Tower Semiconductor Ltd Q1 FY2026 Earnings Call

Good day, and thank you for standing by. Welcome to the Tower Semiconductor First Quarter 2026 Results Conference Call. Operator provided instructions to participants. Please be advised that this conference is being recorded. I would now like to hand the conference over to your first speaker today, Noit Levi-Karoubi. Please go ahead.

Thank you. Hi, everyone, and thank you for joining us today. Welcome to Tower Semiconductor's First Quarter of 2026 Financial Results Conference Call. With us today are Mr. Russell Ellwanger, our Chief Executive Officer; and Mr. Oren Shirazi, our Chief Financial Officer. Before we begin, please note that certain statements made during today's call may be forward-looking and are subject to risks and uncertainties that could cause actual results to differ materially. These risks are detailed in our SEC filings, Form 20-F and 6-K as well as filings with the Israeli Securities Authority, all available on our website. Tower assumes no obligation to update any such forward-looking statements. Our first quarter 2026 results are prepared in accordance with U.S. GAAP. Some of the items that are presented may include non-GAAP financial measures as defined under SEC Regulation G. Reconciliations to GAAP figures and full explanations are provided in today's press release and financial tables. For your reference, a supporting slide deck is available on our website and integrated into this webcast. With that, I'd like to turn the call over to our CEO, Mr. Russell Ellwanger. Russell?

