Celestica Inc. (CLS) Earnings Call Transcript & Summary

Ladies and gentlemen, thank you for standing by, and welcome to the Celestica JDM Roundtable. [Operator Instructions]. I would now like to hand the conference over to your speaker today, Craig Oberg, Vice President of Investor Relations and Corporate Development. Thank you. Please go ahead.

Good afternoon, and thank you for joining us for a discussion regarding our JDM offering. On the call today are Rob Mionis, President and Chief Executive Officer; Mandeep Chawla, Chief Financial Officer; and Jason Phillips, President of our CCS business. As a reminder, during this call, we will make forward-looking statements within the meanings of the U.S. Private Securities Litigation Reform Act of 1995, and applicable Canadian securities laws, including, without limitation, financial projections and guidance as well as statements related to our priorities, strategies, targets, expectations and anticipated operating results. Such forward-looking statements are based on management's current expectations, forecasts and assumptions, which are subject to risks, uncertainties and other factors that could cause actual outcomes and results to differ materially from conclusions, forecasts or projections expressed in such statements. For identification and discussion of such factors and assumptions as well as further information concerning forward-looking statements, please refer to our press release issued on December 2 including the cautionary note on forward-looking statements therein and our most recent annual report on Form 20-F and other public filings, each of which can be accessed at sec.gov and sedar.com. We assume no obligation to update any forward-looking statement, except as required by law. In addition, during this call, we will refer to projected non-IFRS adjusted earnings per share and non-IFRS free cash flow. These non-IFRS measures do not have any standardized meaning prescribed by IFRS and may not be comparable to similar measures presented by other public companies that use IFRS or who report under U.S. GAAP and use non-GAAP measures to describe similar operating metrics. We do not provide reconciliations for forward-looking non-IFRS measures as we are unable to provide a meaningful or accurate calculation or estimation of reconciling item and the information is not available without unreasonable effort. We are also unable to address the probable significance of the unavailable information. Forward-looking non-IFRS financial measures may differ materially from the corresponding IFRS financial measures. Unless otherwise specified, all references to dollars on this call are to U.S. dollars. During this call, we do not intend to, nor, in our view, will we disclose any material nonpublic information. Let me now turn the call over to Rob.

Thank you, Craig. Good afternoon, everyone, and thank you for joining our Joint Design and Manufacturing Roundtable discussion. As you know, JDM is an important part of the Celestica portfolio and has been experiencing very strong growth in 2020, and we expect JDM revenue to exceed $800 million this year, up over 80% compared to 2019. On today's call, we are pleased to have Jason Phillips, President of CCS, join us to discuss the history of JDM, where it stands today, and how this business is strategically positioned for growth going forward. We believe, JDM is a highly strategic asset within Celestica's portfolio. It is an engineering-led hardware-focused business. It has high barriers to entry, long product life cycles, generates margins which are accretive to the company's average and leads to long-term sticky customer relationships. As such, we believe JDM warrants evaluation multiple, similar to that of our ATS business, given their similar characteristics. Before we begin our JDM discussion, Mandeep and I would like to provide a few updates since our last discussion in October. As we look to 2021, we are targeting ATS to grow 10% year-to-year, driven by strong growth in capital equipment, industrial and healthtech, as a result of new program wins and market growth. The strong performance in these end markets is expected to be slightly offset by continuing softness in commercial aerospace, which we expect to remain depressed through 2021. We expect ATS segment margins to continue to recover and ATS to be within its target segment margin range of 5% to 6% by the end of 2021, driven by volume leverage and cost productivity actions. Now turning to CCS. We expect JDM to continue to grow in 2021, albeit at a slower rate compared to the approximate 80% in 2020. As expected, our non-JDM revenue will decline year-to-year, primarily driven by the disengagement with Cisco. CCS segment margin is expected to be firmly in the target range of 2% to 3% in 2021. I would now like to turn it over to Mandeep for a brief update on our fourth quarter guidance and capital allocation priorities.

