NXP Semiconductors N.V. (NXPI) Earnings Call Transcript & Summary

Okay. Why don't we go ahead and get started. Good afternoon. Welcome to JPMorgan's 50th Annual Technology Media and Communications Conference. My name is Harlan Sur. I'm the semiconductor and semiconductor capital equipment analyst here for the firm. Very pleased to have Bill Betz, Chief Financial Officer; and Jeff Palmer, Vice President of Investor Relations at NXP Semiconductor, is here with us today. For those of you that don't know, NXP is the world's largest microcontroller supplier. They're the world's second largest automotive semiconductor supplier. They just put up -- they reported a few weeks back very strong March quarter results, very strong June quarter results, strong visibility for the team. So gentlemen, thank you for joining us this afternoon. Really appreciate it.

Thanks, Harlan.

So let's talk about maybe sort of the near-term sort of supply-demand dynamics. You mentioned during the call that demand continues to outpace supply through 2022. Given the strong visibility, 80% of your products carry 52-week-plus lead times, strong bookings. You see supply improvements on a go-forward basis. When does the team see major improvements in the supply-demand gap? Is it first half, second half of next year? Are we looking at 2024? What's the team's view?

That seems to be the question that everyone is asking us this week and I'm sure for all our semi partners as well out there. I'd say what we said on the call was the second half will be larger than the first half and think about supply incrementally improving, but not solving that demand gap that we have. We believe demand is up here, supply is here, is incrementally improving. If we take out all of the double ordering and triple ordering whatever you want, we believe we have still not servicing about 20% of that demand as we go forward. So it's going to take quite a while to get to equilibrium and not in 2022 for sure.

I believe that -- if I look across your product portfolio, I believe that the highest volumes for the team is centered around from a manufacturing perspective, 55 nanometer, 65 nanometer, 90-nanometer manufacturing technology nodes. Overall, semiconductor industry CapEx is growing strongly, right, low teens percentage type growth rates. But the majority of this CapEx is focused on leading edge, right, not mature or specialty nodes, which is where, obviously, your sweet spot is. In fact, one of your foundry partners recently articulated a view that mature/specialty capacity expansion by the industry is only growing low mid-single digits over the next few years versus a demand profile that's higher. I guess, first question is, do you guys agree with this assessment? And does this mean that mature and specialty node supply will just remain tight for the next few years, especially with some of your outsourced foundry partners?

I'd say yes, and yes, for both those, and we agree with your assessment.

Okay. So on the front end -- and then to maybe make the matter -- the situation maybe even worse is that we cover the front-end equipment supplier, supply materials, Lam, KLA. Similar to your customer base, they're experiencing tight component issues, some supply chain disruptions and so on. And it looks like in the first half of this year, the equipment suppliers are undershipping their customers' requirements by about 10% to 15%. Is this causing a slower capacity expansion profile for you guys, either for your own internal capacity expansion efforts? Or are you hearing that from your foundry partners?

I mean we're hearing both, right? Equipment lead times have extended, both internally and from our foundry partners. Now the way how we allocate the semis is our first priority is in medical and health from any allocation and then followed by front-end processing to get them the chips, to get the equipment into the foundries. For us, again, internally, typically, maybe it used to take 12 months, now it's taken a bit longer, but a lot of the capital we have put in place over the last 12 months to make sure we receive it on time and so far so good.

From a supply side perspective, your channel inventories, your customer distribution inventories are sitting 35% below pre-pandemic levels. Given the backlog visibility, your supply commitments, your channel partners longer-term forecast, I mean, does the team believe they can improve channel inventories to your target levels, which is around 2.4, 2.5 months this year? Or is it really looking like, again, more like next year before you're able to do some sort of inventory replenishment?

Yes. So we would love to be at 2.4 months. Today, we're at 1.6, 1.5 for the past 2 quarters. Just to give you some math exercise, in order to get from 1.5 to 2.4, that would mean we have to ship in if we have the supply $500 million into the channel and assume none of it sells through. So we have a long way to go to get it back to 2.4. And again, this is one of those leading indicators to understand how much of your demand is really real. And this is one of our clear indications that we're far from making or getting to equilibrium.

I know you have a set of direct customers as well? Maybe some of them are on consignment, some of them are not. But do you have a sense do you guys track inventories at some of your direct customers? Are things as tight as they are with your disti channel?