Thank you, Noit. Hello, everybody. Thank you for joining our call today. The first quarter of 2026 was solid, providing a strong foundation for the high growth we expect this year. We maintained strong financial performance with continued execution of our strategic priorities. Our first quarter 2026 revenue was $414 million, 15% year-over-year growth. First quarter net profit was $65 million, 62% year-over-year growth, yielding 16% net margin, up from 11% in the first quarter of 2025. Looking ahead, we guide the second quarter of 2026 to be the highest revenue in the company's history with a mid-range revenue guidance of $455 million, plus or minus 5%, representing a 22% increase as compared to the second quarter of 2025 and a 10% growth quarter-over-quarter. We strongly reiterate our target of quarter-over-quarter revenue and margin growth throughout 2026. We continue to strengthen our alignment and partnerships with our photonics customers through the execution of long-term customer commitments, contractually representing $1.3 billion revenue in 2027 with significantly larger valued contracts for 2028, backed by approximately $290 million in prepayments already received from our largest silicon photonics customers. This reflects the strength of our offerings and our customer partners' confidence in our ability to meet the continued growing demand of next-generation AI data center architectures. Importantly, these reservations do not represent the entire expressed demand of these customers nor the extent of our planned shipments to these customers, and do not include additional wafer shipments to our broader base of more than 50 active silicon photonics customers serving various end market applications. These commitments, together with our continued technology leadership and strategic expansion of 300-millimeter and global manufacturing capacity, provide us with enhanced revenue visibility and confidence in sustained profitable growth. Our recently announced restructuring deal in Japan in TPSCo marks a significant milestone in advancing our long-term 300-millimeter strategy. By transitioning to full ownership of the 300-millimeter factory Fab 7 in Uozu, we are creating a more focused and scalable platform to support growing customer demand, particularly in our differentiated optical photonics technologies. Full ownership allows us to expand and build upon a facility that is running multiple fully qualified, high-volume application flows and importantly, at present volumes is already profitable. The 300-millimeter expansion tied to the expected approval of METI grants is designed to be strategic, operational and capital efficient. With access to adjacent land, we expect to further build out and scale up to four times current levels, generating a meaningful long-term growth engine anchored in high-value technologies. This approach leverages existing customer qualifications with increasing demand, allowing incremental capacity to translate into revenue and cash flow almost immediately as new tools are installed. This positions our 300-millimeter platforms not only as a key driver of future growth, but also as a structurally stronger contributor to profitability, reinforcing our overall financial model and long-term value creation. Additionally, we have entered into a long-term supply agreement with Nuvoton for Fab 5 Tonami. This will ensure manufacturing continuity for our 200-millimeter customers under terms that are mutually beneficial. Moving specifically to first quarter of 2026 performance. This year has begun in a very strong fashion, led by Silicon Photonics with revenue growth of three times year-over-year. All major technology offerings demonstrated year-over-year growth with imagers up 9%, RF-SOI up 12%, power management up 10% and silicon germanium up 24% year-over-year. Please see Slide 4 as reference for Q1 revenue breakdown by technology. Focusing on RF infrastructure, last quarter was truly excellent, both in our team's execution of aggressive capacity expansion as well as in demonstrating new breakthrough technology milestones. First, we continued a strong ramp of 200-gigabit-per-second products for multiple customers while continuing to support strong demand in older products by taking full advantage of new capacity coming online. We are in the midst of a silicon photonics production ramp in each of Fab 2 Migdal Haemek, Fab 3 Newport Beach, Fab 9 San Antonio and Fab 7 Uozu Japan 300-millimeter. Among this, we successfully achieved in Q1 first flow cycle revenue shipments from both Fab 2 and Fab 7, the latter having achieved an impressive 95% yield for the first silicon photonics wafers leaving the factory. Our expansion remains on track to grow silicon photonics capacity five times from the base of our Q4 2025 wafer revenue shipments by the end of this year 2026. In 2027, we anticipate our focus will turn primarily to additional 300-millimeter capacity expansion in the Uozu factory supported by the expected full factory ownership. Next, we achieved a number of next-generation technology breakthroughs working with several of our key customers. This quarter, we announced the demonstration of an all-silicon 400-gigabit-per-lane Mach-Zehnder Modulator with our strong partner and optical industry leader Coherent, which is one of our customers having signed a high-volume, long-term contract. With OpenLight, we recently announced a heterogeneously integrated 400-gigabit-per-lane indium phosphide electro-absorption modulator on our PH18DA platform. In addition, we made strong strides towards bringing thin-film lithium niobate to high-volume manufacturing and announced our partnerships with Lightwave Logic and NLM Photonics to bring organic polymers to high-volume production for next-generation compact modulators. Just prior to the Optical Fiber Conference, we announced our partnerships with Salience Labs and Oriole Networks to manufacture advanced silicon photonics-based optical circuit switches, both using our PH18DA platform with heterogeneous integrated indium phosphide optical amplifiers to achieve high bandwidth and ultra low-latency optical switch solutions for AI data center scaling. Last but certainly not least, our partner SCINTIL Photonics announced availability of the world's first heterogeneous integrated dense wavelength division multiplexed (DWDM) laser sources, designed for near-package optics and CPO-based AI infrastructures. Most market analysts forecast that pluggable optical transceivers will remain the dominant format through the end of this decade. We do see extra dense pluggable optics, XPO, being led by Arista with the aim to extend the served generations of pluggables and highlighted by Andy Bechtolsheim in his Optical Fiber Conference Executive Forum panel presentation, and near-package optics, eventually also co-package optics, emerging and coexisting with pluggables for the next several years. We are thus preparing to ramp these technologies as well. At this year's OFC, Tower Silicon Photonics was on display in leading XPO and near-package optics demonstrations. We are already seeing strong demand for near-package optics products in 2027. Given the strong customer traction, it's our expectation that Tower Silicon Photonics will continue to lead in these new optical form factors. Near-package optics is likely to ramp over the next several years and precede a significant ramp in co-package optics for our primary customers. However, we are investing heavily in several co-package optics technologies, namely in-house 200-millimeter and 300-millimeter hybrid bonding with through-silicon vias to seamlessly attach silicon photonics to electrical ICs, laser sources for both more traditional as well as DWDM architectures for use in CPO implementations, and reduced-size high-performance modulators for use in space-challenged CPO form factors. In addition to silicon photonics, our silicon germanium platform is experiencing unprecedented demand for use in drivers and transimpedance amplifiers for optical transceivers and also for active copper cables that can be an attractive alternative to optical for short-distance scale-up architecture. Additionally, our RF silicon germanium technology is in the midst of a strong ramp for LNAs in a Tier 1 mobile platform. In silicon germanium, we recently announced our partnership to produce high-power U.S.-made silicon germanium beam-forming ICs designed for defense radar and satellite communication applications. Fabricated at our U.S. sites, these chips aim to secure domestic supply chains, offering superior performance for critical next-generation defense systems. As our sites turn to space, whether for data centers or for global satellite connectivity, we see our silicon germanium platform being well suited to support these growing applications. Looking at RF mobile, we continue to move our RF-SOI 200-millimeter technologies to 300-millimeter to take advantage of finer line and other enhanced capabilities offered at 300-millimeter, while repurposing this 200-millimeter capacity for higher-margin silicon photonics and silicon germanium capabilities. Substantial improvements of our Ron-Coff relative to competitors and reduced layer count are creating strong design win momentum that positions our 300-millimeter RF-SOI platform for sustained growth over the next multiple years. In power management, we have seen year-to-year revenue growth in both our 200-millimeter and 300-millimeter BCD offerings. In the last quarter, we announced the release of our latest power platform, Gen3, achieving on-resistance below 1.5 milli-ohm millimeter squared for key devices with operating voltages above 10 volts. Such low on-resistance enables high power conversion efficiency in a variety of applications and places our offering at a very competitive position relative to other foundry offerings. Using our technology, our lead customers have demonstrated a 15% reduction in power conversion losses, quite significant as compared to the highest-efficiency alternatives. Some of the end markets where we have seen revenue growth have included consumer mobile and automotive. In addition, as the AI data center power delivery market transitions to 800 volts DC bus at the rack level, we see a significant growth opportunity ahead in smart power stages and point-of-load converters designed with our BCD offerings. Lastly, in consideration of the value we are offering, our 200-millimeter BCD pricing has increased by 13%. Image sensors: the fastest-growing CIS segments are automotive, industrial, machine vision and high-end video cameras. Growth in each of these areas is concentrated in the high-end portion for high-resolution, high dynamic range with sensitivity to low light, and global shutter technology is required. Tower's global shutter technology, combined with wafer-to-wafer hybrid bonding, provides best-in-class performance in terms of low noise and high sensitivity and allows high resolution. Additionally, we are developing an ultra-high-density in-pixel capacitor to provide best-in-class dynamic range, especially for the automotive market. We won a second high-performance automotive product this past quarter. Significantly, we are fully qualified with the next-generation high-end video sensor with a leading high-end photography camera maker awaiting their product launch. Turning to utilization. For the first quarter, utilization rates for Fab 2 were around 60% as silicon photonics and silicon germanium qualifications continue. Fab 3 operated at 80% utilization. Utilization was slightly constrained due to adding newer silicon photonics and silicon germanium processes. We expect utilization and output to increase in the second quarter. Fab 5 was at 75% utilization. Fab 7 continues to be fully utilized, well above our 85% utilization model. Fab 9 utilization was at 80%. With that, I'd now like to turn the call over to our CFO, Mr. Oren Shirazi. Oren, please?