Thanks, Rob. Looking at our guidance, we continue to project our fourth quarter revenue to be in the range of $1.35 billion to $1.45 billion, and fourth quarter non-IFRS adjusted earnings per share to range between $0.22 and $0.28. Turning to capital allocation. I'm pleased that we are on track to generate at least $100 million of non-IFRS free cash flow in 2020. We expect to achieve the same goal in 2021. As we have shared previously, our capital priorities are to return 50% of free cash flow to shareholders and to invest the remaining amount in the business over the long term. We are pleased with our track record over the last 10 years, making strategic investments in the business, while at the same time, returning over $1.2 billion to shareholders through buyback. When making an investment decision, our approach is well defined. We look for investments that are aligned to our strategic roadmaps with a filter intended to ensure that they drive accretive EPS in year 1 and an ROI that exceeds our cost of capital by year 2 or sooner. We also weigh these investment decisions against alternative uses of cash, such as share buybacks. Currently, we believe our stock price is undervalued, having traded below tangible book value for the majority of 2020. In response to this undervaluation, we launched a share buyback program on November '19. In the near term, if the stock is trading below tangible book value, we intend to be active in the market. While we continue to have an active M&A funnel and have reviewed over 50 companies so far this year, none of those opportunities made it through our robust filter. We will continue to assess M&A transactions against alternative uses of cash. However, we recognize that buying back shares that trade below our tangible book value is an accretive use of cash. I'd now like to turn the call over to Jason to discuss our JDM business in more detail. After Jason's remarks, Jason, Rob and I will be pleased to answer any questions you may have. Jason, over to you.

