ON - NASDAQ

Ladies and gentlemen, thank you for standing by. Welcome to onsemi Third Quarter 2025 Earnings Conference Call. [Operator Instructions] Please be advised that today's conference is being recorded. I would like now to turn the conference over to Parag Agarwal, Vice President of Investor Relations and Corporate Development. Please go ahead.

Thank you, Michelle. Good morning, and thank you for joining onsemi's Third Quarter of 2025 Results Conference Call. I am joined today by Hassane El-Khoury, our President and CEO; and Thad Trent, our CFO. This call is being webcast on the Investor Relations section of our website at www.onsemi.com. A replay of this webcast, along with our third quarter earnings release, will be available on our website approximately 1 hour following this conference call, and the recorded webcast will be available for approximately 30 days following this conference call. Additional information is posted on the Investor Relations section of our website. Our earnings release and this presentation include certain non-GAAP financial measures. Reconciliation of these non-GAAP financial measures to the most directly comparable GAAP financial measures and a discussion of certain limitations when using non-GAAP financial measures are included in our earnings release which is posted separately on our website in the Investor Relations section. During the course of this conference call, we will make projections or other forward-looking statements regarding future events or the future financial performance of the company. We wish to caution that such statements are subject to risks and uncertainties that could cause actual events or results to differ materially from projections. Important factors that can affect our business, including factors that can cause actual results to differ materially from our forward-looking statements are described in our most recent form 10-K, form 10-Qs and other filings with the Securities and Exchange Commission and in our earnings release for the third quarter. Our estimates or other forward-looking statements might change, and the company assumes no obligation to update forward-looking statements to reflect actual results, changed assumptions or other events that may occur except as required by law. Now let me turn it over to Hassane. Hassane?

Thank you, Parag. Good morning, and thank you all for joining us. We are pleased with our third quarter results, which reflect the strength of our strategy and the resilience of our business model. Our third quarter results exceeded the midpoint of our guidance with revenue of $1.55 billion, non-GAAP gross margin of 38% and earnings per share towards the high-end of our range at $0.63. We have been positioning the company for a market recovery, and we believe we are well aligned to benefit as demand normalizes. We're already seeing stabilization in Automotive and Industrial while continuing to grow in AI. Our Treo platform continues to scale across our core markets, and our recent acquisitions are expanding our portfolio and accelerating our road map. We are delivering solutions that help customers scale performance while improving energy efficiency and system cost. The growing demand for high-efficiency power delivery across our end markets of Automotive, Industrial and AI positions us for long-term growth. We remain committed to our gross margin expansion strategy through innovation with both organic and inorganic investments in differentiation and have achieved four significant milestones that I'd like to highlight. First is our Treo platform. Our new products continue to scale, and our design funnel now exceeds $1 billion, driven by strong customer engagement across Automotive, Industrial and AI infrastructure. We remain on track to double the number of products sampling this year. Teledyne Technologies selected our Treo platform to develop next-generation products for infrared imaging systems. Treo's process technology combines precision analog, advanced digital and low-voltage power features to meet the demands of infrared focal plane array systems used in aerospace, defense and security applications. Second is our vertical GaN or vGaN. Last quarter, I highlighted our strategic investment in our generation wide band gap semiconductors. Last week, we announced our vGaN platform developed on proprietary GaN-on-GaN architecture in our Syracuse fab in New York. vGaN conducts current vertically through the chip, enabling higher operating voltages versus lateral GaN and faster switching and record power density. It reduces energy loss by up to 50%, making it ideal for AI data centers, EVs, renewable energy and aerospace, defense and security. Sampling is already underway with lead customers in automotive and AI. This launch expands our leadership beyond silicon and silicon carbide, giving customers a future-ready toolkit to meet rising performance and efficiency demands. Third, our SiC JFET continues to proliferate, and we have been ramping revenue in AI data center for high current workloads. We're also seeing traction in aerospace, defense and security, where our SiC JFETs are now deployed in low-orbit satellite platforms, delivering industry-leading radiation ruggedness and power density. And fourth is our Vcore acquisition. In Q3, we expanded our analog and mixed-signal portfolio with the acquisition of Vcore Power Technology and IP assets from Aura Semiconductor. This transaction accelerates our road map for advanced multiphase controllers and monolithic smart power stages, enabling us to close key gaps in our offering and deliver comprehensive solutions for the next-generation AI data centers and compute platforms. These new products will be integrated into our Treo platform, enhancing performance, reliability and energy efficiency at the point of load and support x86 and ARM-based architectures. Sampling begins this quarter with production release expected in early 2026. Shifting to the demand environment. We are seeing stabilization in the near term with Automotive, which grew 7%; and Industrial, which grew 5% sequentially. And our design wins in both markets continue to reflect a broad global engagement. For example, our industrial image sensor funnel is up 55% year-over-year with traction in factory automation and inspection. We continue to ramp our AI revenue, which again approximately doubled year-over-year in Q3 and is now becoming material with almost $250 million expected in 2025. Regionally, our revenue in the Americas grew 22% sequentially from momentum in automotive and aerospace, defense and security. Japan was up 38% quarter-over-quarter, driven by traction in automotive and image sensing. Europe was down 4% as macro softness persisted, while China was down 7% sequentially. In China, we secured strategic wins in high-voltage traction inverters with a leading Tier 1 for multiple local OEMs. We also expanded our position at NIO with SiC for their traction inverter across their newest brand and with our 8-megapixel image sensor for their ADAS applications. AI is shaping -- is reshaping the power landscape, both inside and outside the data center. The International Energy Agency projects that electricity demand from AI optimized data centers will quadruple by 2030, making power efficiency and density critical differentiators, an area where onsemi leads. Onsemi's intelligent power technologies span the full power tree from solar and storage systems to UPS and rack-level PSUs, optimizing every watt before it reaches the processor. In Q3, we secured strategic wins in solar and energy storage platforms that are foundational to hyperscale AI deployments. Our latest generation of IGBTs and SiC in the most advanced hybrid modules were selected for high-efficiency solar inverters and energy storage systems or ESS, including wins with two of the leading utility solar inverter suppliers in China. We also secured the next-generation large-scale stationary storage with a large OEM in the U.S. as microgrid deployments are rapidly emerging as a key growth vector across our end markets. This business is reported under our Industrial segment, and we expect our latest generation Field Stop 7 IGBT revenue to increase in 2025 over 2024 with continued double-digit growth expected in 2026. Turning to the AI data center itself. At the UPS level, a leading industrial OEM has integrated onsemi SiC MOSFET into their latest 3-phase UPS platform, where superior efficiency and power density were key differentiators. At the rack level, we secured multiple design wins across high-efficiency PSUs with our SiC FETs, T10, Trench MOSFET and SiC JFET into 5.5-kilowatt AI server PSUs, with top global PSU providers delivering best-in-class thermal performance, supply assurance and switching efficiency for hyperscale deployment. At the compute board level, we have introduced high-efficiency smart power stages and secured design wins on multiple platforms with leading XPU providers. The acquisition of IP from Aura Semiconductor further strengthened our SPS and controller offerings for power to the core applications. Our collaboration with NVIDIA is also accelerating the industry's transition to 800-volt DC power architecture critical for next-generation AI data center. These technology achievements and customer engagements reflect the strength of our differentiated power and sensing portfolios and our ability to deliver system-level value in the high-growth segments of our core markets. Let me now turn it over to Thad to give you more detail on our results and guidance for the fourth quarter.