Hello, everyone. Earlier today, we released our financial results for the first quarter of 2026. I will now review the highlights of these results as well as the balance sheet and CapEx investments. Looking into the P&L: Revenue for the first quarter of 2026 was $414 million, representing 15% year-over-year growth compared to $358 million in the first quarter of 2025. Gross profit for the first quarter of 2026 was $111 million, an increase of 52% compared to the first quarter of 2025. Operating profit was $65 million, 96% higher year-over-year. Income tax expense line of $6.5 million in the P&L reflects an all-in 9% effective tax rate, which is better compared to our model for which we estimate all-in tax rate to be above 15% following Pillar Two regulations. The reason for it is the inclusion of a nonrecurring income tax benefit recorded for the first quarter of 2026 in relation to TPSCo, our Japanese affiliate. Net profit for the first quarter of 2026 was $65 million, an increase of 62% or $25 million compared to net profit of $40 million in the first quarter of 2025, reflecting 16% net margins compared to 11% net margins for the first quarter of 2025. Earnings per share for the first quarter of 2026 were $0.58 basic and $0.57 diluted, which is 61% and 63% higher year-over-year, respectively. As we announced, we received $290 million of silicon photonics customer prepayments towards 2027 capacity reservation. These customer prepayments are included in the balance sheet as of the end of March 2026 as short- and long-term customer advances and are included in the cash flow report for Q1 2026 cash from operating activities. Continuing on the balance sheet, our balance sheet continues to be strong, evidenced by the following indicators and financial ratios. As of the end of March 2026, our assets totaled $3.7 billion, primarily comprised of $1.5 billion in fixed assets net, predominantly comprised of fab machinery, and $2 billion of current assets. Current assets ratio is very strong at about 5.6x, while shareholders' equity reached a record of $3 billion at the end of March 2026. Additional evidence of the strong balance sheet and financial position is Standard & Poor's Maalot, an S&P Global Ratings fully owned company, which on May 5, 2026, completed its annual rating review for the company, reaffirming its ilAA rating and raising its outlook for the company from a stable outlook to a positive outlook. I would like now to describe our hedging activities. In relation to the Japanese yen, since the majority of TPSCo's revenue is denominated in yen and the vast majority of TPSCo's costs are in yen, we have a natural hedge over most of our Japanese business and operations. To mitigate part of the remaining yen exposure, we are executing zero-cost collar transactions to hedge currency fluctuations. Hence, while the yen rate against the U.S. dollar may fluctuate, there is limited impact on our margins. In relation to the Israeli shekel currency, while we have no revenue in this currency, since a portion of our cost in Israel is denominated in shekel, we also hedge a large portion of such currency risk by engaging zero-cost collar transactions to mitigate this exposure. Hence, while the Israeli currency rate against the U.S. dollar may fluctuate, there is limited impact on our margins as indeed we have experienced over the past few quarters in which period the Israeli currency appreciated strongly compared to the U.S. dollar. Now moving into the CapEx investment plan. As we announced in recent quarters, in order to support the increasing silicon photonics and silicon germanium demand, we are executing a $920 million investment plan in capacity and capability of SiGe and SiPho in our 8-inch fabs in Israel, Newport Beach, Texas and also in our 12-inch Uozu fab in Japan. This investment is on track in terms of purchase orders issued, technology and process qualification and the ramp plan. Approximately 40% of the above-mentioned $920 million CapEx has already been paid and is included in our cash flow for investing activities for past reported periods, while the remaining 60% are expected to be paid throughout 2026 and 2027. All these CapEx, current and future investments, are fully reflected in the model we presented in February 2026. Under this model, we target $2.8 billion in annual revenue, $1.12 billion in annual gross profit, $900 million in annual operating profit and $750 million in annual net profit. That concludes my prepared remarks. Now I'd like to turn the call back to the operator so we can take your questions. Operator?