It's similar in nature. I mean it's a little bit more harder to really understand the true end-to-end customer of those inventory levels. But we have another means of doing this through customer escalations, which are in the thousands, specifically in our industrial IoT set of business. This business is more supply constrained now than what it was 6 months ago, I would say.

Do we have any questions from the audience? If you do, raise your hands, and then we'll get a mic over to you. Give us one second.

I listened to GlobalFoundries earlier today, too, and I was kind of struck by their comment about a 5% CAGR on supply growth. When you look at someone like Applied Materials, I think they've said that they're now 50-50 in terms of what goes to leading edge versus trailing edge. So that sounds like it's in complete contrast because that number used to be higher, you see a much higher number at leading edge. So it suggests there's more going to the trailing edge, and everyone's got their own data points. It's just very confusing to kind of hear different points of view. Why do you think one group is saying one thing and another is the other because 5% CAGR seems like a very low CAGR?

Why don't I take that, Harlan? I think it could be in the definition of what trailing edge or leading edge is, right? So when we're talking about trailing edge, we mean in process nodes, 28, 40, 55, 90-nanometer. So if you look at some of our foundry partners, they define trailing edge as 28-nanometer and lower. So it could be -- that could be partially the challenge.

Any other questions? So -- is there any question? Okay. Perfect.

Just curious about the current situation in microcontrollers. I would love to hear what -- or maybe walk us through your capacity expansion plans to address the long lead times for microcontrollers, both internally and externally. And if that's something that's actually going to balance out between supply and demand, we're probably in a time frame but it's further out in '23?

So why don't I take that, Bill, and you can chime in if you want. So for us, the demarcation line between what we build internally, what we build externally is about 90 nanometers. So internally, we build any products that use proprietary processes, primarily mixed signal, specialty processes like silicon germanium, LDMOS, other mixed signal processes, 90 nanometers and above. From -- the vast majority, if not all of our microcontroller business, we build that externally. So we are constrained or tied by the foundry capacity that's out there with our partners.

Any other questions? So on the -- okay, that's great Yes.

I have a question on automotive, right? Clearly, the automotive industry is going through a huge change. If you look at regular auto numbers, they look extremely weak. On the other hand, there probably is a lot more content for you as well on the EV side. So can you talk a bit about the content opportunity when you move from regular combustion engine cars to both hybrids and EVs?

Sure. So if you look at the ICE versus xEV split, so average ICE car has roughly $500 of semiconductor content in total. And xEV, based on the research we've done, has roughly twice that amount. Now a large percentage of that doubling is in the powertrain, in the power discrete marketplace, where we don't participate. But a lot of the other functions in the xEV car actually have a higher content versus the equivalent in an ICE version, things like radar, safety features, things like e-cockpit for driver machine interface, comfort creatures -- features, things like that. There's also -- I'm not sure if we're going to get to this now, Harlan, or in a few minutes. But there is a fundamental structural change happening in the auto architecture right now, whereas currently today, most car architectures are flat point to point, almost kind of like a spaghetti, if you will, network, whereas we see a number of large European and large U.S. OEMs moving more towards this idea of a domain backbone, where you'll have -- we've been able to identify 5 different domains in the car. And what you'll have there is more powerful processors that manage that domain and all the leaf nodes below it. And those domain processors will talk over a gigabit Ethernet backbone to kind of a gateway network interface, very analogous to a network -- building network type of topology. And really, what that helps to solve is the problem with software life cycle management. If you think about the current architecture, there's no way car OEMs can update the software efficiently after the car has left the lot, right, or sitting at home. And they clearly want to be able to extend the feature set and the life usefulness of the car once it's left the lot. And to do that, they need to be able to have a better life cycle management of the software. And so that's what the domain processing allows you to do. We have a family called the S32 family, which we introduced a couple of years ago. It's just now starting to ramp into volume production on 16-nanometer FinFET. Currently, revenue of that business is about $300 million in '21, doubling to about $600 million in '24. And that's actually very, very early in the life cycle. The real acceleration curve of that business would be beyond '24 as more and more OEMs draw out this new type of architecture.

I guess to follow up on that because we've heard -- we keep hearing more and more about this zonal/domain, and it makes a lot of sense, right? I mean as the car becomes more compute focused, you want to have this sort of overarching compute/networking layer that sort of ties all the different nodes to the car together. The question I often get asked is given your leadership in automotive compute primarily MCUs right now, does the move to a zonal architecture actually cannibalize part of your MCU opportunity or no?

