ON Semiconductor Corp (ON) — Q2 2025 Earnings Call Transcript

Thank you, Parag. Good morning, and thank you all for joining us. In the second quarter, we made meaningful progress across our strategic priorities and advanced critical initiatives in Automotive, Industrial and AI Data Center. In Automotive, we're helping customers like Xiaomi improve range and enhance the driving experience and we're expanding our engagement with more global OEMs and Tier 1s, including our collaboration with Schaeffler. In AI Data Centers, we are enabling next-generation power architectures that drive efficiency and performance at scale through collaboration with market leaders like NVIDIA to accelerate the shift to 800-volt DC power architecture. We remain focused on making strategic investments to extend our competitive edge and deepen customer relationships to build long-term value. At the same time, we are making structural improvements across the business to enhance efficiency. This, combined with disciplined cost management, creates significant leverage in our model as we prepare to capitalize on a market recovery. On the financial side, we delivered Q2 revenue of $1.47 billion, exceeding the midpoint of our guidance and non-GAAP gross margin and EPS of 37.6% and $0.53, respectively. Turning to the demand environment. We are seeing signs of stabilization across our end markets. We have not seen any pull-ins to date due to tariffs and our diversified manufacturing footprint remains a competitive advantage, providing sourcing options to our customers as they work on optimizing their supply chains. By market, Automotive revenue in Q2 was down 4%, performing better than anticipated and is expected to grow in the third quarter with continued EV ramps. In China, select Xiaomi YU7 electric SUV models integrate our 1,200-volt EliteSiC M3e, enabling better performance and the longest range in this class. China remains a growth driver for onsemi with strong traction in both BEV and PHEV platforms. China revenue in Q2 grew 23% sequentially, driven by silicon carbide with the new EV ramps, I mentioned last quarter. We also expanded our collaboration with Schaeffler to deliver our next-generation traction inverter for a global OEM's PHEV platform using our latest EliteSiC M4T trench technology to unlock higher energy efficiency and reliability. We expect adoption to continue to expand and our customer engagement to continue to diversify as OEMs redesign hybrid architectures to meet the emissions target and extend range. Industrial revenue increased 2% quarter-over-quarter. Revenue for AI data center, which we report as part of our other bucket, nearly doubled again in Q2 over the same quarter last year. As outlined in the President's AI action plan, AI infrastructure has become a focus of national priority in the United States. AI growth will be limited by power delivery rather than compute alone and onsemi is the only broad-based U.S. power semiconductor supplier addressing this challenge with our intelligent power semiconductors, dramatically increasing power density and reducing energy loss. We are actively working with leading XPU providers on smart power stages that address the power requirements of current and next-generation platforms. We're in production on single SPS products in an industry standard 5x5 package and began sampling a dual SPS in the same footprint. As part of our ongoing transformation, we will continue investing in next-generation technologies where we have clear competitive advantages while reducing our exposure to areas with limited differentiation. This includes end-of-life of legacy products, exiting noncore businesses and repositioning our image sensing portfolio toward higher value segments such as ADAS and machine vision. These actions are reshaping onsemi into a company with a distinct value proposition, powered by leadership in intelligent power, sensing and analog mixed-signal technologies. A strong example of the strategy in action is Treo. Momentum continues to build around our Treo platform with a design funnel that has more than doubled quarter-over-quarter as we progress towards our $1 billion revenue target. We are on track to doubling the number of products sampling from last year. Treo's differentiated technology, modular SoC-like design and ability to integrate high and low voltage domains are driving strong customer engagement across all our end markets. An example in Automotive is 10BASE-T1S, where we sampled over 10 customers as they work on their zonal architecture. After delivering our first Treo revenue in Q1, we've also reached an important milestone. We've now shipped over 5 million units from our East Fishkill facility this year. These milestones reflect the strength of our innovation engine and the strategic investments we've made to support long-term growth. By reducing complexity, sharpening our operational focus and allocating capital more efficiently, we are building a more resilient and higher-quality business for the long term. As the global economy accelerates towards electrification and intelligent automation, next-generation vehicles, sustainable energy systems and AI data centers are converging around a shared need for a new era of power solutions. Beyond silicon carbide, our strategic investments in next-generation wide band gap semiconductors have delivered transformative gains in power density, thermal performance and energy efficiency. We started sampling customers on these new breakthrough technologies, and I will talk more about it soon. Let me now turn it over to Thad to give you more detail on our results and guidance for the third quarter.