Thanks, Hassane. Our third quarter results were driven by disciplined execution and prudent management of the business. We have made structural changes across our portfolio and our manufacturing footprint that will enable margin expansion at scale and position us for a market recovery. These initiatives will continue in future quarters, and we are committed to extracting value through our Fab Right activities. Our investments in next-generation technologies, including Treo, Vcore, silicon carbide JFET and vertical GaN are reshaping our mix and strengthening our competitive advantage to further our leadership position. In addition, we continue to return capital to our shareholders. Year-to-date, we have repurchased $925 million of shares, returning approximately 100% of our free cash flow to shareholders. Turning to the third quarter results. We exceeded the midpoint of our guidance with revenue of $1.55 billion, increasing 6% over Q2. Automotive revenue was $787 million, which increased 7% sequentially, driven by increases in Americas, China and Japan. Revenue for Industrial was $426 million, up 5% sequentially, primarily driven by aerospace, defense and security. Outside of Auto and Industrial, our Other business increased 2% quarter-over-quarter with continued momentum in AI data center. Looking at the third quarter results between the business units, we saw sequential revenue growth in all three business units. Revenue for the Power Solutions Group, or PSG, was $738 million, an increase of 6% quarter-over-quarter and a decrease of 11% year-over-year. Revenue for the Analog and Mixed-Signal Group, or AMG, was $583 million, an increase of 5% quarter-over-quarter and a decrease of 11% year-over-year. Revenue for the Intelligent Sensing Group, or ISG, was $230 million, a 7% increase quarter-over-quarter and a decline of 18% over the same quarter last year as we strategically refocused this business. Turning to gross margin in the third quarter. GAAP gross margin was 37.9%, and non-GAAP gross margin was 38%, above the midpoint of our guidance due to favorable mix within the quarter. Manufacturing utilization was up compared to Q2 at 74% as we started to build die bank inventory to support the mass market. We expect utilization to be flat to down slightly in the fourth quarter as we complete these builds. GAAP operating expenses were $323 million, and non-GAAP operating expenses were $291 million. GAAP operating margin for the quarter was 17% and non-GAAP operating margin was 19.2%. Our GAAP tax rate was 6.5% and non-GAAP tax rate was approximately 16%. Diluted GAAP earnings per share was $0.63 and non-GAAP earnings per share was also $0.63. GAAP and non-GAAP diluted share count was 408 million shares, and we repurchased $325 million of shares in the third quarter. Since launching our share repurchase program in February 2023, we have repurchased $2.1 billion and had approximately $861 million remaining on our authorization at the end of the quarter. Turning to the balance sheet. Cash and short-term investments was approximately $2.9 billion with total liquidity of $4 billion, including $1.1 billion undrawn on our revolver. Cash from operations was $419 million and free cash flow was $372 million. Our year-to-date free cash flow is 21% of revenue, and we remain on track to deliver strong free cash flow margin for the full year. Capital expenditures were $46 million or 3% of revenue. Inventory decreased by $39 million to 194 days from 208 days in Q2. This includes 82 days of bridge inventory to support fab transitions and silicon carbide, down from 87 days in Q2. Excluding the strategic builds, our base inventory is healthy at 112 days. Distribution inventory declined to 10.5 weeks from 10.8 weeks in Q2 and within our target range of 9 to 11 weeks. Looking forward, let me provide you the key elements of our non-GAAP guidance for the fourth quarter. As a reminder, today's press release contains a table detailing our GAAP and non-GAAP guidance. Our guidance is inclusive of our current expectation that there is no material direct impact of tariffs announced as of today. We anticipate Q4 revenue will be in the range of $1.48 billion to $1.58 billion. Our non-GAAP gross margin is expected to be between 37% and 39%, which includes share-based compensation of $8 million. Non-GAAP operating expenses are expected to be between $282 million and $297 million, which includes share-based compensation of $32 million. We anticipate our non-GAAP other income to be a net benefit of $7 million with our interest income exceeding interest expense. We expect our non-GAAP tax rate to be approximately 16%, and our non-GAAP diluted share count is expected to be approximately 405 million shares. This results in non-GAAP earnings per share in the range of $0.57 to $0.67. We expect capital expenditures in the range of $20 million to $40 million. To close, we remain focused on disciplined execution and financial leverage. The structural changes we have made across our portfolio, operations and manufacturing footprint are driving margin expansion and positioning onsemi for long-term earnings power. With over 100% of our year-to-date free cash flow returned to shareholders, we continue to prioritize capital efficiency and shareholder value while investing in innovation and differentiation. As the market stabilizes, we are well aligned to scale with demand and deliver sustainable growth. With that, I'd like to turn the call back over to Michelle to open it up for Q&A.