No, it actually is additive. So if you think about -- so we talked about domain processing a moment ago and that helping to solve the software life cycle management problem. The other problem a lot of the OEMs are having is how to deal with the weight of the wiring harness. And so what they're going towards is more of a zonal idea. So while they concentrate a lot of the functions in one zone, have more than just an MCU, let's call it a little bit more higher-performance MCU zonal processing, over time, those 2 vectors of domain and zonal processing will probably asymptote, right? And the way we see it is the number of microcontrollers in a car probably over the next 5 to 10 years goes from about 45 different microcontrollers today to up towards to 60. And then on top of that, you could have upwards of 4 different domain processors. And these domain processors are not just faster microcontroller. These are multi-core heavy-duty processors.

And there, you're also leveraging your leadership on the in-car networking side, right? Because I'm assuming that the zonal architectures, they don't use that sort of proprietary-based connectivity. It's more Ethernet-based connectivity, right?

Well, probably on the zones, the auto industry is probably going to continue to stick with CAN and LIN and FlexRay, which are very known and tried and true for those very known functions in a zone. For the domain communication, it will more likely adopt gigabit Ethernet. So you'll see them live together.

Okay. Perfect. Any other questions? So on the COVID -- the China COVID-19 lockdowns had a direct impact on the team back in the March quarter. You did a great job of recovering from that. But as the lockdowns had spread through March and April, your customers' supply chains and production capabilities were somewhat impacted. The NXP team, you guys accounted for this in your June quarter guidance. COVID cases in Shanghai have come down. Shenzhen remains at low levels. Country is starting to open back up. What's your assessment of the supply chain and domestic demand recovery thus far?

As you mentioned, right, clearly, the COVID shutdowns are having an impact with our customers and our suppliers. And like you said, it impacted us tens of millions of dollars in the current quarter. As we read the news, just like you all, in our double checks, things are opening back up, things are getting better. Related to Beijing, that's still -- nobody really knows. We had that experience, unfortunately, in Q1. It's fully recovered. We had a factory in Tianjin that impacted our revenue in Q1 by $50 million. That's back and fully running. We're very pleased with that, and we'll continue to navigate our way through these supply challenges.

And then in line with some of your diversified analog and microcontroller peers. I mean you guys -- and you mentioned it in some of your opening remarks about seeing a better second half of the year, right? And I know supply is tight. But there are -- by and large, most of your end markets are strong, but I do think that there are some pockets of weakness. And so how should we think about -- sort of on the second half growth outlook that you're expecting? How do we think about the puts and takes between your end markets as you look into the second half of the year, strong areas and weak areas?

Yes, as mentioned, right now, again, during the earnings call, we mentioned that 80% of all our orders coming in are getting scheduled 52 weeks out. So it's strong across the board. Some are greater than others, some are above that 80%, some are below that 80%. We're not ready because, again, it's all dependent on what supply we can get our hands on, and we allocate once a quarter. So the allocation changes every quarter because if anyone is building any sort of inventory, we redirect that material to a customer and need to prevent lines down. So to be honest with you, it would be very hard to tell you exactly the mix in the second half. We'll probably update that more on a quarterly basis as we go forward, Harlan.

Perfect. Let's turn to some of your product categories and your end market segments. So in automotive, you just talked -- there was a question from the audience about the content uplift in electric vehicle. And a lot of investors have continued to ask about the big difference in growth in NXP's automotive business relative to production, auto production, right? And I think the answer that you gave on EV, which is we've got good penetration in EV. We're seeing 2x content uplift. The other thing that Kurt has brought up and the team has brought up is also in this tight production environment, your automotive OEMs are prioritizing premium vehicles. So relative to that $500 sort of mid-range ICE-based auto, if you're -- what is the delta in terms of content uplift if you look at, let's say, premium ICE vehicle? Is that a 2, 3x content uplift as well?

It can be at least 2x to 3x.

So maybe in order to help us understand the delta in growth between your auto business and production SAARs, do you have a sense relative to pre-pandemic, what percentage of auto production today is focused on premium internal combustion and EV, let's say, relative to, let's say, pre-pandemic?

