STMicroelectronics N.V. (STMPA) Earnings Call Transcript & Summary

Excellent. Well, thank you very much, everybody, for joining us after the break and the -- for this session with STMicro. So we're -- my name is Jakob Bluestone. I run the EMEA tech hardware team here at BNP Paribas Exane. And it's a great pleasure to welcome on stage as well, Jean-Marc Chery, President and CEO of ST as well as Remi El-Ouazzane, President of MRDF. So I think what we'll do is same as other sessions. [Operator Instructions] But I think maybe what I'll do is I'll start off with a few questions, particularly may I just set the scene a little bit.

So maybe, Jean-Marc, if we can start talking a little bit about the market outlook. When do you think the analog market will bottom out? And can you think -- share your latest thoughts around inventory drag reaching an end?

Well, I will speak about the market we address, which is mix analog, microcontroller and power discrete and so on. Well, I would like to come back to what we say in April about our situation and market. We say -- we believe that Q1 will be the bottom of our revenue and providing we will not face in Q2 and in Q3, some cancellation of order with a dynamic of order entry from customer in Q1, we believe Q1 is the bottom, and we will grow Q2 and Q3. So what is the situation now, so means week 22. First of all, okay, I confirm to you that in Q2, we will deliver at least the midpoint of our guidance. And we will overperform versus our forecast on personal electronics linked to the success of our main customer. And we will overperform on industrial microcontroller or general purpose and mainly in China. And we will perform as per the plan on automotive. This is the point number one. So the point number two is about customer order entry in Q2, so what we classify the bookings. The dynamic is pretty interesting. Why? Because quarter-to-date, the book-to-bill ratio is well above parity. So I have to say above 1.1. And the second interesting characteristic is that the billable of this booking is spread within 2 quarters. So the current quarter plus the next 2 quarters. So it's a short-term order entry, so 85%. So it's pretty characteristic when you have a book-to-bill between 1.1, 1.2 and you have a short-term booking billable means you are in a dynamic of up cycle. And to be transparent, with the assembly, without the uncertainty related to the trade war and geopolitics, showing the current dynamic of the bookings, I would have to say that Q3, for sure, will grow sequentially versus Q2. But Q3 could have been the turning point for ST to grow again year-over-year. But unfortunately, because this uncertainty, I cannot say it very strongly now, but I am thinking very strong. So this is a positive situation. But we don't disclose, okay, more visibility because on Q4 because of this uncertainty. But the takeaway is we enter up to now in a dynamic that is pushing ST toward a turning point. Honestly, we do not see this positive news on booking driven by anticipation or over inventories related to tariff because our country of origin of our device are mainly France, Italy and Singapore and a less extent, South Korea or Taiwan. And at this stage, we do not see a strong behavior by customers to overstock. I don't say it doesn't exist, but we don't detect it exists significantly. So means, okay, it is a pretty healthy booking. But last but not the least, you know that important part of the business we are doing is with distributor. The point of sales of distributor in Q2 is growing strongly. So above 20%. Even if you have to correct by 5% because of the mechanic effect of the number of working days, Q2 versus Q1. But it is really a strong growth sequential, a strong growth year-over-year. And year-to-date, the POS of our distributor now is close to be balanced compared to last year. So it's also an interesting sign. However, I have to say that it is unbalanced on a regional basis. So strong growth in China, solid growth of the POS in America and Asia Pacific, still a little bit negative in Europe, but I guess it's consistent with the overall economical situation in Europe. So this is the business dynamic we see today for MP. So -- moving forward, under the disclaimer of uncertainty, we are going in a way to turning again on the growth year-over-year.

Very helpful and very clear. Maybe if we can just -- you mentioned inventories. Maybe if we can just touch on what are your expectations in terms of inventories as well and anything you're hearing in terms of channel inventories from your customers?

