Tower Semiconductor - Earnings Call - Q2 2025

Transcript

Speaker 3

Good afternoon, ladies and gentlemen, and thank you for standing by. Welcome to today's Tower Semiconductor Second Quarter 2025 earnings conference call. At this time, all participants are in listen-only mode. There will be a presentation followed by the question and answer session, at which time, if you wish to ask a question, you will need to press star one one on your telephone keypad and wait for your name to be announced. I must advise you that this conference is being recorded today. I would now like to hand the conference over to our speaker today, Ms. Noit Levi-Karoubi, Senior Vice President of Investor Relations and Human Resources. Please go ahead, Madam.

Speaker 0

Thank you, and welcome to Tower Semiconductor's Second Quarter 2025 Financial Results Conference call. Before we begin, please note that certain statements made today may be forward-looking and 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 Semiconductor assumes no obligation to update forward-looking statements. Our Second Quarter 2025 results are prepared in accordance with US GAAP. Some data 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. Please note that we have a supporting slide deck that 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?

Speaker 2

Thank you, Noit. Thank you all for joining us today for our Second Quarter 2025 earnings call. We delivered strong results in the second quarter with revenues of $372 million and a resultant net profit of $46.6 million. We guide our third quarter revenues to be $395 million plus/minus 5% and additionally target a $40 million plus revenue increase for the fourth quarter over the third. Q3 guidance and Q4 expectations validate our onset 2025 target of sequential quarter-over-quarter growth throughout the year with acceleration in the second half. We announced at year's beginning a repurposing of multiple factories, predominantly towards higher capacity for RF infrastructure, namely silicon-germanium and silicon photonics. This is well underway, with Q3 and Q4 expected growth being the first fruits of the execution of this strategy.

Demand not only remains very strong but is consistently growing, as is our increase in both silicon-germanium and silicon photonics capacity and associated customer qualifications. We continue to invest in capacity and also R&D advanced capability CapEx throughout 2025, with further capacity and capability growth planned for 2026, aligned to our customers' forecasted demand, confident in maintaining our No. 1 market share position in this growing and significant optical transceiver market. Let's review our Second Quarter 2025 revenue breakdown along with some context that highlights key trends and momentum shaping our performance for the full year. Kindly refer to slide No. 4 for a detailed breakdown of our quarterly revenue figures. Most notable is growth in our RF infrastructure business attributed to data center and AI expansions served by our silicon photonics and silicon-germanium technologies, predominantly for optical fiber communications.

In the second quarter, RF infrastructure represented 25% of corporate revenues, over $90 million in revenues, up from 14% in the same period of 2024 and expected to significantly increase over the next periods. Specific to silicon-germanium, we began volume production shipments from San Antonio Fab 9 for a Tier 1 customer and as well volume wafer starts in Israel Fab 2 for another Tier 1 customer, providing substantial capacity increase in this growing market on top of the high capacity in our Newport Beach facility, which itself is realizing capacity increases this year. We have also made silicon-germanium design kits available in our 300-millimeter Japan factory Fab 7, for which an additional Tier 1 customer is presently in the design phase.

For silicon photonics, in addition to our existing volume on 400 and 800 gigabit per second, current wafer starts now include a good ramp on 1.6 terabit per second as the industry continues to move aggressively to higher speeds, which is expected to further help market penetration of silicon photonics over legacy solutions, expanding our opportunity due to the cost and strong performance benefits of silicon photonics. In the first half of 2025, we've moved five times more silicon photonics products from pre-production to production phase than in the same period in 2024, already exceeding last year's total. This growth demonstrates our platform maturity, strong customer adoption, and efficient operational scalability. Today, most silicon photonics products we manufacture serve the transmit function in an optical transceiver module.

This quarter, we successfully prototyped a new 300-millimeter silicon photonics technology that enables cost and performance advantage to the receive function in an optical transceiver module. This new technology is expected to see initial production in the fourth quarter of this year, expanding the market served by our silicon photonics technology beyond that which we serve today. By leveraging the majority of features in our mature silicon photonics platform and adding evolutionary customer-partnered improvements to this base process, we've been able to quickly ramp up and achieve high yields for 200 gigabit per second lanes, 1.6 terabit per second optical transceiver products. The transition of 400 gigabit per second lanes, 3.2 terabit per second, requires additional and fundamental device process improvements and new materials, which Tower is aggressively pursuing together with Tier 1 customers. The first of these platforms is anticipated to ramp as early as mid-2026.

Slide 5 shows cumulative wafer build to date for 400G, 800G, and 1.6T speeds. Presently, we are manufacturing similar amounts of each offering. Looking at prototypes, there's further acceleration of 1.6T, with 40% higher year-to-date prototypes at 1.6T above 400G and 800G combined, and this serving multiple large customers. We're seeing recovery in our RF mobile business, specifically in RF-SOI, which showed a Q2 to Q1 revenue increase of over 20% and is expected to show further increases, close to 30% Q3 over Q2, and targeting further increase in Q4. We've gained momentum with a new North America Tier 1 customer, now prototyping several products on RF-SOI in our 300-millimeter facilities in both Japan and Italy. Beyond RF-SOI, we're innovating with other RF switch technologies.