Yes, let me take that one, Jeff. So if we just look back in 2017, 2018, total car production or SAAR was 95 million units. Last year, it was 77 million units, a reduction of 20%. If we look at the xEV, so including hybrids, right, it was about 4% or 5% back in '18 that has now quadrupled to 20% of the 77 million just on xEVs. And then if you add on to the xEVs, the premium back then if you combine xEVs in 2018, plus the premiums, you get about 15%, has doubled. This year is expected to be 30% xEVs plus premium. And in 2023, xEVs by themselves will be 30% of SAAR. That is 1 year faster than what we thought just 6 months ago, driven by higher energy pricing, government regulations, on safety features and so forth.

So let me get the numbers right. So on the 77 million units, 30% of that mix, which carries 2x content uplift, 30% of that mix is xEV plus premium. This is versus 2017, which is -- what was the mix of xEV plus premium?

About 15%.

15%, okay. So basically a doubling of the sort of richer mix?

Correct. Okay. And then even against that and maybe this is more company specific, Harlan, is NXP-specific design wins that we've won in past periods started to also accelerate.

Yes. Right. So in electrification, we and the market tends to focus on your great position that you have in battery management, but you've got a broad portfolio. You've got power inverter -- you've got BMS, you've got domain controllers, DC/DC converters. How was the team leveraging its total solution strategy and sort of your customer relationships to drive higher adoption and more importantly, like higher dollar content per EV?

Yes. So maybe a great way to start with that is to talk about what our solution is in BMS, it kind of gives an example. So our solution for a battery management system is a system approach. We basically offer precision analog devices that sit out on top of the battery pack itself. We connect those precision analog devices together to a processor. And on the processor, we run load balancing software. And the load balancing software manages the charging, discharging of the battery pack as well as the health and safety of the battery while it's in operation. So our approach is very much a full system-level approach. And that is really -- you'll hear this from us many times in different parts of our market. Our customers value the ability to bring multiple different pieces of silicon together, put them together, take care of all the kind of the subsystem interfaces, if you will, provide some firmware wrappers to make them be able to get to market quicker.

And then the other sort of trend in BMS has been sort of the move towards -- moving towards a wireless space BMS system. Does the team working -- does the team of -- portfolio as a team putting effort in R&D in that space?

We have R&D efforts. Nothing to announce today, but I'd say watch this space, but we do have some efforts in that area. Yes.

Perfect. And then in automotive radar, this is another area of pretty strong leadership for the team, pretty comprehensive portfolio. You've got your RFCMOS 1-chip solution, right, integrated transceiver and compute capability on a single piece of silicon. Help us understand the benefits of the move to 77 gigahertz? And more broadly, the team expects to drive kind of 20%, 25% revenue CAGR over the next few years in this segment. What is the team's assumption on unit penetration, number of radars per vehicle as we move forward in time?

Yes. So again, this is another example of a system approach. So if you think about a radar system, it's 3 building blocks. It's a transceiver where the RF signal goes out of the car comes back in. That transceiver device is connected to a back-end processing engine. It processes the signal then tells other parts of the car what to do. Again, a system solution. We started working on 77 gigahertz probably back in 2016 when the market was still just thinking about 24 gigahertz. We knew that we were not going to be able to intersect the market at 24 gigahertz. So we basically said, let's go over the horizon and try to intersect the market at 77. And so our solution -- we have 2 different types of solutions. We have silicon germanium, which gives you higher power, higher -- better distance, if you want, that's usually for front-end radar. And then we also have RFCMOS and the RFCMOS gives you good distance, good power, but it also gives you the ability to integrate that transceiver into the processor itself, so you can build a single chip design. And so the way to think about the radar market, just kind of from an algorithm or how to think about how it evolves, think about it as a 3-way multiplicative effect. More cars with radar, more radar nodes per car. And on each one of those radar nodes, more functionality as we go through time. It's going from standard radar today to imaging radar in the future. At each one of those multiplicative points, we get a higher dollar content.

And just to add to what Jeff. So today, I'd say, on average, there's probably 1 system, 1 node going into the car. By 2024, there'll be 2 systems. In 2027, 3 systems. In 2030, beyond 3.

And so today, our radar business is about $600 million in '21. We expect it to double to about $1.2 billion in 2024, and that's already with design wins we've already been awarded. We just have to execute and deliver to.