Thank you, Hassane. As we continue our transformation, we are focused on creating sustainable long-term value for our shareholders. We have streamlined our portfolio and manufacturing operations to enhance gross and operating margins at scale, and we will maintain these efforts in the upcoming quarters while pursuing value through our Fab Right initiative. Our investments in next-generation technologies will strengthen our position as a leader in power and sensing and shift our portfolio mix to advance onsemi's value proposition for our customers. In Q1, we took significant steps to reduce our manufacturing capacity and restructure our workforce, which will help drive long-term operational efficiencies. By Q2, we began reaping the benefits of these structural changes, resulting in a noticeable decrease in operating expenses. Additionally, we raised our target for share repurchases in 2025 to 100% of free cash flow. We are on track with this target and, after buying back another $300 million in shares in Q2, we have returned 107% of our free cash flow to shareholders year-to-date. Now, regarding our second-quarter financial results, we surpassed the midpoint of our expectations, achieving revenue of $1.47 billion, a 1.6% increase from Q1. Automotive revenue reached $733 million, a 4% sequential decline, primarily due to softness in America and Europe, though this was partially offset by strong performance in China. Our Industrial revenue was $406 million, up 2% sequentially. While our medical and aerospace and defense sectors continue their growth, traditional industrial saw a slight decline in Q2 compared to Q1. Aside from Auto and Industrial, our Other sectors enjoyed a 16% quarter-over-quarter increase, with AI data center being a major contributor. Examining the revenue by business units, our Power Solutions Group reported $698 million, an 8% rise quarter-over-quarter and a 16% decline year-over-year. The Analog and Mixed-Signal Group generated $556 million, down 2% from the previous quarter and 14% from last year. Meanwhile, the Intelligent Sensing Group brought in $215 million, reflecting an 8% decrease quarter-over-quarter and a 15% decline year-over-year. Regarding gross margin in the second quarter, both GAAP and non-GAAP gross margin stood at 37.6%, exceeding our non-GAAP guidance midpoint. Manufacturing utilization remained steady compared to Q1, and following the capacity impairment completed in Q1, utilization is currently at 68% based on our reduced manufacturing capacity. We anticipate a $5 million decrease in depreciation expenses starting in Q4. In terms of operating expenses, GAAP operating expenses for Q2 were $359 million, down from $396 million in the same period last year and decreased sequentially from Q1, which included restructuring charges. Non-GAAP operating expenses were $298 million compared to $308 million a year ago, and these expenses decreased $17 million sequentially, slightly above our guidance midpoint. This was affected by delays in fully realizing the benefits from our restructuring activities, which we anticipate will be fully recognized in Q3. Our GAAP operating margin was 13.2%, with a non-GAAP operating margin of 17.3%. Our GAAP tax rate was 12.6%, while the non-GAAP rate was 16%. Looking ahead, we foresee no significant changes in 2025 due to a single substantial bill, with a positive impact expected in 2026 and beyond, lowering our non-GAAP tax rate to around 15%, down from a previous estimate of 19%. In the second quarter, diluted GAAP earnings per share stood at $0.41, compared to $0.78 a year earlier. Non-GAAP earnings per share were $0.53, a decrease from $0.96 in Q2 of last year. Our diluted share count was 415 million shares. Turning to our balance sheet, cash and short-term investments totaled $2.8 billion, with total liquidity at $4 billion, including $1.1 billion available on our credit line. Cash from operations was $184 million, and free cash flow was $106 million. The drop in free cash flow was due to timing issues within working capital. Year-to-date, our free cash flow is 19% of revenue, and we remain on course to achieve a 25% free cash flow margin for the full year. In Q2, we spent $78 million on capital expenditures, which represents 5% of revenue. Our inventory increased by $9 million quarter-over-quarter and decreased by 11 days to 208 days, comprising 87 days of bridge inventory for fab transitions in silicon carbide, reduced from 100 days in Q1. Excluding these strategic builds, our base inventory stands at a healthy 121 days. Distribution inventory was at 10.8%, up from 10.1 weeks in Q1, remaining within our target range of 9 to 11 weeks. Looking ahead, I’ll share the key components of our non-GAAP guidance for Q3. Today's press release contains a table detailing our GAAP and non-GAAP guidance. Our current projections suggest no material direct impact from the recently announced tariffs. For Q3, we anticipate revenue to range between $1.465 billion and $1.565 billion. Our non-GAAP gross margin is expected to fall between 36.5% and 38.5%, accounting for $6 million in share-based compensation. Our Q3 guidance also includes 900 basis points of noncash under-absorption charges, with utilization expected to remain flat or increase slightly. As for non-GAAP operating expenses, we expect them to range from $280 million to $295 million, including $32 million in share-based compensation. We foresee a net benefit of $8 million from non-GAAP other income, where our interest income will surpass interest expenses. Our estimated non-GAAP tax rate will be around 16%, with our non-GAAP diluted share count projected to be approximately 410 million shares, resulting in non-GAAP earnings per share estimated between $0.54 and $0.64. We expect capital expenditures to range from $35 million to $50 million. As we look to the future, we are actively rationalizing our product portfolio to focus on higher value and higher-margin products. We expect that about 5% of our 2025 revenue will not recur, which includes the phase-out of certain legacy products, ongoing exits from noncore areas, and the realignment of ISG. We continue to implement our Fab Right strategy to align capacity with this transition to improve revenue quality and enhance long-term earnings potential. In conclusion, we are committed to maintaining financial discipline and a strong focus on shareholder value. By taking decisive actions to streamline our portfolio and adjust operations, we are well-positioned for recovery, continuing our investments in next-generation technologies and capabilities to strengthen our competitive edge and support our transformation. With our emphasis on intelligent power and sensing, we are evolving onsemi into a more focused and differentiated company.