Thanks, Mandeep. 2020 has been a banner year for JDM. Like Rob and Mandeep said, JDM is quickly becoming a large part of Celestica's overall business. While the success may appear to have come overnight, in reality, the past 25 years have been leading us to this point. This would not have been possible without Celestica's legacy of delivering enterprise-class quality and execution through deep market insight and technical know-how for our customers. These capabilities and the credibility of our team take time to establish, and our experience in the core CCS business laid the foundation for JDM. In addition, we've been investing in our JDM capabilities for more than 10 years now, spending $20 million to $30 million in R&D per year in a broad portfolio of networking, storage and compute solutions. Over our 10-year journey, we've invested and grown this business beyond the boundaries of the traditional scope of joint design. We've evolved as our customers' needs have evolved, developing a suite of fully realized hardware platform solutions for our customers. When you combine our deep industry expertise and our targeted R&D investments with highly favorable market dynamics, you have a recipe for success. Disaggregation and the rapidly accelerating shift to the cloud opens up opportunities to bring high end solutions to the market when traditional off-the-shelf products do not suffice. Our JDM business started in 2010 from humble beginnings. JDM started with a flagship customer who understood the depth of our engineering strength, they needed a storage solution to meet specific requirements, technically and commercially, and we successfully executed this design for them. Our continued strong performance led to more opportunities, that expanded into other customers and technologies, first, networking and then compute. And while we broadened into new and emerging technologies, we also expanded the scope of our design engagements, moving from joint design to developing fully realized off-the-shelf hardware solutions that could be used as is or customized to meet a customers' specific needs. Our evolution from joint design to hardware platform solutions positioned us to enable our customers' evolving strategies. Advancing technologies were driving our customers to focus on their relevance in the cloud applications and developing strong software and services offerings. They needed to redefine their hardware partnerships in order to keep pace with the rapidly advancing software and dramatically increasing demand for bandwidth, storage and access to critical data. So while our JDM value proposition of faster time to market and lower total cost, while creating and protecting IP was still relevant, customers now needed more from us. And our hardware platform investments had ideally positioned us to meet this need. Celestica had developed an unchallenged, unique value proposition, and we were able to seamlessly take over the task of driving hardware innovation for our customers. Customers were able to use our standard off-the-shelf designs or work with us to customize hardware to the application, software or workload environment. They relied on our proven track record of performance, our reputation for quality and our strategic ecosystem partnerships to underpin the strength of our diverse portfolio. What we saw then and continue to see today is that as applications and software grow more complex, the capability of the hardware that underpins them is critical. Hardware still matters. As I've said, the JDM journey is 10 years along now, and our persistence has paid off. Today, we have more than 500 design engineers across the globe, working on next-generation hardware platform solutions per our comprehensive product roadmaps. Our footprint in China, Thailand and North America allows us to provide white glove service to customers globally. We have over 280 patents safeguarding our hardware, technical superiority, and we drive a robust IP management process. With flexible customer engagement models, we can either share IP development with customers where they can choose to leverage our IP. And we're proud to say that we've launched over 150 programs and shipped more than 2.5 million units to our enterprise and service provider customers worldwide. We are asked a lot, how is JDM different than an ODM model? First, Celestica covers all technologies and form factors in the data center with a focus on the more typically challenging solutions in the storage, networking and compute equipment sets. While ODMs are generally more focused on a common technology, we take a collaborative approach with customers and ecosystem partners on the development of these products and in the development of product roadmaps for the future. But where we really diverge from ODMs is in our customer-driven approach to customization and product life cycle support. We work closely with our customers to customize hardware and product life cycle solutions with the flexibility to meet their needs every step of the way. Finally, and this is an important distinction, we will not directly compete with our customers. We don't brand our products. We don't have a direct sales channel, and we don't have application software for our solutions. We want to enable our customers. We focus on hardware, allowing customers to focus on their core competencies, software and services. We are proficient in software, particularly in BIOS, firmware and porting, but only as far as it enables our customers' success. For the techies out there, we have a very robust product offering with compelling roadmaps for current and future technologies. But for the sake of the non techies, I'll cover this quickly. We do everything from JBODs, RBODs, storage controllers and arrays, including HDDs, flash and converged offerings, to networking solutions from 1G to 400G plus speeds with 800G under active development, including the latest and optical interconnects to 1U and 2U modular servers that can be optimized for acceleration and specialized applications, such as artificial intelligence and machine learning. This slide shows just a few examples of successful platforms we've introduced to market through our customers. I thought we would bring our JDM value proposition to life through a few case studies that best illustrate our JDM capability, offering and how it benefits our customers. Looking at our first case study, a global industry-leading service provider had a long-term relationship with a provider of off-the-shelf products. The current generation of their networking solution was not performing as expected. There were design quality challenges and limited flexibility in delivering what this customer wanted. Ultimately, the customer abandoned gen 1 and started looking for a better solution as they prepare for the next gen product. The customer asked for our help to develop a leading edge, high bandwidth switch design with onboard optics through a collaborative and flexible engagement model. We leveraged our networking roadmap, rigorous design standards in execution, flawless launch process and enterprise-class manufacturing capability to deliver a successful solution that met our customers' needs. The result was the successful deployment of a truly leading-edge product, leveraging the latest optical technology and optimize for our customers' specific applications. And by leveraging our IP, our ecosystem partnerships and the strong performance and efficiency of our design engagement process, the customer saw a dramatically improved time to product availability, and lower development costs. In the next example of our JDM value proposition at work, a leading global OEM and long-standing Celestica JDM customer was aiming to broaden their hardware portfolio while optimizing their R&D investments. Resource constraints and R&D budget overages were limiting their product offerings, driving them to consider other R&D solutions that would increase flexibility, reduce spend and maximize R&D return on investment. We ended up transitioning our customer from a joint design engagement to our hardware platform model, leveraging Celestica's IP to expedite innovation at a significantly lower cost. Moreover, we were able to align our roadmaps with our ecosystem partners to ensure that our roadmap for the future would support the customers' requirements for the long term. As a result, Celestica was able to reduce R&D spend for our customer by over $1 million and enable the redeployment of R&D resources to other high priority projects, all while maximizing supply chain flexibility and reducing time to market. Before I wrap up, I've been living and breathing our JDM story for the past 10 years. I hope you're sensing that I'm a passionate believer in our vision. We've gotten to this point of delivering differentiated value by staying very close to our customers and our industry partners, investing in the right areas, and maintaining a stubborn persistence and faith in our vision for this business. It's an exciting time for our team, and I want to acknowledge their commitment, especially in this unprecedented year. With that, I want to stop here so that we leave plenty of time for Q&A. Operator, I'd now like to turn the call over to you.

[Operator Instructions]. Our first question comes from Paul Steep with Scotia Capital.

Okay. Great. Can you maybe go over a little bit about how we should think about JDM in terms of the growth areas of that business? I know you briefly touched on the portfolio, but maybe just outline what you're seeing the greatest growth in the portfolio, and how we'd we want to think about maybe additions to that portfolio?