[Operator Instructions] And our first question comes from Ross Seymore with Deutsche Bank.

First one, I want to ask about the automotive side of things. Upside in the quarter nicely versus the low single-digit guide that you had. So Hassane, I just wanted to get an update on what you're seeing in that end market, what caused the upside, and perhaps what's the sustainability of that sort of growth as you look into the fourth quarter and then 2026 as well.

Yes. So nothing really out of the ordinary. It's -- you can think about the Q3 and Q4 as I do, which what I've been talking about, I look at the second half versus first half of the year. The quarter-on-quarter lumpiness as customers try, as new designs ramp, I wouldn't read any much into it. What we're seeing in automotive is really stabilization, which is a positive from where we were. So the quarter-on-quarter, I don't think I read anything into it. It's purely between seasonality and ramps. As far as 2026, look, we'll let you know as we get closer, a lot of things going on out in the world. So we're not guiding specifically by market into 2026. But what I can tell you is demand is stabilizing, we're starting to see a seasonal trend. But one thing I would highlight is we haven't seen a restocking cycle yet. So that's still out there.

I guess as my second question, perhaps a little bit of a longer-term one. You, for the first time, I believe, sized the AI business at about $250 million, I think, for this year as a whole. So what is it, roughly 4% of sales. Can you just talk about how ON differentiates in that? You mentioned the collaboration list on the 800-volt with NVIDIA, and that's great to be on that list, but there's 13 other folks on the list as well. So as you look at that market, how do you believe that $250 million will grow? And what's the differentiation ON delivers to drive that growth?

Yes. So that's a very good question. So overall, we expect the AI data center for us to continue to grow. We look at ourselves as really the share gainer from some of the companies that have been in that market longer than we have. So being able to post $250 million or about $250 million of revenue is pretty stellar. You can see our investments accelerating in that across the whole power tree. So from a customer perspective, to answer your question more directly, if you take that "crowded space" of 13 or however many companies and you look at who can go from wall to core, there's only 2. And we will -- we are the share gainer, we're of the two. So the way we differentiate is we're one of the only companies that are able to support the power delivery from the high voltage all the way to the core, all with the product portfolio that we have that we've grown organically or even inorganically with the Vcore acquisition. So that's how we differentiate. We have proven that differentiation through our JFET silicon carbide, through our AMG with the products that they have been delivering and the revenue growth that has doubled year-on-year every quarter in the first 3 quarters to deliver that number I gave in 2025 is really the proof of that.

And our next question will come from Vivek Arya with Bank of America.