Well, inventory, there is inventory OEM at OEM level, inventory at distributor more in our own inventory. Our own inventory, we control will be at our target we disclosed in April. So I do not comment more. At distributor level, again, with the POS we are facing, now we have an inventory situation on general purpose microcontroller across the board overall coming back normal. but unbalanced by region. So in some regions, we are on a situation of -- to be under inventory. So it means, okay, sometimes we miss our capability to answer just in time on the order requested by customers. So we increased a little bit our delinquencies versus the requested date asked by the customer, typically China and APAC. And on some regions like EMEA and America, we are still on over-inventory situation. So on industrial market and distribution, okay, we expect to come back a normal situation of inventory Q4, Q1 H1 next year. Of course, okay, I will repeat each time, we have to pay attention about the extrinsic variable we do not control, the tariff and so on and so forth. So this is an industrial situation. Automotive, we came back normal. We suffer a strong inventory adjustment in Q1 is done. Now, okay, we are coming on a normal situation. And personal electronics, computer, computer peripherals, server, communication, we have no situation of excess inventory.

Very clear. I guess that sort of covers a large part of the volume side. But can we maybe also touch a little bit about pricing? So what are you seeing in terms of pricing? And what are your expectations there?

Pricing, I think we are consistent since the beginning. very early, we say we saw a mid-single-digit pricing effect. this is what moving forward, we'll have this year. And by the way, I saw our competitor now saying the same.

Yes. And if we maybe go a little bit into some of the individual products and maybe start with MCUs, which I guess is right one for Remi. I mean you recently announced a new generation of stellar MCUs. Can you maybe give a little bit more color around this? Where can we see emerging in the sort of space of auto MCUs and what's the timing around that?

Yes, sure. It's a market we care about. It's over time, $12 billion, $15 billion market as a whole. We are today a challenger in this market. There is there is a heavyweight called Infineon, and we -- as I see, we are a challenger. But we have a quite good faith in our ability to challenge due to 2 things. First is the transition that will happen in the car architecture to what's called a zonal architecture. To make it simple, it's really now the car will be organized by geographic region in the car as opposed to domain before. And if you ask why are they doing that, the biggest -- the simplest way to understand it is reduction of cables. You will have shorter cables and you will have less cables, which will translate themselves into less weight. Actually, cables is a big weight factor in a car. And obviously, as you can imagine, with BEV, less weight is more range. So it's a big transition that's happening. And for us, we have 5 elements as to why we think we can compete. First is geographic resilience. We -- our auto MCUs can will be either China for China or non-China non-Taiwan. It can be produced both ways, and that's pretty unique. Second, we have actually -- we are unapologetically Arm-based end-to-end. So it's an open architecture. You need to know that up and until now, the bulk of the automotive MCUs are based on custom processors that are making customer captive. We've decided to go the exact opposite by being ARM end-to-end. Third, those products carries a lot of memory, and we've launched a new class of memory, which we call X memory, which essentially allow -- there is some magic behind that, but allows you during the lifetime of your car to get access to more memory space should you go and bring new feature to the car during the lifetime of that car. It's pretty unique, and I think we are the only one who can do that. Fourth, and we've done that in the industrial space. We intend to do that in the automotive space. We will make AI a first-class citizen in MCU in the car. And the last piece, which is maybe the cherry on the cake is we want to drive the entire ecosystem towards what we think is the most widely diffused software ecosystem on the planet, which comes from the STM32 franchise. We have roughly 1.5 million unique active users every year. It's not bad. Just for your reference, CUDA is 3 to 4. So it's a pretty large space, and we want actually all our automotive products to be able to leverage our ecosystem. So it's those 5 things that are the recipe we're going to deploy as a challenger to challenge the incumbent leveraging this transition that the car architecture will go through in the next 5 years.

You mentioned being a challenge in Infineon. And I guess if I can maybe broaden the question a little bit also to how do you see the competitive environment currently, I guess, for MCUs?

Yes. Well, it's a competitive space, obviously. It is -- when you accumulate auto, general purpose and secure, it's a $25 billion to $30 billion market. So it's a nonnegligible market. Look, we have a firm belief here as to what it takes to win. We think that you need to be in control of your technology. ST Microelectronics today is actually the most -- is going to ship at the end of the year, the most advanced microcontroller on the planet, and it's based on our own internal technology. You need to master your own nonvolatile memory, the memory that is embedded in those products, which we do. You need to have a large software ecosystem. And I've just described that earlier and a few more things that come along. And I believe that because of that, only a few -- if you project yourself over the next decade, I think there will only be a few amount of players that can pull off the toolkit required to make a difference on this market. And I truly believe that ST is one of those companies.