Sure. I'll go ahead and take that one. I would say we're seeing in the storage space a consistent business. There's still a strong place for storage. Some of that is -- of the storage capacity and capability is moving into the compute side of the solutions, but I would say external storage remains a steady business. We're seeing a lot of growth in the networking space. And I think we'll continue to see that as we see data centers continue to expand as everyone continues to work and learn from home and beyond. And compute will continue to grow, I believe, especially as we start to push beyond the data center into the edge.

Great. And then maybe the second one to just give us all perspective, and it's been helpful is just contrast the financial profile of this type of transaction, maybe in terms of working capital, customer commitment. Any other elements you could think of, contract versus the traditional Celestica EMS business? And then I got one last wrap-up.

Sure. Well, it's a very different engagement model from traditional EMS. And it starts with investments in the design through the life cycle solution, and they all have varying aspects. So commercially, we're talking about the investments in the design all the way through the manufacturing, and then it can include services in support of that product through its life cycle. So we typically will customize the commercial and the engagement model, depending upon the customers' requirements and the solution.

Great. And then the last one for me is maybe, Mandeep or Rob, you could talk to the level of investment in this business. I know, traditionally, it's been $25 million a year, and it sort of crept between $25 million and $30 million. Is there an argument given the size of the business today to actually double down and sort of increase that level of investment? Or we think that's sort of the envelope that makes sense?

Of course, Paul, it's Mandeep here. So we've been able to maintain a relatively consistent investment for much of the past 10 years, as you mentioned, in the $25 million or so range and have built up the infrastructure that we have. That being said, with the level of revenue growth that we've seen over the past year, there is an opportunity for us to invest further. And so as we're looking at the business going into 2021, we are making additional investments. And I would say that it's not going to grow at the same level as the revenue. So we are going to have scale benefits, but the business clearly does warrant further investments, and we intend to do that.

Our next question is from James Suva with Citigroup.

This is Jim Suva here. In your prepared comments, you gave a lot of disclosures or commentary about how diverse your product portfolio is. You mentioned servers, storage, connectivity and things like that. I don't think many of us recognize or fully grasped the magnitude of the diversity of your JDM business. So is there a way for you to give us either the numbers or percents or majority? Is it like 1/4, 1/4, 1/4, 1/4 of those different ones you named? Or is it -- do you have your leading edge in servers but you're seeing higher growth like in 5G? Or how should we think about, again, all those product end markets that you laid out, which seems very diverse, the magnitude of each of them versus the others?

Sure. Jim, this is Jason. Yes. Our broadest areas in the portfolio would be in storage, storage controllers as well as in networking, and we have a growing offering on the compute side. And so we have a number of what I would call core platforms across the storage space, whether it would be the RBODs, [ RBOFs ], the enclosures, the controllers that drive the storage systems. And then on the networking side, we've got a range of platforms from Leaf and spine and switching solutions to routing solutions. We do have ruggedized solutions, and there could be variance of each of those. And then on the compute side, I would say in the 1U, 2U modular space, we're more focused on specialized niche applications, things that could be used at the edge that are ruggedized as well or for specific workloads, whether it's in the data center or the edge. So hopefully, that gives you a little bit more color on the breadth of the offering, and with each platform, there can be a number of variants. We talked a lot about our ability to customize solutions. So what we do is we make these investments and we've been making these investments over the last 10 years, is we've built now this comprehensive suite of offerings. And we often will prove out technologies as new technologies come online. We'll develop those into our roadmaps, and we'll prove them out. And as customers see the demonstrated capability, they look to optimize for their specific workloads, their specific applications. And so we will take the platform and create variants for each customer. And sometimes variants will grow into another variant depending upon what problem we're trying to solve, what specific application we're looking to solution for.

Great. And then my quick follow-up is, a lot of the world is moving towards cloud solutions. And you talked about you developing both your own IP as well as jointly IP with your customers. Can you talk about -- do most of your products kind of end up in the cloud? Or do they end up being branded by your customer or more white boxes? Because it seems like there's definitely a big push towards white boxes. And some people may be concerned about the disintermediation that white boxes are bringing to the marketplace?