So Hassane, I know it's a little early, and I'm not asking for a quantitative guidance. But I'm curious how you are thinking about seasonality in Q1 and just growth in '26 overall versus how you thought about it 3 months ago.

No change from where we were 3 months ago. So still with the same outlook, same expectation.

Okay. For my follow-up, maybe on utilization and gross margins. Thad, I think you said something about utilization perhaps flat to slightly down. Anything more to read into it? And if you could just remind us what is your seasonal pattern in Q1? And if there is some utilization headwind from Q4, how does that kind of reflect in gross margins in Q1 just based on historical seasonal trends?

Yes. So our utilization increased to 74% in Q3. As I said in the prepared remarks, we've been building die bank inventory for the mass market. This is a market that we've talked about for probably well over a year about we need to seed that market. We've been doing it through the distribution channel. Now we've got to hold inventory in die bank for kind of quick turn on that mass market that we need to invest in. I expect it to be down, the utilization to be down in Q4 because we expect those builds to be completed. So going back to kind of a normalized utilization rate from that point forward. You can think about utilization, the impact on utilization is having a couple of quarter impact on the P&L. There's always a delay on that, right? So if you think about going into the next year, and obviously, we're not providing any guidance at this point, but we think between Q4 and kind of the next couple of quarters, we're looking at seasonal patterns. If you take the midpoint of our guidance for Q4, it's directly in line with normal seasonality, which is typically flat to down 2%. I think the midpoint of our guidance is down about 1.3%. And to answer your question about kind of what's the normal seasonality in Q1, it's typically down 2% to 3%.

So does that mean slightly lower gross margin in Q1 time? Or just how should we be prepared based on if that kind of seasonality is what actually emerges?

Look, we're not guiding that far out, Vivek, but there's tailwinds as the utilization improves over time.

And our next question is going to come from Chris Danely with Citi.

I'm sure you've seen this ongoing Soap Opera at Nexperia. Have you seen any impact to your business either directly or indirectly? Or do you anticipate any impact longer term from all this stuff going on over there?

Look, as you said, there's a big impact. It's too soon to call anything. We're focusing really on the business that I've really outlined. But what I can say about that, obviously, is we have a lot of the same customers, and we are supporting our customers to the extent we can, and we'll continue to do that with the complete portfolio, not just the parts that may be impacted. So what I can say is I'm not redirecting any changes from where we are, but we are supporting customers as they request it.

Okay. And for my follow-up, so it seems like the Auto market is starting to do a little better than the Industrial end market. We've seen this trend at several of your peers. Going forward, would you expect Auto to keep outgrowing Industrial for the next few/several quarters?

I wouldn't put the two kind of -- I guess I wouldn't compare the two and read anything into a difference in growth quarter-on-quarter. Both of them have growth vectors that we are participating in. But the lumpiness that you see between the two is purely a market timing or a build-out timing. Some of the industrial was from some of the slowdown in the solar deployment in China. That's temporary. As you go from a tariff to a market pricing, there's a shift in there. We see that kind of a temporary and will continue to grow. We talked about some of the industrial growing because of the AI data center power requirements that I highlighted like energy storage system driven by AI, but we called those out in our Industrial market. So that's some of the growth vectors in Industrial. And Automotive, obviously, it's our major market. We -- you know about the growth vectors that we have there. So both are growing, but the delta is purely market-driven and macro driven. So I wouldn't read anything into kind of the deltas in the few quarters here short term.

The next question will come from Blayne Curtis with Jefferies.

I just want to ask about normal seasonal for December. Obviously, your company has gone through a lot of changes. But I think in the past, particularly auto has been up in December. So I'm just kind of curious how you're thinking about this guide, which is down 1%. Do you feel like that is a more normal range for you? Or do you think you're undershipping the market?

Yes. So as I mentioned, our normal seasonal pattern for Q4 is flat to down 2%. I think it's very positive that we've gone from the stabilization to now seeing seasonal patterns. I think that's the step -- the first step to recovery. As we think about the guidance there in Q4, both Auto and Industrial, we think will be down low single digits. The Other bucket will be up kind of mid- to high single digits. Hassane mentioned it, right? I don't think you should kind of read into the lumpiness of the autos just because of the ramping of programs and timing, but that's how we kind of think about Q4 as it laying out right now.

Perfect. And then I wanted to ask you about that AI. I'm assuming straddles, Industrial and this Other bucket you have. So I'm just kind of curious, is there a way to think about as we try to layer on that growth, how it impacts those two buckets?

Yes. So the AI data center is reported in the Other bucket. Everything prior to the data center wall is in Industrial. So think about all the energy storage, energy infrastructure, that's sitting in Industrial. But AI data center specifically inside the four walls of the data center is in the Other bucket.

And our next question comes from Gary Mobley with Loop Capital.

I think last quarter, it was communicated the specifics to the revenue headwind as you exit noncore businesses. If I recall correctly, it was assumed to be a $200 million revenue headwind for this fiscal year, $300 million for next year. Is there any change from that outlook?

