ON Semiconductor Corporation (ON) Earnings Call Transcript & Summary

Great. Well, thanks, everyone, for joining us today. Very pleased to have with us ON Semiconductor. We have CFO, Bernard Gutmann; and in Investor Relations, Parag Agarwal is with us today. So welcome, guys.

Thank you, Craig.

Okay. Just before we get started, just want to remind people. For disclosures, you can see them at www.morganstanley.com/researchdisclosures.

So with that, perhaps we can just start with the current environment, Bernard, just kind of what you're seeing out there? Any particular end markets that are driving the rebound?

Thanks, Craig. I'll be glad to answer that question. Obviously, I cannot give you a mid-quarter update, but I can tell you that we did see some nice recovery after a big decline in automotive in the third quarter, and we guided for that to continue being up in the fourth quarter. We expect automotive in terms of units to be down in that high teens for this year. And for us, we expect to be down 10% or less. So that basically confirms that we keep having a content story where our revenues are down high single digits, less or are up high single digits, more than the end market. We also have seen, in general, some optimism, and I believe that's reflected in all the sell-side models that show that 2021 points to a nice recovery, and that would be kind of normal when you look at what has happened with -- in previous down cycles. Current sell-side model show about a $5.7 billion revenue level or about a 10-plus revenue recovery in 2021. And when we look at our secular growth drivers, we think that, that is a very plausible scenario. We also -- I can point you also to other macroeconomic factors like global GDP and PMIs that are also showing some very nice signs of recovery going to a good 2021.

Got it. Anything to note just from a supply chain perspective? I know you guys brought inventory down the channel last quarter, but just what you're seeing from a lead time or inventory perspective more broadly?

So indeed, you're correct, we did bring our lead -- our existing inventory down. And now we're at the -- within our comfort range at the lower end of it. So we're in a good starting position. We don't have excess inventories in the pipeline from the channel point of view. We still have a little bit high inventories on the internal front. We improved them from Q2 to Q3 levels from 140 days to 133. Nevertheless, we hinted that the fact that in Q4, we expect to bring them down a little bit more, we're getting us closer to our target range of 115 to 120. Lead times continue being pretty stable. We have seen a little bit of elongation in those, but still all within normal levels. We have seen a few tightening going on foundry and subcons that is showing signs of also of good recovery.

Got it. Well, maybe we can segue over to gross margin. I mean it was -- it's challenging through the down cycle in 2019 and '20, but now it's rebounding. So if you can kind of just walk through for ON, what are the most important levers as we think about the recovery in the gross in your margins?

Sure. So they are the same levers that we enumerated at our Analyst Day with the difference that having had 2 years of down in revenues versus an expectation of 5% up, we expect to see even more leverage on the revenue side. We are running our facilities, our factories right now at somewhere around 70%, way below our sweet spot of 85%. So we expect a sizable margin contribution as a result of increasing our utilization and running our factories at the higher level. We are still subscribed to the 43% target model we have out there at $7.1 billion in revenue. And as we get closer in that revenue level, we will be showing some really nice rebounding on our gross margin. Having said that, we have also -- because of the fact that we invested in a very heavy way, CapEx-wise, in our capacities, and revenues didn't materialize to a level that we thought, we have also done some self-help, announced over this last year, the intended sale or closure of 3 manufacturing facilities, 3 manufacturing assets, mostly subscale, 6-inch facilities. The savings of which, by the time we're all done, should be about $75 million, so 1.3%, 1.4% of gross margin. The Rochester one is a closure, and that will occur as we speak in the next week or so. So by early next year, we should see about $50 million annually of savings flowing through our gross margin for that particular facility. The other 2, the timing is undefined, and it depends to some extent, to a big extent to what kind of selling arrangement we make in terms of manufacturing services agreement and timing of the transfer. But having said that, when we're done, we should generate $60 million from these other 2 facilities. And these savings of $75 million that I'm talking about is mostly is simply just the elimination of the fixed cost infrastructure that comes from eliminating a site. It doesn't incorporate the additional benefit of better factory utilization on the recipient factories that then we should be getting from that. The third element to the same one that we had in the Analyst Day is mix. We do continue expecting to get a tailwind from mix by virtue of higher growth on our end markets that we're targeting, all of which command better than corporate average gross margins, while the ones that we are deemphasizing and expect to decline in revenue over time, those have below corporate average. So it's still volume, the self-help on manufacturing and mix, with volume or revenue becoming now a bigger component as we're starting from a much lower point.