Hassane, the first one's for you. And I guess just kind of 2 parts to it. You sound better cyclically than you have in a while. So the first part is what are you seeing cyclically? And where are there still headwinds? Where are you seeing mainly the tailwinds? And perhaps more importantly, when we move to the secular side, you talked about the AI data center side, the Treo side of things, but we still have the offsets of businesses you're exiting, like Thad just mentioned probably a 5% headwind. Can you talk a little bit about the traction in those secular drivers and when the good ones are going to offset the exits?

Yes, Ross, thanks. That's a great introduction to the main topic of the call. We are observing stabilization compared to the last three to five quarters, which is a positive sign. We mentioned that Automotive hit its low point in the second quarter, and we expect it to improve in the third quarter. While we are seeing signs of stabilization, I wouldn’t call it a recovery just yet, as there is still a lot of uncertainty and customers are being cautious. However, I feel cautiously optimistic about the future. We will continue to manage the company carefully until we have a stronger foundation for recovery. We are controlling our operating expenses and being disciplined with investments aimed at transitioning the company to high-value products and revenue. Regarding AI data centers, we have begun introducing products that are gaining traction, not only in analog mixed-signal areas but also with silicon carbide JFETs. This sector has doubled year-on-year compared to the same quarter last year, demonstrating the expected momentum. Our Treo product is also seeing differentiation, and we reached a significant milestone last quarter with revenue recognition that came earlier than anticipated. Initially, we expected revenue in the second half of this year, but we reported the first revenue in Q1 of 2025. The next milestone we’re focusing on is increased volume, which reflects the broad interest from customers. All of our recent investments are contributing to our longer-term vision, emphasizing the need to reshape the company into a high-value product firm, which should enhance our revenue quality, margins, and capital allocation. Everything we have been working on is starting to align with market traction, providing a solid basis for our future direction.