No change. For Q3, we exited about $45 million of noncore exits. That leaves about $55 million here for Q4. That's right in line with our expectations. And then you nailed it going into '26, there's about 5% of the 2025 revenue that doesn't repeat. So no change from what we're talking about last quarter.

Great. And I guess there's been some news, maybe it's a few months old now about the big analog player raising prices. How do you think that impacts sort of a pricing reset as we transition to the next calendar year?

I think we're expecting normal pricing behavior. I don't know if the other company you're talking about is something specific to them or not. But obviously, things can change. We're monitoring the situation always. As you can imagine, it's very dynamic out there. But right now, we're not expecting any of that in 2026. So you can think about it as if anything does happen, it will be upside.

And the next question will come from Quinn Bolton with Needham & Company.

I'm wondering if you could give us a little bit more detail on the Vcore Power, exactly what comes into the business with that acquisition. I think in the script, you mentioned Vcore for x86 and ARM processors. Obviously, there's a huge number of voltage regulators on the XPU side. Does Vcore help you on that? Or does that come from the existing ON product portfolio? And then I've got a follow-up.

You can think about it as a combination of both. So the way we look at the acquisition is it complements the product offering that we're already offering with Treo. It provides products also in the short term. I talked about revenue generation coming here in 2026. So that gives you time to market while we integrate those architectural and product function into our base Treo platform. So it's a very synergistic approach that gives us the acquisition itself, time to market. And in long term, it gives us an architectural advantage from a performance perspective once we leverage the performance of Treo from a technology base.

So reading between the lines, are you taking those Aura products as they are today into the market for '26, but longer term, you'll redesign them using the Treo platform to get better performance?

Yes, yes.

Got it. And then just you guys mentioned the entry into the vertical GaN market. GaN to date hasn't been used that much in the high-power segments of the market, I think, because of reliability issues. Can you just address how do you feel the vertical GaN technology compares with lateral GaN on reliability? And can you give us any sense on when you think that might start to go into production?

Yes. So it's -- so I'll tell you, vertical GaN is better on the reliability side. It has all the inherent features from the lateral GaN, but better on reliability from a die size perspective. The one thing you need to understand the barrier for lateral GaN to be used in high-voltage application has really been the fact that lateral GaN to get it to high voltage, you have to go laterally, which makes the die size not competitive versus other similar functions. When you look at the vertical GaN, the current goes vertically, which means that we can go higher and higher voltage without increasing the die size. So not just from a performance perspective, but also from a commercial competitiveness perspective, not just the reliability. So we believe we've solved those. We're sampling. We have lead customers in our -- both in AI and automotive. So we're excited about that, that we crack that code. It is a breakthrough technology. I don't believe anybody is able to sample such technology outside. So it gives our customers the optionality to have really a broad portfolio of high voltage, high-efficiency products. So anytime you need high voltage and high switching frequency, vertical GaN is the solution, the answer.

And our next question will come from Joe Quatrochi with Wells Fargo.

I was wondering with a quarter to go, any sort of color you could provide on your expectations for silicon carbide revenue growth this year?

We didn't provide any guidance on silicon carbide, but I'll tell you silicon carbide is coming in exactly where we expected. We continue to gain share in our end customers. And our position in China remains unchanged as new products are ramping. I mentioned a couple of examples here. One is the NIO launching a new brand where we were designed into that new brand with silicon carbide. And a broader deployment now in China EVs through a leading Tier 1 in China that gives us really exposure to beyond just the top 10 OEMs that we've been engaged to. So that gives you a little bit of an outlook or a feel into our penetration of silicon carbide will continue to increase, and we will continue to gain share.

And as a follow-up, I was wondering if you could talk about the rate of short lead time orders that you're seeing and how that compares in the third quarter relative to prior quarter? And if you're seeing any increased visibility?

So our lead times actually pushed out slightly. We're kind of in the mid-teen weeks. We're up around 20 weeks or so now. I don't think there's been a significant change to the short lead time orders at this point. Customers are layering in backlog as they have visibility. We probably have seen order patterns that continue to improve, which gives us that confidence in the stabilization right now.

And our next question will come from Josh Buchalter with TD Cowen.

I was hoping you could provide a little bit more color on the revenue by geography. It seemed like there was a lot of volatility this quarter with Americas up so strongly and in particular, China down. Could you maybe elaborate on some of the drivers there? Was the Americas strength led by your lead customer? And what's going on in China?

Yes. So there's -- I think in our prepared remarks, we laid out the quarter-over-quarter changes on each of the markets. Now there is some kind of movement of orders between some geographies as well. A large customer is now placing orders out of Japan versus Europe. So I think if you normalize for that, the Japan comes down slightly, Europe goes up a little bit. The rest of it, I think, is just kind of what we're seeing as a normal pattern at this point. So not a lot to read into those bigger swings.

Okay. And then I was also hoping you could elaborate on why -- what you're seeing now and why it's the right time to start building up die bank inventory and taking utilization rates up, especially ahead of a couple of down seasonal quarters. Maybe how we should be thinking big picture about your capacity planning with those utilization rates?