Sure. So a lot of our solutions will end up in both, Jim. So we have -- when you look at our service provider customers, we've talked about supporting and being engaged with 8 of the top 10 hyperscaler customers globally. We're doing business with 4 of the top 5 storage players. That would have branded products, if you will. That gives you a sense of where our solutions end up. In terms of white box, our hardware platform solutions do align or can align with a white box offering, which is ultimately a bare metal hardware coupled with software. But when we look at our hardware platform solutions, we look more beyond just the hardware at the total solution. So again, it can be customized hardware, wrapped around a customized engagement model. That could be coupled with services as an example.

Thank you so much for the time and the details and clarifications. This was a great learning experience. Really appreciate it.

Our next question is from Robert Young with Canaccord.

I was hoping you could put a little more precision around the growth expectations in JDM. I think the goalposts are pretty wide. I think you said it would be less than this year, and it will still grow. So it's somewhere 0% to 80%. Is there any -- maybe a little more information you can give there?

Rob, this is Jason. Yes, we continue to see strong growth in the business. I think it will temper a bit as we've seen the pandemic accelerate, a lot of data center expansion, but there's no question data centers will continue to expand. And with that expansion, we believe our JDM business will also continue to grow.

Rob, it's Mandeep here. I would say just to your point, the goalposts are quite wide, and the 80% is on the back of a level of significant wins that we had in 2018 and in 2019. If you were to think about it as a high single-digit growth rate going into 2021, that would be more appropriate for next year.

And then maybe long term, is this something that would be better growth in the ATS business in general? Or would you just think of it as sort of a high single-digit sort of normalized growth?

I'll talk -- I won't compare the 2, but I'll talk about them independently. So on the ATS side, as you know, the markets that we're aligned with are growing anywhere close to the mid-single-digit range. But with the level of outsourcing that is increasing, that gives some accelerated growth in the various end markets. And that's why we continue to feel that the 10% long-term growth rate for ATS is appropriate. On the JDM side, you don't have those same market dynamics happening. So right now, we feel that a high single-digit growth rate for next year is more appropriate.

Okay. And then I think you said something in the prepared remarks around not having a direct sales force supporting this. I don't know if I heard that correctly, but I was curious, how do you find opportunities especially if opportunities are coming from a new class of customers that you haven't engaged with in the past?

I'll let Jason take that one on the sales channel.

Sorry, Rob, could you repeat that again, please?

I'm not sure if I heard it correctly, but I think I heard it said that you didn't have a direct sales force attached to this in the prepared remarks. And if that's the case, how do you find -- how do your customers find you, especially if they're not existing relationships like the hyperscalers I'm presuming?

Yes. No, we do not have a direct sales channel, Rob. And we typically -- it's a direct engagement with our customers, which lends itself to the customized approach and solutions that we that we typically provide here.

Okay. And then maybe I'll just ask 1 last question around the -- I think in -- said there are 500 design engineers and there's some split in the business between sharing IP development and then selling the efforts of that IP development as a product. Is there any way to sort of split that or to maybe describe those 2 business models?

Sure. In terms of the IP question, we create our own IP, and then we leverage it in solutions, and sometimes we will co-develop IP with a customer. And there are certain instances where when we do co-develop IP, a customer will allow us to leverage that IP in a customized commercial model we might establish into other designs. So hopefully, that sheds some light on the IP question. I'm not sure I quite understood the question on the resources.

I'm just trying to understand how the utilization of those design engineers, like, are they fully utilized on developing product all the time? Or like is there a way to split it between shared IP developments and developing product that you're building for your own inventory?

Yes. Yes. So if you think of the resources in total, there is a significant portion focused on the hardware piece in executing the hardware elements of the design. There's a significant portion that are focused on architecture of the design in terms of how it functions, generating the market requirements for the solution, the product requirements as a result of the market requirements. And then I'd say another significant portion is focused on software. While we're not in application software, we are very much involved in everything from what connects the hardware to our customers' application software and everything in the middle, and there's a lot of resources focused in that space as well.

Okay. Maybe 1 last little one. You said you had M&A targets that you'd evaluated, but would any of those 50 have been attached to this JDM effort? Would there be M&A that would supplement the growth of this business?