I guess as my follow-up, one for Thad, on the gross margin side, again, a little bit of a near-term long-term balance. In the near term, why is it flat to slightly down if your revenues are up? And then longer term, especially given the changes in mix that you're talking about, what are the key levers you think you can pull to get to the 53% long-term target, if indeed, that is still the target?

Yes, the key to increasing our margins is all about utilization. We've mentioned this for a few quarters. As we see a recovery, utilization will improve, which will positively impact gross margins a couple of quarters later as we work through our inventory. In the short term, we are being cautious. We have inventory on the balance sheet that we need to reduce. We are streamlined in our operations and distribution, positioning us well to increase utilization when the market improves. Therefore, we anticipate our performance will be flat to slightly up in Q3, and as we gain more clarity into Q4 and early next year, we will consider increasing utilization, which should provide a favorable boost as we move toward 2026. Regarding our goal of reaching a 53% margin, we have about 900 basis points of underutilization charges in our Q3 guidance, consistent with Q2. Each point of increased utilization translates to an improvement of 25 to 30 basis points in gross margins, and as utilization increases, we will recover that 900 basis points. Additionally, we are benefitting from the monetization of divested fabs as we bring production back in-house, and our ongoing work on the Fab Right initiative is expected to contribute another 200 basis points. As Hassane mentioned, the new products we are ramping up come with favorable margins. When you consider these factors, we can get closer to our long-term target of 53%. However, in the short term, the focus remains on improving utilization, which is our primary driver.

For the first one, I just wanted to go back to Q2 results. So Industrial was a bit softer than I think you had thought before. So what drove that? And then the Other was much stronger. Is it really the data center part of that Other? And if that is the case, how large is the data center part of your Other business right now?

Yes. On the Industrial, it wasn't up as much as we were expecting. That's primarily because of what we call the traditional industrial. It was down just slightly. When we look at our traditional industrial, I would say it's kind of bouncing across the bottom. We think we've stabilized, but there's going to be some ups and downs here. It was down slightly and down more than we thought it was going to be. So that's the primary driver on the industrial. On the other market, yes, it's the AI data center and the opportunity that we have there. We talked about year-over-year that, that business has doubled. So still a small piece of the overall company, but growing nicely.

Okay. And for my follow-up, Hassane, where are we in the automotive recovery cycle for ON? I think you mentioned China appears to be strong for you. Is it really the weakness outside of China, especially at the North American EV OEM? Because I'm trying to contrast what you're seeing versus what several of your analog peers are seeing. Some of them are within, I think, 4%, 5% of their prior automotive peaks, whereas ON's auto business is still, I think, 30% of your prior peak. So why is the automotive recovery so slow for ON? And when do you think that your Auto business could start to regrow year-on-year?

Yes. In the automotive sector, particularly outside of China, both Europe and North America are experiencing weakness. There is significant uncertainty in the automotive market. Our situation hasn’t changed much except for the portfolio rationalization we've undertaken, which won't be repeated in the future. The ramp-up in electric vehicles is ongoing but not at the anticipated pace. The unit volume is below expectations, and the differences in performance compared to our peers are mainly due to our market mix and exposure. I can't speak to what our competitors are experiencing in the automotive space. However, we reached a low point in the second quarter and are beginning to see growth. We anticipate growth in Q3, and we'll monitor developments over the next few quarters to get a clearer idea of where this will lead.

I actually wanted to ask you about the ISG repositioning. And I thought you said the 5% was the end-of-life products, and that's what you've been saying for a while. So maybe you could just walk me through, I guess, what are you repositioning? Why? And is there a revenue number tied to that repositioning in ISG?