Got it. Thanks for those details. There's been some investor talk about even perhaps a fab-lite. So can you maybe just talk about ON, like what's the right balance for you as you look to, as you say, shut down sub-scale factories, which makes sense, but how do you view kind of the manufacturing strategy, bigger picture in terms of what you own and what you can outsource over time?

So our strategy has not changed on that front. We believe that with the size and scale that we have, we can get some really good benefits for manufacturing certain parts of our business. We do not manufacture parts where the economic investments in capacity surpass the economic benefit of doing that and that in most cases, is for either very advanced nodes on the fab side, 22 nanometers, 45 nanometer, 65 nanometers, we all outsource all of that. We also outsource really old stuff where it makes no sense for us to use our footprint and our floor space for these very old assets in sunset dose by outsourcing. And then we have a third element is mainly flex capacity where existing products that we do in-house, we also like to have a way to outsource. Having said that, when we look particularly at the power discrete space, which is one of our biggest technologies and products that we run, we see our competitors in that space doubling down on investments in manufacturing capacity, primarily on 300 mm side. So we believe that we will need to be along the same lines on that front. So the fab-lite, for that piece of the business may not be appropriate. However, for certain pieces, and by the way, today, if you look at our total mix on the front end, we are outsourcing about 1/3 of what we're selling and making 66% internally.

Got it. And that's probably a good segue to the Fishkill 300 millimeter that you'll be ramping. How does that change things in the coming years in terms of how meaningful it will be for ON from a percent of production? And what that you can grow into with Fishkill?

Sure. So as to remind people on East Fishkill, we stroke a deal with Global Foundries to buy their 300 millimeter East Fishkill 45 and 65 nanometer production facility in Upstate New York. The deal, we made an initial payment, but we don't consummate the deal until beginning of '23. Between now and in that time frame, we have a foundry arrangement with Global Foundries that gives us the benefit of redesigning our existing production and new production from 8-inch to 12-inch in qualifying all of the different required technologies and customers to these -- in this new facility without having to suffer the underutilization charges that come with the mega fab like that one. With the intention right now in this interim foundry relationship, not to maximize volume, but to maximize the qualification of all the required technologies. So by the time we pick up the facility, we're able to run it at the very nice utilization levels. When we did the announcement of the deal, we talked about this facility being able to generate up to $2.2 billion of revenue. That's at 100% utilization, which we normally don't run. So if you take an 85% utilization, it's probably a little bit under $2 billion. And that compares to, let's say, in 2023, '24, $7 billion of revenue. So that gives you an idea that it is a pretty meaningful-sized facility that will allow us to get some really nice volumes and benefits to it. Also as a benefit, which we are seeing right now, having that flexibility of having this capacity coming on board gives us the ability to be able to do these musical chairs within our existing footprint to close down the less efficient and lower geometry facilities as kind of a side benefit. So it is a -- we believe it is going to be a key transformational fab. It will also to allow us to have our own sandbox to design stuff in 45 and 65 nanometer, which we believe, with us being more lagging edge technology, is advanced nodes that will give us a nice runway for many decades to come.

Great. We touched on gross margin, maybe we can touch a little bit on OpEx. You guys had kind of signaled into Q1 that OpEx will come up as variable comp comes back. How should investors think about it as you progress through the year? What are the puts and takes to spending levels?

So our goal -- our long-term goal for OpEx is, we're still committed to the 21% of revenue. Obviously, as we step up variable comp, which has been virtually nonexistent for the last 2 years, we will see a negative impact as a percent of revenue. We will spend at a lesser pace on OpEx than the revenue growth, and over time should be coming down as a percent of revenue. When we look in absolute dollars for '21, we expect to see the step function increase in Q1 and then after that, pretty stable and relatively flat compared to what we will see in Q1, obviously, with expected revenue growth over time. Our R&D investments, we keep that in that 11% to 11.5% rate and expect no changes on that with a heavy focus on things like silicon carbide, ADAS, LiDAR, East Fishkill, industrial IoT, servers, 5G. Those are kind of the main areas where we're going to spend. But long term, our goal is to achieve the 21% that we have enumerated in our Analyst Day model.

Got it. Well, you touched on a couple of interesting technologies, so maybe we can segue over to the automotive end market and would love to get an update of just how you're doing. CMOS image sensors has been an area of strength for the company, and you have really nice leadership. What's the growth looking like and as well as the competitive landscape?