Maybe I can add that you have a major change in the world of microcontroller. Now the design platform cost is at least 10x more expensive than the one 15 years ago or 10 years ago. And you know that the microcontroller is a market which is really fragmented. So if you want to have the payback of a platform development, you need to be a generalist in terms of microcontroller. So tackling okay, general purpose, but automotive and secure microcontroller. This is ST. And by the way, this is the reason why we changed our organization more than 1 year ago in order to create one unique entity general purpose and automotive microcontroller in order to maximize the synergies, in terms of design cost, to tackle in the most efficient way the market we address.

And I think for MCU specifically, you've talked about winning back market share. Is that what you're still continuing to see? And I guess, are there any other areas as well where you'd highlight particular market share gains or losses?

I'll let Jean-Marc give a more global answer. I will tell you on general purpose MCUs, we are -- like as Jean-Marc was describing, we are cautiously optimistic. This quarter will mark the first quarter since we are actually growing back year-on-year. The last time it happened was 7 quarters ago. And we have also a very strong pipeline of products. Just to give you a little bit of intensity, we are producing a brand-new generation of products or tapping it out, like we say, to manufacture the first samples. We are doing that every 4 weeks right now. So we have a very, very strong pipeline of products covering the bottom of the range to the highest possible range. I feel pretty strongly about our ability actually to get back to our nominal market share minimally. And that nominal market share, we nominally situate it around 23%. And I feel quite good that we will be able to get there. And we'll do our best to try and exceed that.

Can we maybe talk a little bit about the competition from China as well? I guess what are you seeing there in terms of competition from Chinese players? And if you can also explain how do you defend yourself?

Specifically for microcontroller?

I'd say probably more broadly for both, more generally.

For microcontroller, there is a well-known company called Giga Device, which is very active. More than we know that there is at least 70 design hours making microcontroller in China and trying to get advantage of the mainstream technology capacity that are put in a brand-new fab in China. But clearly, at this current period of time, the most active competitor we face in China on microcontroller are TI and Renesas and not the local player because, again, to support this product, you need to have an ecosystem. And moving forward, okay, you have more and more feature in this device, especially when you will provide AI at the edge capability. The cost of a design platform now is starting to be a triple-digit one. So you need to be a big player. Model on analog and it's pretty similar, I have to say, general purpose analog or application specific. The main competitors are still American one on DI/TI. And that's the reason why you saw recently, you have seen some tariff implication impacting the American in China where Chinese have pulling business because of the tariff, but not on us because we are a country of origin free of America. Power and discrete. And clearly, here, there is a strong competition on silicon carbide. On material first from OEM, not semiconductor company, but OEM like BYD, they have developed their own capability on MOSFET, and they are the most aggressive competitor, in fact. On raw material, it is clear that now China is equal or better the Western competitor on silicon carbide. More than the other MOSFET on silicon, IGBT, vertical integrated power, competitors are Infineon and so on. And China players are there, but not the main one. MEMS, on the sensor, high-end sensor, we have no competition from China. It's more on the low end. And for sure, a custom design imaging solution, we have no competition from Chinese.

Interesting. Can we talk a little bit about silicon photonics, and I'm not sure who would rather take this one. But can you maybe help us understand the nature of the sort of silicon photonics opportunity that you've recently communicated with the customers for this? And how big is the revenue opportunity from your point of view?

