our customers. Where the hyperscalers need someone to produce products for them with a direct contract manufacturer relationship, we think we can do a good job for them.
Got it. A lot of the discussion with investors on the optical front is on OCS products at this point. How are you thinking about the opportunity? What are the breadth of engagements do you have with customers on that product? How should we think about once those engagements turn into wins, how soon can we see revenues related to the OCS products?
I think, you know, for us, OCS, we're not really going to opine on whether OCS is a winning product or not a winning. For us, it's a product we're very capable of manufacturing. A lot of the technology that underpins OCS we're very comfortable with, we're very familiar with. Our role really will be to help, you know, our customers to ramp up that capability. It does look to be very promising, that's for sure. There looks to be a lot of demand there and a lot of volume there. We'd be, you know, we'd be very excited to work with a couple of our customers to get some of the products off the ground and ramp them for them. Yeah, we think, you know, there's a few product areas.
OCS is one, ELSFP is another. There's a few, I would call them new-ish product areas where we're really focused on winning, and OCS is certainly one of those, and we're excited about it.
Good. Last one quickly for Csaba. Just given the capacity expansion plans that you have, how should we think about free cash flow for the medium term? Do we see a change in the free cash flow conversion rates on account of the capital plans that you have?
In the past, we have been very strong free cash flow generator company. As you can see, our balance sheet, we have a close to $1 billion cash on the balance sheet. In the last two quarters, we have seen some pressure given the capital expansion and working capital build-out. In the near term, we do anticipate that to continue. The good thing is that we are able to still generate 40% ROIC on this investment. We look at it as a fundamental strategy of our capital allocation, investing in our growth and using our own cash to build out this capacity. In the short term, I think it will be somewhat compressed, but nevertheless, we are still not compromising on the growth and the ROIC that we are returning on these investments.
Okay. Great. I'll wrap it up there. Thank you. Thank you both.
Thank you.
for coming to the conference. Thank you to the audience as well.
Thank you.
Thank you.