Competitive landscape has not changed. It is mainly consists of us, Sony and OmniVision as we look -- directly adjust image sensors. We have complemented our offering by also now having LiDAR and RADAR and ultrasonic, which we have had for a while as part of the offering, which these competitors -- other competitors don't have, which we believe will give us a nice edge. Competitive landscape continues being the same. We have a first-to-market advantage in automotive and industrial and pretty strong technology differentiators, including the barrier to entry of software drivers that are needed to connect our devices with the microprocessor. So we expect the growth to continue being in the double digits for this subset of products, and we have also augmented our offering by now offering the power management that will do the powering of the microprocessor that the ADAS -- complex ADAS systems are required. So pretty nice business, it's -- and the content keeps going up as you move from level 1 to level 2 to level -- all the way to level 5. So it's quite exciting, and we -- our design win activity tells us that we continue having a pretty good market share in that space. It's pretty good.

Got it. And silicon carbide was one area you mentioned from an R&D perspective, you're focused on. A lot of attention on that market this year as kind of EVs reflect. And I think you have a good perspective because you do IGBTs and silicon carbide as well. So what's your sense in terms of one technology versus the other kind of where things are? What's customer interest level for silicon carbide?

We have seen very good interest in silicon carbide. We talked about a design win with a global -- with one of the top 5 global OEMs in automotive, 401 platform in silicon carbide. That should generate revenues in the back half of '21 and in a more meaningful way in 2022. Having said that, we still see a good amount of interest for the older IGBT technology. We expect for next couple of years that it's still going to be dominant. And as time goes by, silicon carbide, which has some pretty compelling performance advantages, will be moving more towards that technology. But we believe that the 2 will be going to at the same time for a while. We were asked if that's a cannibalization of ours, and we say it's probably a welcome cannibalization because the delta, the ASP that you get on silicon carbide is pretty nice. We have talked about the -- one of the key elements for silicon carbide is the raw wafers, the substrates. And we have worked on the R&D side and now on the manufacturing process side to have our own capability of doing raw -- of doing substrates, but have not spent the CapEx on that because we want to really assess whether it's worth that CapEx investments to have the vertical integration on that front. And the good thing is that there is several outside alternatives available, and we have the internal capability also available. So it's going to be more an economic choice for us at the appropriate time.

Got it. Do you feel like is there a certain period of time that, that will be a more important decision in terms of to move forward with CapEx for outsource on the wafers?

Yes, it will. To me, it will depend on the economics of the overall what other mainstream silicon suppliers will decide to do whether we enter into this market or not, and there will be ample supply, and it becomes more commoditized like regular silicon or if it becomes still the big differentiating factor, at which point of time, we would trigger the CapEx investment.

Got it. Great. Maybe we could segue over to the industrial market, which is obviously much more fragmented and there's a lot of different subsegments. And so which one of those segments, as you look at your industrial business, do you feel like has the best kind of secular growth prospects?

I would give you 3 different secular growth elements. One, and it's fairly pervasive across a lot of our industrial offering, is automation. For automation, you need a sensing element, and that's where our image sensors come into play, and you also need a power management to power-manage these automation features. So we're seeing a significant push in that direction. It can come in many different flavors, in as of late, we're seeing a big push towards warehouse automation as our e-commerce customers are wanting to automate more and more of these. So warehouse automation, robotic delivery, all of these things are giving us some really good tailwinds. But it can also come in things like factory automation, machine vision, building automation, smart metering, robotics in general, drones, all of these things are part of the automation team. The second driver in industrial, which is also fairly pervasive across many of the submarkets in industrial, is energy conservation. And we participate there in 2 different types. One is directly in providing chips that go into alternative energy generation in solar farms, wind farms. And the second piece is through providing chips that efficientize the use on the running of motors. With our portfolio of power discretes, we have a pretty strong medium, high voltage and even low voltage products that can fit that space. The third tailwind -- or the third secular growth driver completely separate and different is medical. It is a business that's growing at a very, very nice pace. It is also very good gross margin, has similar characteristics as the rest of automotive and industrial, i.e., very long design cycles with regulatory approvals. But once you're in, it's a very nice, call it annuity, for many years to come. It used to be 1% of our revenue, now it's passed 3% in Q2 is growing. And you look at the aging global population as well as the movement towards personal diagnostic tools, it suits our portfolios very well. So those are the main drivers. We have other small -- other businesses within that industrial, but they are probably not as -- they don't have as stellar of a secular growth driver.

Right. And maybe just for a minute, going back to silicon carbide, I mean, EVs, of course, get a lot of the attention in terms of the size of the market, but there's also applications in alternative energy and industrial broadly, what's the customer response to silicon carbide in kind of the industrial space today?