Yes. Our strategic focus has been on the machine vision segment. As competition increases, we're prioritizing value and high-quality revenue driven by our technology and unique offerings. This differentiation aligns more with machine vision products rather than human vision products. For instance, in reverse parking scenarios, the quality of the camera is less relevant due to dirt on the lens, which means we won't engage in those designs. However, for advanced driver-assistance systems (ADAS), where clarity and image quality are crucial for safety, our vision products provide significant value. This marks a shift from our previous strategy of targeting market share across all sectors to a more concentrated and value-driven approach that we have embraced for several years. We're confident in our strategy moving forward and in the investments we're making to sustain our competitive edge in preferred markets.

And Blayne, to answer the second part of your question in terms of the revenue impact, we think for 2026, it's about $50 million to $100 million that doesn't repeat from the 2025 baseline.

Is that incremental to the 5% from the end-of-life stuff?

That's inclusive of 5%.

Got you. And then maybe, Thad, just on the distribution inventories, I guess, it kind of came in the high end of the range in June, kind of what's your expectation for distribution in the September guide?

Yes. I think it's going to be right in this range, let's call it 10 weeks plus or minus, right? We came in at 10.8%. It's in our sweet spot of 9% to 11%. So we don't expect any material change one way or another.

We don't focus on quarter-on-quarter details as long as the numbers stay within the expected range. There are ramps that we experience in the third quarter; we deal with those in the second quarter, and then we reach a steady state. As long as the changes remain within our range, we don't feel the need to control quarter-on-quarter variations, provided we maintain our customer ramp strategy.

Just a couple of questions digging on the gross margins. Is the 5% of business that's going away next year? Is that going to be gross margin accretive? And if so, how much? And then how does the silicon carbide business fit into the overall gross margin ledger? Are those gross margins lower than the corporate average now? And then how do we get those back to the 53% target?

Yes, Chris. Currently, the gross margin for silicon carbide is below the corporate average mainly due to underutilization. To bring those margins back to the corporate average, we need to increase volume and fully utilize our manufacturing capabilities, which will happen as this business grows. Concerning the exits, in the long run, they will negatively impact margins. Right now, the margins are around the corporate average, but our goal of achieving over 50% gross margin cannot be supported by this business. That’s why we are choosing to exit it now. We may have misjudged the timeline for this exit in the past, expecting it to happen sooner. While it will have a dilutive effect in the long term, in the short term, it will remain neutral since it is currently at the corporate average. We will also adjust our manufacturing capacity accordingly as we move through this transition. As I mentioned in my prepared remarks, we are aligning the right capacity with these planned exits.

Okay. And for my follow-up, just real quick. What percentage of your auto business is China? And then how would you expect that to trend over the next couple of years?

We do not break out the Auto business specifically for China; we report it as a whole. However, we see China as a key market for us. We currently hold a 50% market share, which we anticipate will continue to increase alongside our revenue and unit sales. We are pleased with our standing in China. Our strategy there focuses on enhancing efficiency and range. Each success in China is linked to the quality and performance of our products, which helps us stand out not only against Western competitors but also against local companies. We intend to maintain this differentiation. I mentioned earlier that we have already introduced our Trench silicon carbide technology, which has led to some successes. This is how we plan to stay ahead of the competition, both in China and beyond.

I was curious about the Q3 guidance regarding Automotive, which you mentioned would increase this quarter. However, it seems you're less certain about the Industrial and Other segments being up. Could you provide some insight into what you anticipate for Q3 in those areas? I would guess that the data center component of Other will perform significantly better. Please clarify whether you expect Industrial to be up, flat, or down.

Yes. So the comment I made on Automotive is exactly that, but I made it specifically on Automotive relative to the second quarter being the kind of the trough as we ramp. But just to clarify my comments, we expect every end market, Auto, Industrial and Other to be up in the third quarter. Other will be up higher driven by the ramps that we see in these markets that we put under Other, which includes AI, of course.

Yes. And to give you a little more specific expectations there, we expect Auto and Industrial to be up low single-digit percentages. We think Other is going to exceed that will be up in the mid- to high single-digit range.