Any questions from the audience? Industrial and IoT team is a top 5 market share leader in industrial microcontrollers, in line with your Analyst Day guidance. Your industrial and IoT segment is driving -- I think it's driving the fastest year-over-year growth alongside your auto franchise. Last year, it looked like about 60% of industrial and IoT was compute, like your microcontrollers, your crossover MCUs, your processors with the other 40% being connectivity, analog, security attached. Wondering how this mix is changing, especially as your customers focus more and more on your system solutions and reference platforms. And from a market perspective, what's driving more of the growth this year? Is it industrial and edge compute? Or is it smart home? What's driving the strong demand for IoT?

Let's talk about what industrial is from an end market perspective. So every company has different end markets definitions for industrial. For us, industrials is a long tail, tens of thousands of customers, primarily serviced through the global distribution network. There is no one customer or one product type in industrial IoT that makes up more than a few points of revenue. So very, very diffused product set and customer set. Probably about -- if we had to be -- really put into different spots, about 60% of that revenue is industrial. So factory automation, building automation, traditional true industrial-type applications. The other 40% is IoT or things like home electronics, smart home type systems, some wearables, some health, some point-of-sale type of items in there as well. I think you've got it pretty well done, Harlan, in terms of the mix of business there. It's primarily a processor business. Over the last couple of years, we've really driven the sales team to the minute they win the processor opportunity, they pull along the PMIC, the power management for analog attached because in many cases, those 2 devices are a match pair, if you will. We're starting to find a lot of our customers, they may be experts in their end markets, but they don't want to have to put together and stitch together all the different components to make a subsystem. So what we started to do is, first, you win the processor because that's where they're going to write the software to. You pull along analog attach for power management. You then can pull along options for connectivity, whether that's wireless, Bluetooth, ZigBee, NFC as an opportunity. And then you also -- certain customers want security. Security is much more of a concern now. We have certain processors where the security is embedded in the processor and somewhere the security is discrete, and we can offer both. And really what's distinguishing us, I think, with our customers, we're able to bring all those products together into a secure connected edge processing solution. And we have customers who they don't want to spend the time picking different sockets and stitching them together. And that's really been -- I think when we bought the Marvell asset a few years ago, that was kind of the missing building block that really completed that story.

On -- moving on to your mobile segment. In mobile segment, strong leadership in NFC secure wallet. I think attach rates are closing in on 50%. Now you're extending that leadership right with next-generation UWB capability, right? The team has laid out some OEM platform wins in mobile, in auto with your UWB solution. So how do we think about sort of the long-term attach rates of UWB in auto, in mobile, in IoT? Is it -- does the opportunity look similar to NFC?

It does look similar. We don't think it will take as long to get to this 50% attach rate as it did for NFC. For NFC, it took us probably 7, 8 years. Really, it started slow. We think it will be a little bit faster, not dramatically but probably not 7 or 8 years. But what's really interesting, our ultra-wideband solution is an intersection of the mobile market, the automotive market and the industrial. And really what we're offering is secure access via the phone. So the phone becomes your virtualized key or your virtualized identity and then we'll be able to enter -- you'll be able to use it to enter the car, to start your car or enter your home or you're building what have you. And so that's really dependent on not just the ultra-wideband RF device with the security that's in the mobile wallet. That mobile wallet, think of it as owns as a security complex in the phone that we're attaching other functions, too. And so last year, we were probably just under about $80 million on ultra-wideband revenue, very early days. We're focused primarily on the smartphone side of things in the Android market because the other large operating system vendor has their own ultra-wideband device, where we also offer the mobile wallet. We think that business will grow to about $400 million by '24. So the first couple of years will be primarily the mobile phone. Every major auto OEM who is working on ultra-wideband-based secure access solution, we've won that design. But the design to revenue cycle in automotive, as you know, are 2 to 3 years. So it won't be till late in that horizon '23, '24 before you start to see some of these models out in the marketplace. We can talk about BMW and Hyundai Kia, who've adopted our technology, but there's quite a few others as well.

And then finally, before we talk a little bit about the financials is your communications infrastructure business. We all know NXP for its leadership in LDMOS and power, right, in comm infrastructure. But I think since then, I mean, you have built out the portfolio, both on the power side, but also on the compute side as well. Talk about the comm infrastructure portfolio and the leverage to 5G and the leverage to the 5G build-outs, let's say, ex China?