We will be delighted to. We -- first of all, to go and put things in perspective, silicon photonics is really about backhauling data, no more of copper wire, but over fiber optics. And you do that for 2 main reasons. Why -- one is because the speed at which you need to transmit that data just keep increasing because the processor consuming it just keeps getting stronger. And the second reason you're doing that is because actually copper comes with 2 disadvantage when it's implemented. One is power, but it's also actually gauge the gauge, the size of the copper cable. And you are getting at a point now in AI infrastructure where you are being limited by the numbers of cable that you can put within a rack. I believe it or not, we are building 1 nano processors or whatever TSMC is doing. And right now, we're being limited by how many cables that you can put behind the rack. That's the reality of the situation. And by the way, similarly, power is the next big issue in AI infrastructure, and maybe we can talk about that. As a matter of fact, we have decided to reactivate what we actually pioneered 10-plus years ago, which was silicon photonics, except that when we did that at the time, the market was not quite mature. We are actually right now acting mainly as a foundry. To calibrate you, we are -- when you're looking at the pluggable optics or fiber optics, it is comprised of a microcontroller, where we are rapidly shipping STM32. It's comprised of an electronic IC that is doing a lot of controlling function and is comprised of the silicon photonics chip, which essentially translate electronic into photon or reversely depending on where you stand at the end of the cable. This market as a whole as a foundry market, us acting as a foundry for this market. We think this is a $2 billion market by 2030. And we have clearly established the fact that we intend to reach 25%. Now you may know -- you may have heard that there is another market above and beyond pluggable that is getting created, which is called co-packaged optics, where essentially now the optical link is no more in a cable. It's actually sitting on the GPU or XPU processor in an AI data center. And that's a market for which we are also very actively engaging with from a technology development standpoint, but for which we have nothing to announce. What we can tell you on the first market, the pluggable market, we have announced a collaboration with AWS, Amazon Web Services that has actually committed themselves to leverage and make this a platform for all their suppliers moving forward. And the first production is slated for us to happen by the end of the year, beginning of next. We're not fully in control of when this is going to happen. It's actually done through partners. But we are right now on schedule to go and deploy that technology to the market.

You mentioned AI, and I guess from some of your peers, we've heard particularly around AI power. I guess you're more of an entrant within that segment. I'd be interested in sort of understanding what is your position within AI power? And do you see room to grow within that as well?

We have contrary what I've heard, ST, we have all the technology capability to address from what is from the grid to the processor, each power stage. we have, we have IGBT. We have silicon carbide. We have the driver, okay, we have controllers. Then it's a matter of product design bandwidth and to accelerate our product offering. But it is clear that up to 2 years ago, our main focus, but just consistent with our strategy was to say we want to be broad range leader on automotive and industrial. We focus our resources on these verticals and not on computer and server. Now we are more balancing and diversifying our design resources, also thanks to the organization change we have done early '24, again, to maximize our design resources to have better synergies. And now we consider ourselves as a challenger attacking this market. More recently, we have been retained by NVIDIA. We didn't claim it because we feel that, okay, we will develop our offer step after step with resource allocated. We have put in place, some task force to accelerate our time to market. And I'm strongly believing that '27, '28 ST will be a very strong and serious challenger to the best-in-class, which are, yes, Infineon and MPS.

I want to open up a little bit to the audience as well. But before I open to the room, I got one which is maybe a slightly more hypothetical question, but what sensors and controllers could you contribute to an augmented reality glasses device? And if you got designed in by a major producer, how should we think about the value per unit? I realize it's a bit hypothetical, but is there anything you can share on that one?

Yes, it's difficult for me to answer because for this high-tech device, the OEM, they are super attentive that we do not disclose any specific innovation in advance. And I cannot comment honestly on this. But it is clear that ST with our MEMS and with our optical sensing solution, time-of-flight, this kind of augmented reality, virtual reality, but I would like to expand to humanoid robotics. We have a huge market opportunity to be a growth driver for ST in the near midterm.

Understood. Are there any questions from the audience? Happy to take some. If not, then maybe I can ask you about cost cutting. You've guided for high triple-digit cost-cutting ambitions I guess if you could help us understand what are some of the drivers behind that cost-cutting ambition? And how do you get down your OpEx to sales per your midterm guidance?

No, we already disclosed a number that exiting 2027, basically, we target $300 million, $360 million cost savings on a yearly basis. But we have already engaged because this is pretty under our control, and we leverage the automation process we have put in place and the fact that we have to, let's say, modify the profile of our resources in the field of support function, overhead and including some design, R&D activity. Then on cost of goods sold, clearly manufacturing, we are engaging a reshaping plan that will call to close wafer fab production line, 8-inch and 6-inch. It is pretty clear. Overall, between expenses and manufacturing, we will have to move out on a voluntary basis, I have to say, 2,800 people on top of the attrition. Attrition, okay, we know that we -- 2,000 people will leave. So at the end, it's 5,000 people will leave the company within the next 3 years. We, of course, in each country, we have to discuss with the related stakeholders and authorities based on the local labor laws or rules, the implementation of the plan. We are on time with this discussion. So this is good. In terms of timing, we are not late. Well, I don't want to hide that is one country specifically is harder and most likely, could delay a little bit our speed of implementation if they continue to behave like that. But okay, this is my duty to fix it.