That's very helpful. And then maybe if you could give us any kind of color on within Auto, you talked about the U.S. and Europe being a little bit weaker. I don't think that's particularly surprising to anybody. But can you maybe talk about the reasons for that? Is it purely tariff-based uncertainty in terms of their end market uncertainty? Or do you think there's a little bit of excess inventory or buffer stock still trying to work down internally?

I would say it's a combination of factors. I can't pinpoint one specific reason. It's not just about the industry or the market. Every customer faces their own challenges. Some have inventory issues, but we don't view that as a widespread problem. There's the impact of tariffs and general uncertainty regarding end market demand. As a result, customers are hesitating to place orders until the last minute. That's the cautious approach we've adopted, and we've generally been correct in our assessment. Therefore, we will continue to manage based on the visibility we have. When it comes to the auto sector, particularly in Europe and North America, it reflects all the various issues we've been seeing in the news.

I wanted to revisit the topic of utilization impact. You mentioned that each point of utilization equates to 25 to 30 basis points, and that there are around 100 basis points of underutilization charges factored into the guidance. This suggests that to achieve the full 900 basis points, we need to reach 98% utilization. However, I recall you previously indicated that full utilization was typically in the low to mid-80s range. Could you clarify your view on what constitutes full utilization?

Yes, that's a good question. Previously, we indicated that full utilization for us was in the mid-80% range. After the impairment we recorded in Q1, it's now in the low 90s, around 92%. When you calculate the math on the 25 to 30 basis points for each point of utilization, it adds up to about 700 points of improvement. Additionally, there are another 200 basis points from our Fab Right initiatives that are still in progress. Together, these will bring you to approximately 900 points. So, our current expectation is that with the reduced footprint, full utilization is now in the low 90% range, showing an improvement.

Got it. And then the rest sort of from the 46 or so percent to 53%, that would all be mix and new products?

Yes. Yes, exactly. Exactly. Sorry, one more thing. As well as the fab divestitures, right? So you got another 200 basis points of the fab divestitures that we'll recognize.

Got it. Okay. Hassane, you mentioned the smart power stages, both single and dual phase that you are currently sampling. Can you provide insight on whether the qualification process will take about a year? Is there a potential to accelerate that timeline? Additionally, could you share your expectations regarding the 800-volt rack opportunity? Do you anticipate that it will ramp up around 2027, considering what NVIDIA has communicated about its own 800-volt rack timeline?

Look, given the sensitivity with mentioning customers, whatever we have specifically on that opportunity is listed in the press release.

Okay. How about just the power stages then?

The power stages follow a standard design cycle, so we anticipate a qualification timeline of around 12 to 18 months for production. This is dependent on market adoption and any potential changes in the roadmap. The key point is that we are aligned with the XPU suppliers on the specific roadmap. As they begin to ramp up and deploy their new platforms, we will be ramping up alongside them. However, we are already in production ourselves. The single SPS is ahead as we begin sampling, and our qualification occurs with the customer. These processes occur at different stages. It’s important to note that we have a roadmap in place and are maintaining a steady flow of new products aimed at the AI data center market, spanning from high-power JFETs to SPS products positioned close to the GPU or XPU. We've mentioned in recent years the need to enhance our new product engine, and that enhancement is underway, serving as proof points of our progress.

I wanted to ask about inventory levels. I believe last quarter, you mentioned that you were expecting them to peak in 2Q and then decline through the rest of the year. Is that still the right way to think about inventories for the year? And any amount that you expect to take out through the balance of 2025? Like how should we think about utilization rates? I guess I'm a bit surprised to see them up when you're trying to bleed inventory given revenue was up modestly in the guidance for the third quarter.

Yes. Now keep in mind, the utilization is flat quarter-on-quarter. The calculation now is 68%. But if you normalize to pre-impairment, it's 60%, right? So it's flat quarter-on-quarter. The new calculation gets you to 68% utilized given that we have a smaller footprint and less capacity. First part of your question...

Inventory.