Operator provided instructions to participants. Now we're going to take our first question and it comes from the line of Mehdi Hosseini from Susquehanna Financial Group.

Yes. A couple for me. Russell, if you just think about the big picture and the longer-term trends, is there any way you can help us understand how opportunities for Tower as it relates to silicon photonics and silicon germanium changes as we migrate from a pluggable transceiver to optical? In other words, how does your content change as you make this transition from a traditional pluggable transceiver to optical?

As we go forward, as far as content itself, certainly you'll still need lasers. We're driving integrated lasers. We're driving different material modulators for higher speeds, and we're also driving through-silicon vias to be able to do some 3D packaging and tie the silicon photonics to the electrical ICs. With XPO there are many more channels, and with MPO there are again many more channels, so greater modulation.

Sure. I asked the question because the way we, in the investment community, see the optical transceiver manufacturing evolving, there are several different solutions by different foundries. I was just trying to understand if Tower has become a leading foundry partner for transceiver manufacturing for pluggable, how does that change as we move into optical where there are alternative technologies?

We expect that what we have with pluggables will transfer nicely into near-package optics and will extend as well into co-package optics, which will be several years down the road. If you look at the silicon photonics ports itself, from 2025 the amount of silicon photonics ports was $30 million, driving to 2028 silicon photonics ports to $137 million according to a LightCounting report. At the same time, total ports go from $90 million to $205 million. This is a function of data center build-out, but the amount of ports becomes much greater. Pluggables are not going away; pluggables will stay extremely strong at least through 2030. We expect that the first things to come at a higher rate are near-package optics where we have multiple design wins presently and expect a reasonable volume that we'll be shipping in 2026 and 2027. We also have a very strong position in XPO that was highlighted at OFC, with two strong demonstrators using our silicon photonics. So the growth of silicon photonics ports is projected at 4.5x from 2025 to 2028. The growth is substantial, and I believe we are by far the leader in pluggables. I don't see any reason that that should change. In the short term, XPO and MPO form factors should transition nicely with Tower maintaining leadership.

Great. And just a quick follow-up. I understand investment communities focus on silicon photonics. And thank you for identifying revenue opportunities, especially what you have contracted for 2027. Is there any way to think about the ratio between silicon photonics and silicon germanium because silicon germanium is also used in transceivers. So if you have extended visibility on silicon photonics, would that give you visibility on silicon germanium? Is there a ratio that we could use to understand opportunities that you have focusing on silicon germanium?