Like every company, you have a kind of a catch-all end market, if you want. The comm infrastructure is a bit of our catch-all. To just kind of educate everyone, what's inside that end market is really 3 sub portfolios. It's our RF power business for the base station marketplace. There is a legacy multi-core processor business that came to NXP via the Freescale acquisition. And then there's our security card businesses that just we didn't know where else to fit it, so we put it there. I would say from a growth perspective, we said really focused on the RF power business. We think we've got about 40% market share. We think that business will grow at about a 15% CAGR from about $500 million last year onward to 2024. Really what we offer there is a full suite of products depending on what frequency and power output the carrier needs. So a lot of people think that LDMOS will go to 0 because it's old and nobody likes it. The reality is LDMOS is very, very well known, operates very well, very efficiently up to about 3.5 gigahertz. From 3.5 gigahertz to 6 gigahertz, we offer a GaN solution. It's actually a hybrid GaN solution with LDMOS and GaN in a combined module. And then for our millimeter wave above 24 gigahertz, we have a silicon germanium family. So we can offer a full spectrum of products depending on what the carrier wants to deploy. So right now, tactically, we're seeing decent traction in North America with the C-band build-out, seeing a little bit slower build-outs in Europe. But as you all know, the base station market is very, very lumpy, very choppy. You'll have a couple of good quarters, then you'll have a couple of pause quarters. It's just -- that's the nature of that business.

Perfect. And then on the financial front, the team has done a great job in driving gross margins, driving even better leverage to drive strong operating margin performance. It does seem though as you move into the second half of the year, there's a little bit of a ceiling on gross margins because utilization benefits are sort of maxed out. But the team has been driving incremental gross margin fall-through of around sort of mid-60% over the last few quarters. So I guess the question is, is that how we should think about your gross margin profile for the business as we move into 2023 and as you unlock more supply growth that on incremental revenues, the fall-through would be around sort of that mid-60s range?

Yes. First off, we're not going to guide 2023 gross margins. Just to clarify that. But the way to think about it is our long-term model is 55 to 58, as like you said, we're approaching 58. We said recently on our earnings call, we'll be in this range where we are plus or minus 50 basis points. What really gets us to that 58 consistently and above 58 is really our new product introductions. And that's just going to take a while because we grind it, we have things falling off. We have new products coming in, and then any given quarter can move around a bit. But when I talk 58 and above, that's more of hitting it every year. But yes, we do get -- expect it, a little fall through on the higher revenues, for sure, what you're saying, just scale over our fixed costs. But again, there's so many moving parts inside our gross margin at any given quarter.

Is there any way for you to quantitatively, qualitatively sort of profile your design win pipeline, right? Because as you mentioned, mix is going to continue to be a big factor of the gross margin. So if you look at your design win pipeline, I mean, I don't know what percentage of the wins are above sort of corporate average versus at corporate average.

I think maybe an important thing to realize is back in 2017, about a year after we did the Freescale merger, there was a decision made that any new product that was to get funded had to be at least or above corporate average. So the bar -- and the bar continues to get raised, right? That's what Bill was talking about. There's a pipeline of design wins that are at least at corporate averages if not above, but they'll take time to grind into the revenue stack. Simultaneously, you'll have older products, legacy products that maybe were a little dilutive to margins will start to roll off. It will be a grind. It will be a little like watching paint dry. I don't think we're going to give you a dollar size of our design win pipeline. As you know, Harlan, it's not one of my favorite topics. I would say last year, we had record design wins across the board. So far this year, things are going very, very well. We're very encouraged by the key accelerated growth drivers, continue to get great customer traction, radar, BMS, the zonal domain processors, the crossover processors, the connectivity, all going very well. Ultra-wideband design wins going well, things are going very well, but it does take time from design win to revenue in our business.

Perfect. We're out of time, but I wanted to get 1 more question in capital return. You guys have done a great job, 100% free cash flow return on a go-forward basis. How do we think about the balance between dividend growth versus buybacks?

Yes. Our capital allocation strategy has not changed. Our first prior already is organic growth. So as you know, a 1% change on the top line is better than 1% change in the margin. We'll continue to focus on growing the business. The remaining part, after we source the funds needed for the company to grow, we plan to continue to drive our dividend towards 25% of cash flow from operations. Last year, it was at 18%. Related to the buybacks, we returned 121% when you take the dividends plus the buybacks in Q1 and then all -- the trailing 12 months was around 185% when you combine both. Think about as us going forward, that we will continue to return 100% of our free cash flow back to our investors.

Great insights, guys. Thank you very much for your participation.

Thank you.

Thank you.