Yes. Look, we've been -- I think we've been very disciplined on utilization versus inventory versus outlook and demand. I think we've proven that the formula works. We're not sitting here on a ton of inventory. Our inventory -- our base inventory is actually closer to the low-end of our target, about -- which is 110 to 120 days. I think we're sitting at like 112. So I think Thad mentioned that the die bank inventory we're building is really for the mass market. For the last kind of 2 to 3 quarters, we have been consistently talking about how we are going to be growing. Our customer count increased almost 20% year-on-year just in the mass market. Therefore, the demand is there for that, and we will make sure that we have it in die bank internally so we can respond to changes in demand that usually come from the mass market. So I think we do see the business justification for it, but that doesn't mean that we're going to be building blind. We will maintain our targets. We will maintain inventory and all of our metrics within the range that we've previously outlined. So I don't -- I see this as business as usual, really.

Yes. And just to point out also that even with that die bank increase, our inventory actually declined quarter-on-quarter, $39 million. So it's a mix shift within our base inventory. So to get a better profile of inventory for that mass market.

And the next question will come from Tore Svanberg with Stifel.

Hassane, with the recent acquisitions, I know you have a slide that talks about power delivery from grid to processors, and the content per rack going from maybe a few thousand dollars today to maybe as much as $50,000 by '27 or so. I mean, do you have all the IP and all the building blocks right now to get there? Or is this sort of more of an opportunity and you still need to build out a few more things before you get to those types of numbers?

I think with whatever we need, call it, in the next couple of years, we either have it or are working on it, both organically and inorganically. Obviously, the ecosystem is evolving. Things that are needed 3, 4 years from now are slightly different. We believe we have a very full portfolio of the IP that we need, and we will be creating products very quickly based on that IP. So we have -- you can think about it as we have built a toolbox with all the IP and technology, and we are quickly deploying products. I mean you've seen us double the number of products in Treo overall year-on-year, which we remain on track to do. You're going to see kind of that same mindset on AI data center along with Automotive and so on. So we do have the toolbox. We do have the IP. We developed it internally and/or acquired it. And we will be deploying it to win in these markets to capture a lot of that share from the dollars you mentioned on the rack.

Great. And as my follow-up, and I want to just take a step back on vGaN. So could you just give us a little bit of history here? I mean I know it's obviously in your own Syracuse fab, but how many years has this been in development? Maybe back to Quinn's question, when do you start to expect some revenues here? Because obviously, this is a very unique approach to GaN. So any sort of historical context and future revenue contribution milestones would be great to know.

Sure. So we started working on it through acquisition of IP and assets back in 2024. Since then, we've "turned on" the fab, launched the first products, first products from a, call it, electrically speaking or yielding or functioning. Therefore, we were very aggressive in our deployment with samples to customers. We have lead customers in our major markets of Automotive and AI data centers that are currently evaluating the first-generation samples, and we're already working on the second generation. We expect revenue, you can think about it in the '27 time frame.

And the next question will come from Christopher Rolland with Susquehanna.

So yes, my questions are really around AI as well and this what seems like a bigger push over the last few quarters. Just some of these applications that you mentioned, I wanted to know if you could address, could you do things like solid-state transformers? It sounds like you're in the PSU 48-volt bus converters. I guess the last one would be hot swaps as well. Do you address these? Or do you plan on addressing these over the next few years?

So we addressed every single one of them already. So when I refer to -- and we have it online, too, when we refer to our ability to address the power tree, that was my answer before as far as how do we differentiate. Our ability to address already today the whole power tree, including all of the IP and functionality required that you have mentioned some of them is the differentiation we bring. So the answer is yes to all. We do that today, and we will continue to expand that portfolio as we gain share.

Excellent. And Hassane, secondly on silicon carbide. As we kind of digest that growth outlook, perhaps you can talk about some of the moving parts like geographically or even across industries. And lastly, do you have the ability to convert to 300-millimeter wafers? We're hearing about the potential for new applications on 300.

Yes. So well, first off, there's a lot of changes in the silicon carbide as new opportunities open up. For example, a few years ago, silicon carbide in AI data centers was not even a conversation point. Today, it is, and we are gaining share and really design into the PSUs with our JFET and even our silicon carbide MOSFETs. So those are new applications that our legacy with silicon carbide in automotive allowed us to really tackle very quickly and gain share with products we already have. In Automotive specifically, the silicon carbide approach was for battery electric vehicles or BEVs. As now you see a resurgence of a mix into plug-in hybrids or range extender EVs, silicon carbide is now getting designed in even in plug-in hybrids, which historically has been assumed to remain on IGBT. That's not the case, and we are gaining share in the plug-in hybrid market with our silicon carbide. So within the market itself, there's new opportunities and really breadth of opportunities that just a few years ago, when we started on this journey, was not part of even our addressable market because it wasn't there. As far as geographical, I would say I don't expect a change in the geographical outlook for silicon carbide specifically because to a first order, it's going to match where the electrification, whether it's full electric vehicles or plug-in hybrids is going to come from and where the AI data center deployment is going to come from. And that puts it strong in China and the U.S. And following behind that is Europe and Japan.

