ON Semiconductor Corporation (ON) Earnings Call Transcript & Summary

All right, everybody. We'll get started with the next fireside chat. We have both the CEO, Hassane El-Khoury and Thad Trent from onsemi, the CFO. And so guys, thank you very much for coming down to Dana Point. Why don't we start with just kind of a lay of the land cyclically and where we are. You guys have been, relatively speaking, more cautious on the slope of the recovery -- well, the length of the downturn and the slope of the recovery, and that's proven to be wise because we've had many companies guide for an improvement and then have to backtrack on that over the last nearly 18 months. So as we stand today, it sounds like things are getting a little bit better and maybe the question is on the slope. So how are you guys feeling about your business today?

Look, it's -- I think the answer is I have to give it in relative terms. When I talked about the last few quarters, we're seeing signs of stabilization. I take that as a positive stabilization from where the business -- the business was prior to that. So stabilization is a positive. We're still seeing it that way. You can talk about visibility is slightly better. we still have turns into the quarter. But if you think about where we are today from the same time where we were last quarter, we need less turns. So on a relative basis, I think you can talk about that being an improvement. We've also said the second half of this year is going to be better than the first half. So that gives you also a stabilization. And I'm very cautious about not talking about a recovery because it's not really your normal cycle where after X, A, B and C happens, you get a recovery. You typically will have an inventory burn and then you have a replenishment cycle and then you have a demand. Well, we're not in a typical cycle because some of the factors that are changing the demand environment are not really only market-driven. You have a lot of the geopolitical overhang on it. So that's how I remain looking at it. And like you said, we've been more right than wrong as far as how we look at it. We're going to focus on the things that we manage and the things that we control. And that's the operational excellence of the company, whether it's managing inventory, managing customer relations, managing new products and managing cash flow. Those are all things we can manage to in that uncertain environment.

So when you talk about the turns percentage decreasing, the turns needed to make your guidance and those sorts of things. So how does that metric compare to what you would define as a normal period? And I know it's been a long time that normal has occurred, but not the time in COVID where you didn't need any turns or the time a year ago where you needed all turns. Are we closer to the normal or still closer to the trough?

Look, it's hard to say just forget about the normal because of the market kind of cycles that we went through. But even for us as a company, we haven't really gotten to a normal because you can't compare to where we were prior to 2019, if that's kind of your time frame because we're a very different company, totally different mix, right? So our mix of products has been different. Our market approach has been different. I mean, right now, over 80% of our business is auto industrial. It used to be, I think, about 60% back then. So I can't give you like how does it compare to a normal because we haven't landed on a normal as a new company with the new mix and the new exposure to the markets that we have. So that's the fairness that I would say. That's why I keep answering in relative terms because those are, relatively speaking, their comps. But from a longer period, that becomes a little [ mighty ].

So if you had the time when you had 0 visibility and everything was turns or the extreme of that to what you would desire -- realistically desire. So take the historical anchoring out of it.

My desire is full visibility [indiscernible] quarter.

[ Hence ] realistic.

No, I think we're not where I think we could be. When you have auto and industrial business and so on, you typically have a longer visibility, which is why we like that market. So typically, we will -- there's still improvements to be had there, but we're not going to have 100%. Those days during the shortages and so on, I would consider that a realistic normal.

Right. So how are -- you mentioned the geopolitical side is abnormal, and I think we would all agree on that. How are you seeing the tariffs in the geopolitical side affecting your business? Are you seeing more pull-ins, pushouts, people just with paralysis because there's just too much change, and it probably differs across geos and end markets? So whatever color you could give.