Yes, you're correct. It's very much hand in hand. As the ports increase and the movement to silicon photonics increases, silicon photonics grows. As the ports increase, the need for TIAs and drivers increases at roughly the same ratio. The difference between the two comes into the advancement of the silicon photonics technology and the fact that although they both demand very good margins, silicon photonics commands a higher margin than silicon germanium. So silicon germanium revenue growth is lower than silicon photonics revenue growth, but as far as the amount of units, they go pretty much hand in hand.

Now we're going to take our next question, and the question comes from the line of Cody Acree from Benchmark.

Congrats on just a stellar quarter and guidance. Russell, maybe could you just give a quick clarification on the $1.3 billion commitment — is that for wafers to be delivered in fiscal 2027? Or is that wafer starts that then would extend delivery to 2028? Just trying to get my model correct.

Wafers delivered in 2027. What I stated about a higher volume for 2028 is wafers delivered in 2028. It is very important to note that the $1.3 billion is a contractual commitment. It does not represent the full volume demand from those customers with whom we have the contractual commitments. So the $1.3 billion, if you look at it compared to 2025 where we were around $230 million, is a large increase, but that $1.3 billion is not our full forecast for silicon photonics in 2027; we are forecasting substantially higher. But to answer your question directly, that is for wafer shipments.

I guess that then just begs the next question. With that level of visibility, what degree of visibility would you have to feel comfortable increasing your recently increased long-term model?

Good question. When you say long-term model, you might be referring to the 2028 model, which I view as somewhat midterm. Our focus for 2027 will be increasing 300-millimeter silicon photonics capacity, and the silicon germanium capacity as well on 300-millimeter. That is not included in any model right now. The visibility will be fairly short-term. The timing of updating a model to higher numbers, I would believe, will be within the next quarters.

Well, that's excellent, Russell. And a correction: long-term versus midterm. And then one last question, if I may. What is your view on integration of indium phosphide lasers? I know there's been a lot of industry concern about the reliability of integrated lasers as we move towards CPO versus an externally sourced laser alternative. I know that's a strength of yours. Can you talk about the reliability steps you've been able to make?

We see no reason that the integrated laser is any less reliable nor any more reliable than a discrete laser. The major focus is strong integration of the laser into the silicon photonics IC so that there is no potential downside of integrating the two. The activities with OpenLight and their platform are going very well. From our standpoint, we're very bullish about the integrated laser and additionally about an indium phosphide integrated modulator for 400G. Our platforms can allow both approaches: if a customer wants a form factor with external lasers, great; if they want to go with integrated lasers, equally great. We're somewhat agnostic to the customer's direction, other than the fact that we're pushing very strongly on very advanced 400G modulation. For advanced modulators, indium phosphide has tremendous advantages on form factor.

And Russell, since you brought it up, I'll sneak one last one in. In the subsequent press release to your earnings release, you mentioned a wide variety of modulation technologies you're pursuing. Would you handicap or prioritize any one of those over the other as being a better fit for Tower in the intermediate to long term?

No, I think we're confident in all. Thin-film lithium niobate can be architected in multiple ways: a separate lithium niobate modulator wafer side-by-side to the PIC, or integrated as a chiplet onto the PIC. We're pursuing both form factors with different customers. Indium phosphide is integrated onto the PIC. We're somewhat agnostic as to which one a customer would use. However, as you get to many channels, indium phosphide offers benefits on form factor.

And now we're going to take our next question, and the question comes from the line of Richard Shannon from Craig-Hallum Capital Group.

Apologies for the ambient noise here, I'm just about ready to board a plane. So if it's too loud, I'll try to call in later. But I have two basic questions. The first is regarding your capacity expansion in Japan. How should we think about potential revenue capacity scale for silicon photonics as you get the first tranche? And do you have any worries about being short of capacity in silicon photonics before that Japan capacity expansion has started?

Excellent question. We're looking and hoping and believing that we'll receive METI approval in the very short term. We've already begun with contractors to get planning done. Once we have METI approval, we'll most likely put in for permits. It's probably a 1.5-year timeframe between breaking ground and having a facility that can accept tools and start ramping. So best case for that capacity is in the first half of 2028. Regarding worries, it is an exciting problem. We're increasing capacity within the existing footprint at Fab 7, where we've approved capital spending for an initial phase, which we haven't publicly detailed. We also have the ability to grow capacity relatively quickly within existing footprint by using another factory within the TPSCo complex, an RI factory previously shut down and under Nuvoton Winbond ownership, where we could place a certain amount of tools to grow capacity quickly. Our plan is not to miss out on any upside or to be unable to meet customer demand in 2027 while developing the larger shell. Any tools we buy initially are applicable to the entire build and can be placed in interim facilities and later moved as needed. Hopefully that answers your question.