And I'll let Parag help me with this one, but a couple of comments. One is, actually -- we actually started first selling silicon carbide into industrial before we went into automotive. So we have been selling it for a couple of -- a good couple of years. It is -- it provides a sufficient efficiency difference that it has nice and good applications in industrial. Parag, any more details that you want to give?

Yes. Yes, Craig, any application where there's a need for efficiency, it could be motor control, it could be alternative energy or it could be the chargers for EVs. Silicon carbide is a good material to use, and we have seen a very good traction in EV market worldwide, and we expect that growth to continue, at least for the next few years. So we are very upbeat about that market.

Great. Thanks. Well, maybe we can move over to the comm market. First, I'd like to just kind of start with Quantenna, just how that business has kind of worked into the fold, and I know not just for comm, but there's -- part of the thought process was taking some of those products and really addressing automotive industry sales. So any update there?

We are still on track. We're on track to deliver new products that will fit into the industrial IoT, where we're going to combine the Wi-fi capability that came from -- the premium Wi-Fi capability that came from the Quantenna acquisition with our owner, Bluetooth and ZigBee and put that and on the -- in our low voltage capability to provide the system on a chip that will be targeted for industrial applications, industrial IoT. And when we acquired Quantenna, we said that, that's an 18 to 24 months from the acquisition, and we're on target to deliver goals. We have completely integrated that acquisition, have merged the R&D team. So we have teams that are now jointly developing these industrial IoT applications.

Got it. And how about 5G? I know that's a growth driver that you guys highlight. Certainly, the whole industry has seen headwinds from Huawei recently. But if you put the Huawei headwinds aside, how is the 5G market developing for ON? What's some of the content and growth that you could see in that space?

Sure. So in 5G, we provide -- no surprise -- power management of the base station or the infrastructure side of it. We have participation in content and contracts with pretty much all of the manufacturer of base stations. So the fact that the share is shifting from one to another as a result of these geopolitical pressures in the long run should not be a differentiating factor as we have, again, good presence in all of them. The content per macro cell is about $150. It actually has increased even more as we have now provided beyond power management some additional analog parts for these. So it's pretty high content, and we see that as this infrastructure is deployed over the next 5 years, so we should be able to get our fair share of that market and get some really good tailwinds in that revenue. Huawei, that you -- to close on that, was a meaningful customer. We never gave details about how much, but we don't we have any customer that's greater than 5%. And if we look at our guidance for Q4, the Huawei guidance was 0.0%. So it can’t get any lower than that, and it's only up from there.

Got it. And then I think servers is another area from a data center perspective that's been a nice growth driver. Any update there in terms of your positioning for server power management?

Yes, very similar to your point with the 5G. It is the power management of servers. The content there is around $60. And with the addition of a controller for the next generation, we can add another $15 to get to about $75 per server. We play in both the enterprise and the cloud side. So we have good coverage from the customer point of view. It is a new business for us, or relatively new. We didn't participate in that before we acquired Fairchild. And we have been able to -- with the performance and the cost and the quality, we have been able to elbow our way into that market even, say, call it, late-entrant and have been able to get and continue growing share to quite a meaningful one as of right now. So we expect that, that will continue, and we are well poised to continue delivering on those power discretes.

Got it. All right. Coming up, we have about a minute or 2 left. I did want to address just China broadly, not just for ON, but the semiconductor industry has been very topical in terms of in-sourcing. I think in the past, you've said markets like consumer or computing, maybe a market that goes -- some business goes to China, but any updated thoughts in terms of what you're seeing in China? And what the implications are for your markets?

So for the mainstream, power discretes and analog and sensors, we are not seeing increased competition or increased in-sourcing. It looks like these are technologies that are difficult enough that throwing money at them is not the only thing that's needed to be successful. You need significant cycles of learning and building a business one by one. So at this stage, we're not seeing any incremental in-sourcing push from -- in the China market. We continue competing successfully in the -- even in the consumer and client computing space with our products competing against Chinese indigenous manufacturers. We have a lot of our manufacturing for these products also in China. So therefore, we can compete easily on cost. And our quality is superior, so we get that added edge that keeps that business from being in-sourced completely. So we don't think there is any major headwinds that will be coming from this in future.

Okay. Well, excellent, I think we are coming down on the half hour here. So Bernard and Parag, really appreciate the time spending with you today to dig into ON.

Appreciate. Thank you, Craig.

Great.

Thanks, Craig.

Thanks. Have a good day, everyone.

Bye.

Good bye.