Excellent. And 300-millimeter?

300-millimeter, we've seen it, but my point is it's too far from now. I don't think 300-millimeter opens up new applications. It just -- it's a different, call it, throughput, just like 6 to 8. I've always said 6 to 8-inch provides us an additional capacity from the number of die per wafer. We see the 300-millimeter the same, but it's very, very early in development today. I wouldn't put that in any short-term models or anything. But today, we have been -- just I'll use the opportunity to give you an update on our 8-inch. Our 8-inch is in production. We're running 8-inch in our fab at 350-micron thickness, so best-in-class, and we will be shipping production on track in '26. So the 8-inch is full on, and then we're always looking at what's next to come both from a device like the SiC JFET or MOSFET, but also from a technology.

And the next question will come from Harlan Sur with JPMorgan.

Back to the mass market strategy, your long tail of small- to medium-sized customers, this has been a bright spot for the team, right, solid customer count improvements. It serves through distribution, rich gross margins. How big is this segment as a percent of your total distribution revenues? And how did this subsegment do in the September quarter relative to your overall disty business?

So let me give you a breakdown of the distribution revenue that may help you get there. So roughly about 58% of our business goes through distribution. About half of that is fulfillment, half is demand creation, right? So if you think about that half, not all of that's mass market. When we think about mass market, we're thinking small customers, right? We at onsemi, maybe don't know their names, right? They are emerging customers. The distributors do a good job of identifying the opportunity. So you can think about it as being a subset of that half. Maybe it's 25% of the total distribution revenue, somewhere in that kind of camp if you think about it. If they're a medium or large customer of that distributor, we still have -- we still track that. I wouldn't put that in the mass market.

Got it. Okay. And it was good to see the technology and portfolio expansion on the wideband gap with your vertical vGAN technology. As you mentioned, I think, Hassane, looks like this was the technology that you acquired through the acquisition of NexGen late last year. Did the acquisition also include the DeWitt Syracuse fab facility? Or was that already a part of ON? And then it looks like they were able to develop this very differentiated technology, but not able to commercialize it. So what has the onsemi team done to take the technology beyond proof of concept to commercialization?

Great question. So yes, the fab was not part of our base fab. You can think about it as a fab that came with the technology given the differentiation of the technology. You're absolutely right on -- it's such a breakthrough and differentiated technology, very difficult to make. What the onsemi team has brought is our ability to manufacture wideband gap and the team's capability to be able to scale new technologies very quickly and reach maturity very quickly than, call it, a start-up. By the way, I will mirror this to what we've done with GTAT and silicon carbide. If you recall, same questioning, same conversations, can you guys pull it off? Why would you pull it, the others didn't? And look where we are today. You can think about it, our capability has already been proven with the GTAT acquisition and building a franchise in a couple of years that gives us leadership. You can imagine that same muscle, that same knowledge and that same team is going to do exactly that with vGaN.

And the next question will come from Jim Schneider with Goldman Sachs.

Hassane, you talked about the fact that customers are not willing to restock at this point or you're not seeing that effect. Can you maybe talk a little bit about when you speak to OEMs, what they would need to see to get more confidence to restock? And is that broadly applicable to the distributor side as well?

Yes. Look, well, first, I'll answer the distributor. I think distributors are -- from a mass market, Thad said it, we are increasing our die bank internally because we want to be able to make sure we see the, call it, the shelves as customers pull on the mass market. The OEM is slightly different. The OEMs, what they need to see, one is a credible demand signal. Think about it consumer level confidence, consumer level demand signal that people are going to buy cars or people are going to buy power tools or whatever the market is. That has to be seen. And the biggest thing that they want to see, which we do also is stabilization in the geopolitical aspect of it. As they're working on shuffling and changing logistical models and manufacturing sites and so on, they're not going to be replenishing given the changes that they're going through. So what I would say is consumer confidence and geopolitical stabilization will start adding more and more confidence for OEMs to restock.

And then maybe as a follow-up, give us a little more visibility on what's happening with your Other segment for a minute. It sounds like data center is doing very, very well for you. Maybe talk about what some of the offsets are that might be headwinds you saw in this quarter and then maybe what you're seeing going forward?

So for Q4, I mentioned that we think that Other segment is going to be up mid- to high single digits. Now there's some normal seasonality in our noncore markets there that helps that you have AI data center that's growing as well in that market. I think those are the big drivers if you sum it up.

And then, of course, we have the exits that a lot of it lands into the Others bucket that's offsetting the growth. So net-net, growing is actually means the strategic market like AI, data center and so on within that is growing very, very nicely.

And the next question will come from Joe Moore with Morgan Stanley.