It does. My second question is on your announcement on the 400-gig-per-lane silicon modulator with Coherent. Can you use this technology or something similar with other customers? And across the 400-gig generation, how much of your business will be silicon versus thin-film lithium niobate versus indium phosphide versus other modulator technologies?

On the first question, the 400G modulator we press released with Coherent was based on Coherent's design. Our know-how and IP are available for any customer, but that specific modulator performance was due to Coherent's design on our platform. Other customers could design to our platform to get similar capability, but this particular design is Coherent's IP. On the second question, my view is that lithium niobate will come in fairly strong for one generation, but I don't think it will last for many generations; I think it will migrate to indium phosphide over time.

Now we're going to take our next question, and it comes from the line of Lisa Thompson from Zacks Investment Research.

I was wondering if you could talk a little bit about gross margins. It seems surprising how much they increased in the first quarter given revenues were sequentially down. Can you talk about how the margins went up? Was it more product mix, higher prices? How did that turn out?

Yes. It's consistent with the model we published last quarter. In our model, we assumed incremental revenue would come at 59% to gross profit over a baseline, which was 20% a year ago. Last year Q1 2025 gross profit was 20%. We assumed incremental revenue would come at 59%, and the fully built-out model will reach 39%. This quarter we achieved 27%. So we've moved from 20% to 27%, and that's the linear progression we expect toward 39% when we achieve $2.8 billion revenue.

So it's really just selling newer, higher-margin products?

Yes.

Okay. And then going back to the technology: is there anything you're concerned about regarding getting parts or being constrained? I know indium phosphide material is hard to get and constrained. Is that going to change anything or do customers just switch to something else?

You are correct that indium phosphide is constrained right now. In our case, it's not necessarily the finished laser product that's the constraint but the starting material. We have good plans and supply chain activities so that whatever constraint issues there might be, we're working through them. As these products ramp, which they should this year, I believe we'll be well positioned to meet demand.

Now we're going to take our next question, and the question comes from the line of Krish Sankar from TD Cowen.

I had two questions, Russell. One is where do you think your silicon photonics market share is today and where do you think it will be in the next couple of years given your longer-term contractual commitments from customers? I'm asking in light of GlobalFoundries ramping scale and TSMC ramping their COUPE platform. Where do you think your market share could head from today through the next two years? And then I have a follow-up.

That's a very interesting question. Others claim very high market shares, but I don't see how that's possible. I'm not going to give a percentage market share at present, but I think we're certainly the leader and by far the leading market share in silicon photonics presently, and I see no reason why that should change.

Got it. And then on advanced packaging: you have experience with hybrid bonding and CMOS image sensors. When do you think that ports over to the silicon photonics side? Is that a real advantage in the short term or more of a longer-term story?

I think it's a very strong advantage mid- to long-term, not so important for this year.

Now we're going to take our next question, and the question comes from the line of Mehdi Hosseini from Susquehanna Financial Group.

Just a couple of follow-ups. RF mobile was down almost 36% quarter-over-quarter. Obviously, it sets up a low base, and I expect growth for the remainder of the year. How should I think about RF mobile in 2026 versus 2025? I have a follow-up after that.

Indeed, Q4 2025 was an exceptional growth in RF-SOI compared to previous quarters. Year-over-year it was not reduced; it was a specific ramp in Q4 2025. Russell will address 2026.

As I mentioned, we're moving away from our 200-millimeter RF-SOI and transitioning to 300-millimeter designs to take advantage of finer line and enhanced capabilities, while repurposing 200-millimeter capacity for higher-margin silicon photonics and silicon germanium. The design win cycle is extremely strong. Looking at the multiyear forecast, RF-SOI on 200 plus 300 will increase as a total, but 200-millimeter is almost gone. In the very short term not everything has transitioned to 300 yet, so you are correct that it was down. If I look at the full year, I would see the whole year being down against the previous year even at 300-millimeter. Then I would expect that in 2027 and 2028 it will be record growth for RF-SOI at 300-millimeter. We have multiple design wins that have been awarded, but they're for phone models that first come out in 2028, meaning they would start needing wafers in the third quarter of 2027.