Yes. Rob, so we haven't had very many targets that have met our strategic filter thus far in JDM, but it is an area that we are open to. As we've spoken about before, the majority of the targets traditionally land in the aerospace and defense or in the healthtech sectors, but we are open to making inorganic investments into the sector when appropriate.

Our next question is from Matt Sheerin with Stifel.

Yes. And doing this. And just another question on the JDM business regarding margins. Could you talk about the margin profile relative to the core business? And your gross margin for Celestica has improved in each of the last few quarters year-on-year. I know some of that is mix of business, divesting some lower-margin businesses. But has JDM been part of that positive impact on margins?

Matt, Mandeep here. So the answer is yes. JDM is part of the story on why our margins have been expanding. The business is a -- has a heavy fixed cost structure because of the investments that we are making, which we've talked about in the $20 million to $30 million range. The business today is operating margins that are accretive to the total company and therefore, accretive to the CCS business as well. With the growth that we've been seeing in JDM as well as the reduction in more fulfillment EMS type of CCS business. It is one of the drivers that has led to the very strong margin profile that CCS is having. The other thing, of course, is improved commercial terms with certain customers as well as the productivity actions that we've taken over the last 18 months. But from a pure mix perspective, the JDM portfolio is helping. As we look forward, we expect that this strong margin profile will continue. So one of the earlier questions in the sense of will we continue to invest in this business, we will. But when you look at those investments in the context of being able to continue to have a strong margin profile in JDM. So the business has to pay for it, frankly. And the outlook that we have right now, that's possible.

Okay. Understood. And you did comment, there were some comments earlier about the customer mix. You talked about hyperscale, traditional OEMs, but could you give us a sense of that mix what percentage of that growth that you've seen, those hardware products end up on-prem versus in a public cloud environment? And is there any customer concentration within that?

Matt, yes, we won't talk about customer concentration. I would say we've seen a lot of growth with our hyperscaler customers. And I would say our enterprise OEM customer base has been fairly steady. I would answer it that way.

Okay. And just lastly, could you give us a sense of the typical product life cycle of these products across the portfolio relative to, I guess, the -- would be your compute or your Datacom legacy businesses, those traditional businesses? What's the difference?

Sure. On the compute side, you're going to see faster life cycles, you're going to see, call it, 12 to 18 months, depending upon the technology and how it's evolving on the compute side -- or on the networking side, you're going to see anywhere from, call it, 24 to 36 months. And again, it's technology dependent. And some of those technologies will have longer lifespan depending upon where they're where they're playing inside the data center in the market. And then on the storage side, I would say 24 to 36 months is another good range as well in terms of a product life cycle. And again, as we create these platforms, we will leverage some of the technology that we develop from one generation to the next. That can be applicable in next generation. And that's why over the last 10 years, we've been able to build up this comprehensive suite of offerings that build upon one another, if you will.

And Matt, maybe just 1 thing I'm going to add on to 1 of your earlier questions in terms of the customer diversification. If you go back over the past 10 years, this really did start off with our traditional OEM customers, EMS customers who knew about our skill set and who we were able to partner with. And then as we were able to build out the JDM business, we become very well-known in the space, and it also goes to an earlier question from Rob Young, where we're partnering with ecosystem. We have ecosystem partners who are also engaging with us in developing new business. And so one of the things if you look at our business today is while we've been able to expand our customer base to hyperscalers, we also have a diversity of programs with many of these customers. And so we're not just doing a single type of product with many of these customers, we're doing multiple products. And so we have diversification within our customer base in addition to OEM data center type of customers.

Our next question is from Paul Treiber with RBC Capital Markets.

Just wanted to follow-up on your -- the comment you made on the growth outlook for JDM. The -- I was hoping you could put it in context with the historical growth rates that you've seen for JDM? Is it typically lumpy? Or is it more consistent? What's sort of been the average growth rate over the last couple of years?

Paul, I would say, as we started in the business, we started with -- in the enterprise OEM side. And the growth rates were good. They were, I would say, strong on a year-over-year basis. They would -- I would describe them as steady. As we've gotten into the hyperscaler space more and more and started to grow that business, we've seen stronger growth. And then, of course, we've seen the pandemic, which has come along and accelerated a number of the programs that we've won over the last, call it, 12 to 24 months and has really helped fuel some of the growth we've seen there. So I would say steady on the enterprise customer side, steady growth year-over-year. And then on the hyperscale side, we've seen it -- we've seen some stronger growth as the data centers globally continue to expand.