So I can give you from what we see for us as a company. We haven't seen any signs of pull-in. Actually, the growth that we've seen in certain markets or certain submarkets like our business in China is doing well. We could tie directly to commentary I mentioned in the first quarter about, for example, we went to the auto show, we saw all the platforms that we're in, and those platforms are going to ramp. Well, fast forward, the platform ramped, and we delivered the results that we expected. So the results that we've posted and our outlook all have been what I would call a natural demand, meaning consumption side at the end markets. So what does the tariff overhang kind of do? It's more of the paralysis. And then what do we do about it is offer customers the flexibility. And let me break those 2 into kind of what I mean. From a customer perspective, if customers are placing order, the last thing you want is the customer places an order when there's a tariff and then the tariff goes away and now their COGS are higher. So they're waiting until the last minute when they need their product to say, okay, now ship it to me because the window between order and build and sell is much shorter. So it's less likely. So that's what we call the paralysis. And at the end of the day, you still need to end consumer confidence to increase a little bit to drive that demand. So that's from the demand environment and customer behavior has been a little bit short. You hear some of my peers talk about short lead time orders and so on. That's all within that realm of customers don't want to -- don't have the visibility, so they're just doing very short term. From the -- how we're doing for customers is really the flexibility. So think about it this year. We have a pretty broad manufacturing footprint. We do about 65% of our product internally with a broad global footprint of manufacturing, whether it's Japan, Southeast Asia, North America, Europe. So when customers see that, they're more comfortable with our ability to manage as they change supply chain to navigate through their tariffs. So we offer that flexibility because of our global footprint. And that's been seen as a favorable competitive advantage for customers because they don't have to make the choices now on us as a supplier. They don't have to worry about us as a supplier. But while we can help them figure out how to change their supply chain to have the least amount of impact where we can overlap with them. So that has been a very collaborative approach with customers. But at the end of the day, we do need stability because customers need the stability. Nobody is going to wake up and move a factory overnight just on a -- without the regulation being stable or at least visibility on what it is and it is not.

How is inventory at your customers in the channel, et cetera, enough of stuff internally. We can talk about that later. But how is inventory playing out as kind of that stabilization dynamic? Is it still a headwind? Is it normalized?

So inventory, I wouldn't say there are some pockets that are still draining inventory. But the first order, it's less than it was. So it's continuously improving. In certain areas, we are shipping to natural demand. So I wouldn't say this is a regional answer or an industry answer, meaning it is a customer-by-customer basis. In the case of automotive, there are some Tier 1s that still have inventory that they're trying to burn, while others are already down to 2 weeks of inventory, which, in my opinion, is kind of marginally dangerous. So -- and that depends on what OEMs they were exposed to. And again, at the end of the day, it's demand. So the inventory burn is tied to demand. So we're still in that. We're monitoring that with our customer. We're monitoring their ordering patterns, but that's definitely improving.

Yes. And I would also say that the industrial end market is probably more close to natural demand at this point. It was the first to get soft, the first to stabilize. And I think we're probably shifting closer to natural. I think in auto, there's definitely pockets as Hassane described. But probably by the end of this year, hopefully, we're back to natural again across the board on auto as well.

Any sense at all about anybody in kind of restock mode? Or is that volatility you're seeing in end demand and geopolitics and all of that, precluding anybody from restocking?

Unless there's an imminent ramp like we've seen in some of our customers that are ramping like in automotive in China, EVs or some of the medical customers, nobody is stocking. So it's more on behavioral, which is -- which -- talking about normal, which is normal. If we have a ramp, then we will ship ahead of the ramp to stay ahead of our customers. That's a normal, call it, restocking. But the restocking that I think you're referring to, which I call the replenishment cycle, that has not been happening -- that has not -- in my opinion, that has not started yet.

Do you think there's been a structural change in the amount of inventory that the customers will hold? If I asked you that 3, 4 years ago, the argument would be, will the Tier 1s and everybody in the auto ecosystem hold more because they learned their lesson on shortages. From what you just said, some of the folks are down to 2 weeks, which is more dangerous than anything else. It doesn't seem like much has changed.

Nobody is learning a lesson.

Right. Maybe they can't afford to learn the lesson.

That's the reality. Look, if you look at the margin profile and just think about balance sheet. A lot of the companies cannot afford to hold because it's very thin margins. So from that side, I get it. From the other side is they pay for it on the other side when the market does recover, and we talk about shortages and we talk about at this point, I'll take those days any minute.

The last cyclical question for you. One of the responses to the shortages during the lockdowns and all those sorts of things in the pandemic was to add a ton of supply across the entire industry, the capital intensity across analog discretes, broad-based semis, everywhere so that we wouldn't have this problem again on the shortages. Now of course, the argument would be maybe we have too much capacity. Do you think there's any sort of structural excess across the semiconductor space? And does that lead to some of the negative outcomes on pricing pressure and those sorts of things?