But I guess the other big driver for margin is the utilization as well. And as you've described, we're sort of still coming out of a low demand environment, but it seems like that's hopefully recovering. Just as we sort of think a little bit forward, can you help us think through as the utilization recovers, where is the utilization today? And as it recovers, how much could your margins recover as well?

Today, we are running at 65% of utilization. and we translate it in 420 basis points of unused capacity charges. To come back to utilization rate that is close to 90%, which is the one we consider where unused capacity is marginal, means it's mix dependent. The company will have to come back a quarterly run rate of revenue between $3.6 billion, $3.8 billion. So we will have a double benefit. Of course, we will wash this 420 basis points of unsaturation charges. But on top of that, when you run a fab at 65%, you are not, let's say, washing the inefficiency all in unused capacity. There is a part of efficiency that is remaining, and it is basically 170 basis points. When we will be close 90% loading, this inefficiency will be washed. So you see we have a potential of 550, 600 basis points from the manufacturing efficiency and on saturation charges in overhead on top of the Engage program, reshaping plan and so on.

And I guess, I mean, when we sort of think about where you can potentially get back to the guide you've given was the CMD guidance back in November, since then, has anything changed? We've had tariffs that have come in, the outlook on inventories is maybe taking a little longer. Is there anything that you sort of feel has changed from the CMD in terms of where you might get to? Or is that still very much your vision of where you can end up?

CMD objective, okay, in the time frame we provide and so on. At this stage, we don't change, okay? Please let me come in July to say this is a turning point and understand, okay, where the cycle now is going. And then, of course, because today, as I said, Q4 is difficult to forecast. We are confident and positive on the dynamic of Q2 by fact. the KPI I have in front of my eyes is making me confident on Q3. I cannot say the turning point of the cycle because of this uncertainty. But if it is removed, it will be the turning point. And then, okay, we will go back on a path of growth where basically we will have the cycle impact plus, what is related to our programs. And ST, again, we have sensor, we have analog related to high-growing application like personal electronics, okay, robotics, this kind of stuff. Electrification of the mobility maybe will not grow at the expected level 3 years ago, but will grow significantly. We are our microcontroller, we come back, okay, to our normal market share. And on top of that, we are introducing, as I said, many new products with many new features. We will come back on automotive microcontroller as a very competitive Tier 2 at a transition phase, and we will come back on AI as well. So all these ingredients when I will be capable to assess the speed of the market up cycle will make us confident to confirm our objective.

And what is the key thing that, hopefully, in 3 months, you're going to say this is still flashing green. This is the thing that will give the confidence. Is it the inventories of the customers? Is it visibility over tariffs? What is the key missing piece at this point?

No, the first order is booking and the customer behavior with this booking definitively and our capability to support this strong dynamic as whole economical actors, okay, we should expect at a certain moment less oscillation I've not spoken about macroeconomic issue, but less oscillation in the decision in order to trigger less emotion about decision. So some stability on the tariffs, okay, and so on because it could delay decision on investment, this kind of stuff. So some stability agreement on the trade between the various countries. So this is the second point, more the third point that some people let the management work and focus.

Very clear. I think we've got time for one final question. Maybe if I can just ask you about some of the new programs kicking in, in the second half of the year. I think you said that there's several hundred million relating to new programs. And I guess it would be helpful just to understand what are some of these new programs coming in? What can you share about the types of activity?

I confirm strongly that this program will bring USD 20 million additional growth H2 versus H1. I can confirm that some will perform better and some will perform less. But I don't want to disclose the name of who is performing less. I guess you understand why, but you can feel.

Excellent. Jean-Marc, Remi, I would like to thank you both very much, and...