And then shifting gears to the competitive landscape. I was hoping you could give us a sense for what is the competitive environment for JDM in that -- what's the sales cycle? Is there typically like bake-offs for these engagements? And who do you typically see when you're competing for them?

Yes. In terms of how we compete in the space, a lot of it has to do with your demonstrated capability and your roadmap. You've got to have the right platform, the right offering to demonstrate your competency and capability in that technology. That is a big driver of how we win in the space. And then as you get interest with a platform that's solving a problem or enabling a customer that then turns into the engagement model, both commercially, technically, at an engineering level as well as how you're going to support the solution throughout its life cycle. So that's where we focus. That's how we compete. That's frankly, how we're different than you heard us comment on in terms of some of our competitors in the space.

And just lastly, in the prepared remarks, you mentioned that the barriers to entry are high and customer engagements are long and sticky. The -- could you elaborate on why that's the case? And if there's any sort of anecdotes you could provide in terms of the length of customer engagements?

Sure. So ultimately, we're aiming to become extension. I mean, when we started in the space, we became extensions of our customers' engineering teams, ultimately, working closely with them, developing these platforms. And over time, as our customers shifted more focus into software services and their strategies and trust started to really grow between us, they allowed us to drive more of the hardware innovation for them. And so we started taking over more and more of the hardware requirements and supported their strategies and their offerings.

Our next question comes from Thanos Moschopoulos with BMO Capital Markets.

Maybe just expanding on the competitive question. In most of the time, are they evaluating your capabilities versus competitors? Is it typically more of the case that they're evaluating whether to try to develop something internally versus using you guys? Or are they sort of trying to look for off-the-shelf solutions, which maybe haven't met their needs and that's how discussion begins? What would you see you see in the majority of cases?

Thanos, it's Jason here again. Actually, we see both. We see both. We see in the -- where we started was jointly designing, and we still see a lot of that. We're -- customers want a very unique or custom solution for a very specific requirement, workload or application. And then we'll see some customers that are the platform that we've developed and the capability of the platform works well as it is, and they can integrate it rather quickly, plug and play. And then we see variants or derivatives in the middle of that, where we'll take a platform and customize it to certain degrees, whether it's the IO, whether it's the control plane, whether it's the power packaging and cooling element of a design to meet the requirements that the customer has for its use case?

But I guess in terms of evaluating your capabilities versus competitors, would it be sort of a minority of cases where it ends up being competitive? Or how frequent does this competition come up?

Yes. When a customer goes from 1 generation to the next, if it's a current customer, we're working on next generations together. So we're building the roadmap, the future roadmap together jointly with our customers and our ecosystem partners, and it tends to be very sticky, and we're focused on obviously, maintaining those engagements and growing those engagements for the future. When it's a new customer, it's really -- we're up against -- sometimes it's -- a lot of times, it's ODMs, as an example. And then it gets into how the platforms compete, how this solution competes, how the commercial and the engineering engagement model competes. And again, I've mentioned those a few times as areas where we look to differentiate. And that's where we focus when we look to compete in this space.

Great. And then finally, we've talked about JDM in the context of CCS. Are you doing much on the ATS side? Is there an opportunity there? I mean, you alluded to ruggedized solutions, for example. So is there an opportunity with an ATS?

Yes. There is an opportunity for taking these technologies and applying them into the ATS space. And a lot of it would be form factors, ruggedization. But when you look at the electronics, the designs, the functionality of the platforms, there is opportunity there. It's going to be -- and again, how it applies on the ATS side and where [ it ] would go ruggedization, form factor, power, or things of that nature.

But are you to clarify you doing much in ATS today? Or is that sort of maybe an area where there might be more opportunity to further penetrate?

It's Mandeep here. We're not -- we don't have any JDM sales in ATS right now, but it is an opportunity for our ATS business to leverage a lot of the design work that our existing organization has done. So it's more of a -- something that we would continue to explore.