Yes, we are on track with our inventory management. We expect to reach a peak in inventory during the second quarter, followed by a slight decrease in the third quarter. I anticipate that this downward trend will continue into the fourth quarter as we reduce our strategic inventory. This has always been our approach, using inventory to support the fab transitions in silicon carbide, but we will be working through that now. This will contribute positively to our cash flow.

Okay. I appreciate the color there. And I'm sorry to keep picking on gross margins. But I wanted to ask about pricing. I think last quarter, you mentioned pricing a bit more aggressively to defend market share. How did that develop in the quarter? And any changes in the pricing environment that you've seen over the last 90 days?

Yes. No, no change to the pricing environment. It's actually stable, within our expectations. So there's nothing new here.

I believe you said roughly a $300 million revenue headwind in fiscal year '26 or roughly 5% of revenue as you exit noncore business. How much of a headwind is it for fiscal year '25? And should we think about it as spread over 8 quarters, roughly $30 million to $40 million per quarter headwind? Or is it linear like that? Or is there any sort of step function?

Yes. For 2025, we're expecting around $200 million in exits. Year-to-date, through Q2, we're on track for $100 million. As I mentioned earlier, we believe we'll achieve these exits faster, which means that $100 million will extend into next year. This is the impact for 2025.

Yes. So from East Fishkill, obviously, it's the utilization impact overall that Thad mentioned from our fab network. East Fishkill is part of that fab network. So it comes with utilization. The important thing is we already have our power products, our silicon power products qualified and shipping from there, high voltage and medium to low voltage. Our image sensor qualification is on track as expected. And the Treo has started, I wouldn't call it a soft ramp at 5 million units already, but that's, again, a brand-new product that we ramped up with a brand-new technology in East Fishkill. So the fab is running and qualified where we want to qualify. And right now, it's primarily driven by utilization. But from a technology perspective, we're pretty happy with the performance.

Just a quick question on the Section 232, 301. Obviously, a lot of noise around that. How are you guys preparing for that? What are you expecting in terms of when this happens? And I have a follow-up.

I mean, I don't know, Vijay if anybody told you I have a crystal ball better than my peers. But I think the way we prepare for it is we remain focused on what we can control. We're talking about our footprint, our fab footprint, our manufacturing footprint as a competitive advantage. That has been recognized by customers, especially as the whole tariff talk has been for a few quarters now. The 232, I believe, will be very similar to the changes that we have to go through with customers. The thing is there's no planning to be done here because we don't know where it's going to land on either side, except the fact that we need to maintain flexibility, and we need to maintain the focus on what we can control.

Got it. And then in terms of silicon carbide, obviously, good to see you guys picking up some share there. One of your peers declared bankruptcy. Just wondering how that's playing out from a business risk perspective, obviously, people might be moving or reallocating. But just wondering how that's kind of playing out on the roadmap.

Yes. Look, from a roadmap perspective, that doesn't have an impact. I've always maintained my focus on we're going to win. We're going to win because of our own products and because of our own investment, not because one of our peers are struggling. So we're going to win because of things we are doing. Having said that, the changes in one of our peers with the bankruptcy, obviously, that is not the trigger that forces customers to think otherwise. We all know they struggled way before the bankruptcy filing. I think a lot of roadmap changes from our customer, a lot of sourcing decisions have already been made. So that to me is all, I would say, part of the baseline, part of the funnel, part of the ramps. I wouldn't call the bankruptcy as a trigger.

I'd like to also dig into silicon carbide for a moment. I think one of the things that came out in the last quarter or so, Hassane, is that perhaps for some of your customers, there's been an interest in purchasing chips instead of modules. Modules had been the story. Now we're hearing more about Trench and some other maybe aspects of the individual chip design. Can you comment on that change? And if customers continue to choose chips over modules, does that influence your long-term thinking about the attractiveness of this market for you?