Got it. And then one follow-up. Looking at the overall analog foundry industry outside of RF infrastructure, utilization rates are improving and some peers have talked about higher wafer prices like-for-like. How should we separate mix effects, given RF infrastructure has higher growth and higher margin? Outside of RF infrastructure, should we assume there's some pricing power coming to you?

Pricing power comes from having best-in-class platforms that give customers advantages. If you have something that allows the customer to gain advantages using your platform, you can charge a premium. We don't raise prices indiscriminately because of capacity constraints; that's not a partnership model. There are times customers might need to go on allocation due to demand, but it's not good practice to raise prices simply because of capacity constraints. Our pricing benefits come with new generation platforms for new technologies. Every time we put out a new platform, the starting price point is higher than the previous platform. I did mention that our 200-millimeter power management pricing increased by 13%, which was a reevaluation of the value of that platform, not a reaction to a capacity squeeze.

And now we're going to take our last question for today, and it comes from the line of Cody Acree from Benchmark.

Just a couple of quick ones. Thematically, can you talk about the move to CPO and the industry's larger integrated players like TSMC and their CMOS integration of silicon photonics and their leverage as a one-stop shop for somebody like NVIDIA long term? How does Tower compete with that on a long-term basis with your hybrid strategy?

TSMC is a one-stop shop and there's nobody who can compete with them on deep digital integration. But NPO, XPO and pluggables will remain primary demand into the beginning of the next decade as CPO gains traction. On the CPO side, the benefit we add is by having PICs that are highly performant. In theory, there's no reason TSMC wouldn't use our PICs for their COUPE if our PIC was superior. The key is that we could become a reference design for major integrators, and that could be part of what they'd use with TSMC. We focus on modulation performance and advanced photonic components where we add unique value.

All right. And then maybe lastly for Oren: any thoughts on future tax rate, interest income, noncontrolling interest lines, any of the ancillary items of the income statement?

Financing and other income, which is about $10 million a quarter, I would expect to remain the same. Tax: on a regular model, we would expect at least 15% because of Pillar Two, and up to maybe 18% because we have some regions with higher rates like the U.S. at 21% and Japan. So assume between 15% and 18% for tax. Noncontrolling interest should be similar to prior periods. This quarter we had a specific upside related to a Japan-related income tax benefit, which is a nonrecurring item. Excluding that, noncontrolling interest should be similar to previous periods and is a small amount.

Cody, I wanted to clarify: when I said there's nothing that would prevent TSMC from buying our PICs, I included the scenario where we become the reference design for major integrators and that our PICs would be used in their integrations.

And Russell, can you expand on your relationship with NVIDIA at 1.6T? They've historically used TSMC but are also partnering with you going forward.

I wish I could provide more detail, but I don't have freedom to talk about specific customer programs beyond the press we've released. It was clear in the PR that NVIDIA referred to us as a development partner.

I have no further questions for today. I would now like to hand the conference over to Russell Ellwanger for any closing remarks.

Yes. Thank you very much. I appreciate your continued trust and support. I want to thank our teams around the world whose dedication has made the progress we reported possible and the progress we expect to have over the next year and years possible. Success begets success. The more success we have with our customers, the more excited they are, and that inspires our people and creates strong partnerships. We have extremely good relationships with multiple optical customers. When the customer is really happy with your performance, interactions become inspiring for next-generation work. I thank our teams and our customers. To our equity stakeholders, I truly appreciate your continued trust and support. We look forward to seeing you at a variety of events planned in the coming months: on May 18 we'll participate in the 27th Annual Oppenheimer Israeli Conference in Tel Aviv; on May 27 we'll attend the 54th Annual TD Cowen Technology, Media and Telecom Conference in New York; on May 28 we'll participate in the 23rd Annual Craig-Hallum Institutional Investor Conference in Minneapolis; and on June 9 we'll attend the 2026 Mizuho Global Tech Conference in New York. If you have availability to be at any of these, we'd love to meet with you. As always, our Investor Relations team is open to accepting calls and setting up video calls with anyone who would like further updates. We believe we have an extremely exciting and rich roadmap. Thank you very much.

This concludes today's conference call. Thank you for participating. You may now all disconnect. Have a nice day.