It's an emerging area for us, and we're making some investments in that area to bring some products to life.

And our last question from Ruplu Bhattacharya with Bank of America Merrill Lynch.

Jason, thanks for all the details on JDM. I have a couple on JDM, and then I'll ask Mandeep a question on CCS. On the JDM side, is your revenue mostly with customers in North America, can you give us a sense of the geographic mix of your customers across regions?

Ruplu. Yes. So we have, I'd say, a good blend, a good mix of customers across the globe. We do have a large concentration of customers in North America as well as in Asia.

Okay. Great. And I'd like to maybe step back a little and maybe, Rob or Mandeep, feel free to chime in on this as well. When you look at JDM as a business, I mean, what's management's long-term goal for this business? I mean, is this something that you think can be incubated and maybe eventually, if this becomes large enough and meaningful enough, is this something that you can spin-off as a separate entity? Or do you think that this is really integrated into a Celestica, and this is something you want to grow organically within the company? So just as a long-term vision, how is management looking at this portion of the business?

Yes, Ruplu. Right now, at the scale that JDM is, we think, the integral part of the business, and it's -- we're stronger together than apart. So we're looking at continuing to invest in JDM and growing our JDM business inside Celestica, at least at the current time.

Okay. Got it. And then maybe on the CCS side, I know for ATS, you said it's going to grow 10% year-on-year in 2021. Looking at the CCS side, can you just give us an update on the pruning that you've been doing in that side of the portfolio. Is that more or less done? And how should we think about growth on that side of the business? What are the puts and takes that we should keep in mind?

Yes, Ruplu. So I would say that as we talked about, we are expecting that the JDM portion of the business to be over $800 million this year, and we will have -- we are targeting high single-digit growth going into next year. For the remaining part of CCS, we do expect it to be down. Cisco is going to come out. We're expecting essentially 0 revenue for Cisco next year. So that's clearly going to weigh on the overall growth rate. That being said, there are other parts of the EMS portfolio within CCS where we have a number of wins, and we are focusing on ramping that. So right now, CCS is expected to be down double digits next year as we ramp some of those new wins that we've been having and the reduction, again, is going to be largely because of the Cisco impact. In terms of portfolio shaping and how we look at it, it's always an ongoing discussion with our customers. It's nothing that is really just confined to the last 18 months. It's a part of how we do business. Sometimes we are looking at the right mix of customers within our portfolio. Sometimes we're looking at the right programs within a customer. And so what I would say is that those conversations will continue nothing to announce at this time, but it's normal course for what we do.

Got it. And maybe I'll just sneak one more question in. I think you -- from a free cash flow standpoint, you talked about doing -- or generating another $100 million or so next year as well. Can you remind us on your targets for working capital? Like what are the metrics you're targeting for days of sales and inventory? And is there any improvement there? And in terms of CapEx, what is the expectation for next year?

Yes. So Ruplu, we are targeting $100 million or more of free cash flow next year. That is our strategic goal on a year-over-year basis. And we're pleased that we're on track so far. Still got a month to go, but we're feeling positive right now [ of ] 2020, and we think that, that's the right goal for 2021. We haven't given a specific working capital cash day target for next year. But I'll talk about it in overall dynamics. We're really pleased with the improvement that we've seen in accounts payable over the last 18 months or so. We think that, that performance will continue. We've always had very strong receivable collections and that performance we expect to continue. And then on the inventory side, the constrained material environment is improving. But as you know, there is a significant amount of volatility that continues with the COVID cases around the world. And so there are constrained material still. We are targeting to improve our inventory performance next year, but it won't all go straight into cash because, as you know, the teams have done a lot of work with our customers. To secure deposits during this time of elevated inventory. So some of that inventory unwind will be offset by deposits being unwound as well. But when you put it all together, again, it gives us comfort at this point to set a target of $100 million or more for next year.

And this concludes the Q&A period. Now I'll turn it back over to Robin for any closing remarks.

Thank you, operator. So many of you have expressed interest in learning more about our JDM business. And as such, I hope this session was helpful. Thank you again for joining us today, and we look forward to updating you on our progress in the new year. Happy holidays to all, and please stay safe.

This concludes today's conference call. Thank you for your participation, and you may now disconnect.