Yes. The shift in the mix between modules and what we call die has been happening for a few years and is already part of our baseline. Regarding your reference to Trench, that aspect is independent of die versus module. Trench represents another step in device design. Our planar technology outperformed others, and we've recently introduced a new trench design that is succeeding due to its performance. Whether we sell it as die or as a module, we win based on the performance of that die, and we often achieve better outcomes when it's integrated into our own module, as we can design it specifically to fit our die. Many of our projects, even in China, include a combination of die and module sales. We adapt to what our customers need and their core competencies. Some customers can create a module, while others prefer to purchase ours. Our goal is to provide flexibility and deliver the optimal solution to address customer challenges and deliver value. This value is tied to efficiency, which is reflected in our trench technology on the die side. Our perspective on the market remains unchanged. Our results and the market share we've gained indicate our success, regardless of the mix of sales. We will continue this approach, focusing on maintaining R&D and investing in key areas for differentiation in our markets.

If I can follow up in a similar area. One of the big consumers in silicon carbide has discussed a migration to hybrid inverter from silicon carbide to silicon carbide combined with IGBT. I think they said that they intend to reduce SiC by 75% in that traction inverter. But I think they haven't done it yet. I wonder if you're seeing that influence your go-forward view of this market, if it changes anything for you?

No change. This is not just the perspective of one customer. If you look back two years to our Analyst Day, Simon already demonstrated how the transition will favor higher-end, higher-performance silicon carbide. As we move towards more mainstream applications, we will see a hybrid approach, while IGBT still has its place. These points are not new; I've discussed them repeatedly. Our perspective on the market remains unchanged. Our competitiveness lies in our ability to offer a wide range of technologies, including silicon carbide, trench planar, and uniquely designed IGBTs. Some vehicles still utilize IGBT on one axle and silicon carbide on another, or feature full IGBT systems. This does not alter our outlook on the market. We will continue to succeed based on the performance of our products.

Question for maybe Hassane. Hassane, as you think about the recovery and maybe the fuel for recovery, you talked about 900 basis points underutilization. Can I ask if there's an element of written-off inventory here as well that might come into play as you look at recovery?

No, no. When you talk about like reserve inventory or written off inventory, no, because we have a very disciplined approach on what you reserve and what inventory you write off. Inventory write-off does not have demand. So it cannot be part of the recovery. So when we talk about purely the 900 basis point being on utilization, it is purely demand-driven that will drive utilization with healthy inventory that we build and ship to customers. If you recall, we saw the shift in market before a lot of our peers. So we took down our utilization ahead of most of our peers and ahead of the market really softening. So that puts us in a much better position from the inventory we have. We are already draining our inventory. And as demand picks up, we can pick up more utilization on the fab that will help the margin, help cash flow when we shift the bridge inventory. So everything we've done and the discipline we've shown all the way here is what's going to drive that healthy recovery without any kind of side notes.

Yes, Harsh, looking at our base inventory without the strategic items, it stands at 121 days, which is within our optimal range. This is quite healthy, and we don't perceive any risks associated with inventory. The strategic items are simply a matter of time to sell as demand returns. These products usually have a longer lifespan. Therefore, as we evaluate our current situation, we do not foresee any substantial risk of having to write off inventory.

Understood. And then as my follow-up, can I ask. You've got the legacy piece that you mentioned, I think, $50 million to $100 million, you'll peel off this year. You've also got some growth areas, which you called as other that are showing outsized growth. I was curious if you can help us understand what are the major components of the other outside of the AI data center piece and then maybe how big it is?

Yes. Look, the primary focus for the others bucket that I will talk about is really the AI data center. We have some client in computing, they're a very small part of our business. That's why we put it in Other. We're not breaking those out. As far as AI data center, it's still showing a lot of growth. It's showing per expectations because we're investing in that business. As we get more and more into that business and it becomes sizable, we may talk about it in more detail. But for now, we're just keeping it in Other, and that's really the driver for the growth you're seeing there.

Thank you all for joining us. I want to take a moment to thank our global teams for their relentless execution, our customers for their continued partnership and our shareholders for their ongoing support. Together, we are building a more resilient and higher-quality business and I'm confident that the strategic progress we've made this quarter positions onsemi for long-term success. Thank you.