I wonder if you could give us some sense of the Automotive market by region. Any sort of different behaviors that you're seeing? And I guess, particularly on China EV, there's been sort of a lot of noise in both directions. Can you just talk to the health of that market?

Yes. Look, I think from a market, of course, we've always expected adjustments in that market. I've always said there's over 100 brands. So between consolidation, between success and not success, the only strategy we have, which we've been executing to and it's worked very well for us is customer diversification. So you've heard us always adding new and new customers, leading customers in the top 10, which drive a lot of volume. And then secondary is trying to reach into that tail of OEMs. So we're not sitting here picking winners or not winners in China. We want to have the majority market share across the market. And as shares shift between them, our customer diversification strategy will work to our advantage. We've proven that very well over the last few years. We're gaining share consistently across a broad range of OEMs and brands have worked for us to really derisk the lumpiness that you're referring. But I don't see that as any change from the headlines. So our strategy is working, and we'll continue to execute to that while we kind of fine-tune it as things change because things do change rapidly.

Great. That's helpful. And then you addressed the Nexperia's situation. But I guess I'm just trying to figure out why that isn't a bigger deal. We've listened to some of the Tier 1 auto suppliers, and they seem quite anxious about the situation. Like shouldn't that be the catalyst for them to start building up inventory to sort of deal with the geopolitics of the situation? Or just why isn't that something that's a bigger deal for you guys in the next quarter or 2?

Well, we're here to support, but I'll make a comment on the Tier 1s panicking. I've been saying that inventory is low for the last 2 years, and we're draining inventory below critical levels. Whether Nexperia or not, any trip in the supply chain is going to cause a chain reaction, and this is the proof. The only way out of this is place the backlog with visibility and we will start planning and shipping. So we are seeing it. We are responding to it, and we will keep supporting it. But regardless of how the next few quarters go, we need the replenishment cycle. We need to make sure that the Tier 1s and the OEM have safety stock in order to buffer any disruption. That's the only solution. We've learned a hard lesson in COVID, and here we are again.

And the next question will come from Harsh Kumar with Piper Sandler.

Hassane, if I can dare say that you seem somewhat, somewhat cautiously excited about your end markets for the first time in a long time. So if I can ask you a question on Auto, it's a two-parter. Could you give us a hint of maybe what backlog or bookings were? I'm trying to gauge that relative to your stabilization comment. And if you are talking about stabilization in Auto, then I'm looking at your 6% odd growth that you put up in the September quarter, is that seasonal growth? Or is that better than seasonal growth? And if it's better, then, of course, what drove that?

Yes. Look, I think I'll go back to the prior answer that I gave earlier. I don't look at the quarter-on-quarter. I would recommend you shouldn't either. I got to look at it first half, second half. And we've always said the second half of the year is going to outgrow the first half of the year in our end markets. Remember, Auto, we said the bottom was going to be in Q2. So that has been the case, and we're going to grow from there. Grow meaning closer to demand, but no restocking yet. So that's coming in exactly as we expected. So that quarter-on-quarter, I wouldn't talk about seasonality within markets and so on. I would talk about the lumpiness in project ramps, some projects ramped in Q3 versus ramping in Q4. Those, I don't think, are a read on how the market is doing. Visibility, however, with stabilization, we get better visibility. Not where we would like to see it, but it improved and we're getting better visibility. But again, there's more work to do to get the visibility. So that's what I can tell you about where we are in Automotive. I'm cautious, but I am also looking at the data in order to sound like I do. Our work is not done. It's not all behind us. But I think what you've seen from us is we will manage to what we see, and we will deliver the results that we promise. That's the consistency. Of course, we all wish it were different. We all wish it was way better than sometimes it is. Some of my peers did, but we've been very consistent, and we're going to continue to manage the company with discipline, whether in inventory, cash flow or really R&D investments in differentiated technologies.

Fair enough. Maybe one for you, Thad. As sort of you look at stabilization, I understand you'll need to ramp up your factories and fabs to be able to get to that 40% level. But is there a revenue number that I can think of where you start to get close to that 40% number? Or is it just purely a function of utilization and it ebbs and flows depending on how much die bank inventory you're building?

Yes, it's utilization driven, right? So we talked about every point of utilization is 25 basis points to 30 basis points of gross margin improvement. That math still holds. So as we look into the '26, utilization is going to drive the margin.

This will conclude today's question-and-answer session. I would now like to turn the call back over to Hassane for closing remarks.

Thank you all again for joining us today. Before we conclude the call, I want to recognize the outstanding efforts of our global teams. Their focus and execution continue to drive our results and help us deliver for our customers and shareholders. We're encouraged by the signs of stabilization across our core markets and remain focused on delivering differentiated solutions and operational excellence for our customers. We are committed to being a reliable and trusted partner and continue to raise the bar on how we support their success through technology leadership, responsiveness and a deep understanding of their evolving needs. We appreciate your continued support and look forward to updating you next quarter. Thank you.