Yes. So it depends. I would say it depends where other companies have added capacity. I'm not worried about it in products, for example, for us like Treo. It's not about the capacity. It's the products that you put into that capacity. So from that perspective, I'm very happy with where we are and the capacity, and we're competing very, very well. And I talked about that 60% to 70% margin that we're capturing with our Treo Platform. Now where capacity starts to put pressure on growth or start to put pressure on margin is in a lot of the, what we call the dual sourcing products, which we've been exiting over the last kind of 3 to 4 years, exactly for that reason because when you have capacity, it becomes a price elasticity. And typically, in an environment like today, it is a price headwind in order to get the volume in order to fill the fab. Our approach has been walk away from that because it will never change. Whether it's good times or bad times, there's always going to be that dilutive aspect of that margin and focus our capacity on the Treo like products that regardless of capacity, it's the value of the product that allows you to win. And then the delta between them is take that capacity off-line to position the company for margin expansion when the mix shifts to the higher margin. That has been our strategy, and that's what we've been executing to even -- I mean, we walked away from -- we divested 4 fabs when everybody was dying for fabs, not because of the moment, because of where we are today. We're better off for it today because we divested 4 fabs when the peak was happening. And that's how you can think about our management style is that strategic approach.

That's a perfect segue from those cyclical questions I want to start with to more structural questions. A question I get from investors a lot was on your last earnings call, you talked about potentially exiting about 5% of your revenues from now through calendar '26. A lot of folks hoped you'd be done with that by now. Why are you still exiting products at this point? And what makes that choice for those products occur?

Yes. So let me break that down. So the -- what we said is 5% of our revenue in '25 won't repeat in 2026. So if you break it down, there's 3 components of it. And this isn't new. We've been talking about this for a long time of these exits and the planned transition as we move up the value chain for our customers, moving to more differentiated products. So it's everything of what we've been talking about. So if you take the 3 components: one is the businesses, the dual-source businesses that we've said we're going to exit. We've been exiting those over the last several years. We entered the year thinking we would exit $300 million. Halfway through the year, it's been $100 million. So for the year, we think that will be about $200 million rather than $300 million. So that pushes about $100 million into next year for that exit. So that's just one. Nothing new there. It's just timing. This is business that we've said we would only maintain if the margins were at the corporate average. And we believe as the market back to what you're saying as pricing declines, customers will just naturally leave us, and we're not going to chase that business. So good business to lose, not a high-quality revenue, right? We want to focus on high-quality revenue. The next piece of this is the repositioning of our image sensing business. So we're moving to more machine vision away from human vision where we can differentiate, drive higher margins. That's about $50 million to $100 million for next year that won't repeat for next year. And then the third piece is stuff that we're end of lifing over time, and it's been ongoing on stuff that we don't want to repeat because as we move to that higher margin, let's call it, 50%, greater than 50% gross margin, that's going to be dilutive if we maintain it. So there's nothing new here. What we -- the reason we put kind of a box around this on our last call is because we didn't think the Street was modeling this correctly. Even when we've been telegraphing this for over a year, we wanted to make sure the Street understood what was ahead of us, right? And I think now the investor base has kind of gotten that. I think it's a good thing that we laid it out there. But it's all consistent with what we've talked about. It's just those 3 components are going to happen.

Got you. So it seems like the new one to me out of that, and maybe it was my mistake that it wasn't really new, was the image sensor side of things. I had thought that the majority of that had already transferred to being a more automotive and industrial mix. It sounds like now there's a subset of the automotive mix that is no longer desirable. So what's really changing in that market? And what can you do to stop it from creeping into more and more of that business?

Yes. So if you think about also, that is not new. If you think about our last Analyst Day, we talked about refocusing that business on machine vision rather than human vision. And so let me put it in terms we all live every day. You have the autonomous side of it where the cameras feed into a central compute. There's no human making decisions. The computer will make decisions for the car going left, right or stopping. That's the focus for us in automotive. The human vision side of it is your reverse camera. The car is not making a decision. You're looking at the screen of how far you are and you're making that decision. So that's the difference between the 2. Now why is one more important than the other for us? I always talk about the value that we provide and focusing on value, which means the margin side of it, comes from the value of the technology. So we have the perfect technology with the perfect pixel in all light environment. If you're driving in a tunnel and the sun is in your face, we're able to see a red light on the other side. Almost a human cannot see it. That provides a ton of value for cars. Take the same camera with the high-performance camera and put it in reverse. When was the last time you put your car in reverse and there was not dirt on your lens for your reverse that you can't even see what's behind you almost. So the camera is irrelevant in this case. The quality of the image is irrelevant because there's -- the lens is bad or dirty or cheap or whatever it is. So the customer, why would the customer pay a ton of value for the best camera out there when there's -- it doesn't really matter. So that's the difference between machine vision and human vision. Our focus on machine vision has been a strategic focus. So nothing changed back to your point, but it takes time for that to play out. And that's why it's nothing new, but we put a wrap around and saying this is the number so people can model it properly.