This quarter, we won the IMS Best Paper Award with PCEMI, a Maratha fully owned company, for our PCEM phase change material switch technology, achieving a 15 femtosecond RONCO figure of merit, being a 4X improvement versus state-of-the-art RF-SOI. These switches are being prototyped for both low and high-frequency millimeter wave applications. Importantly, this quarter received a Best Supplier Award from Wysov, a major Korean RF front-end module provider. Looking at power management, the computational complexity of AI processors is increasing, leading to a corresponding rise in power requirements. To meet this need, as shown on slide 6, we provide a variety of power management solutions with switch devices that have ultra-low resistance, advanced digital logic integration, and manufacturing options in our 300-millimeter lines in the U.S. and Japan. Lead customers are now designing to our high-efficiency power delivery solutions for this rapidly growing market.

We deliver best-in-class power transistor performance and continue to advance this offering, having released device optimization for higher switching frequencies this past quarter. For sensors and displays, we expect a revenue increase of about 20% in the second half of 2025 against the previous quarters and previous year's run rate, primarily due to increases in the machine vision market. We have also begun substantial new activities with several customers, including a leading automotive imager provider and an OLED-on-silicon supplier, the latter of which has prototypes already in test. These new activities are expected to fuel additional future growth. Looking at utilizations, in the second quarter, Fab 2 in Israel and Fab 9 in Texas both operated at about 60% utilization while repurposing tools to now load high levels of silicon-germanium and to begin the manufacture of silicon photonics. Fab 3 is fully utilized at our 85% utilization model.

Fab 5 was at 75%, with rising demand for high-voltage power management. Fab 7, 300-millimeter, was fully utilized, well exceeding the 85% model. In summary, our business continues to advance with significant progress across key platforms. We are well positioned for continued growth and success in our target markets. We are successfully executing to a clear strategy that is translating into tangible financial results with expanding customer engagements and measurable operational progress. We are committed to delivering sustainable and long-term value to our stakeholders. With that, I'd like to turn the call to our CFO, Oren Shirazi. Oren, please.

Speaker 1

Hello, everyone. Earlier today, we released our financial results for the second quarter of 2025. For the second quarter of 2025, we reported revenue of $372 million, representing $21 million or 6% year-over-year revenue increase compared to the same quarter in 2024, and a $14 million or 4% quarter-over-quarter revenue increase compared to the prior quarters. Gross profit and operating profit for the second quarter of 2025 were $80 million and $40 million, respectively, each higher than the prior quarter by $7 million. Net profit for the second quarter of 2025 was $47 million, also $7 million higher quarter-over-quarter, representing $0.42 basic and $0.41 diluted earnings per share. Financing and other income net in Q2 '25 was $14.4 million compared to $10.6 million in the prior quarter.

This $3.8 million increase was mainly attributable to gains recorded as a result of zero-cost collar transactions we executed to hedge future foreign currency risk. In this respect, I would now like to describe our currency hedging activity. In relation to the Japanese yen, since the majority of Tower Semiconductor's revenue is denominated in yen and the vast majority of Tower Semiconductor'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 exposures, we are executing zero-cost collar transactions to hedge the currency fluctuations. Hence, while the yen rate against the U.S. dollar may fluctuate, the impact on our margins is limited.

In relation to the Israeli shekel and euro currencies, while we have no revenues in such currencies, since a portion of our cost in Israel is denominated in the Israeli currency and a portion of our cost in Agrate, Italy is denominated in euro, we also hedge a large portion of such currencies' risk by engaging zero-cost collar transactions to mitigate exposure resulting from our NIS and euro-denominated costs. Under GAAP, the fair value of such transactions is recorded in the P&L, which drove the previously stated financing and other income net increase. Moving to our balance sheet and future CapEx and cash plan, our balance sheet remains very strong, as evidenced by the following indicators and financial ratios.

As of the end of June 2025, our assets totaled $3.2 billion, primarily comprised of $1.4 billion in fixed assets net, predominantly comprised of fab machinery, and $1.8 billion of current assets. Current assets ratio is very strong at about 7X, while shareholders' equity reached a record of $2.8 billion at the end of June 2025. Our strong financial position allows us to invest in strategic opportunities that support our corporate vision as follows. We have committed to pay up to $300 million to acquire equipment and CapEx in the 12-inch New Mexico fab, 20% of which has been already paid, while the remaining 80% are forecasted to be paid as we ramp up capacity and technology qualifications over the next two and a half years. As part of our STMicroelectronics partnership, $500 million in cash is allocated for the Italy 12-inch fab equipment.

We've already invested 85% of it, while the balance is expected to be paid by mid-2026. In addition, for our high-margin silicon photonics and 5G business, we announced plans to invest $350 million to expand our capacity in our 8-inch fabs in Israel and Texas, and in our 12-inch wafer fab in Japan. This CapEx includes a large portion of capability CapEx for advanced development and high-end RF technology-related projects. 40% of this amount has been paid to date, while the remaining 60% are expected to be paid by the end of 2026. All of these investments, including the 5G and silicon photonics CapEx, are fully reflected in our previously presented strategic and financial model.

Under this model, we target $2.7 billion in annual revenue at full loading of our existing fabs, including a grateful favor in New Mexico, $560 million per annum of operating profits, and $500 million per annum of net profits. That concludes my prepared remark. Now, I'd like to turn the call back to the operator so we can take your questions. Thank you.

Speaker 3

Thank you. Ladies and gentlemen, we'll now begin the question and answer session. As a reminder, if you wish to ask a question, please press star one, one on your telephone keypad and wait for your name to be announced. Listening by will compile the Q&A queue. This will take a few moments. If you wish to cancel your request, press star one, one again. Now we're going to take the first question. The question comes to the line of Cody Acree from Benchmark. Your line is open. Please ask your question.

Hi, Cody. Yeah, thanks, guys.

Speaker 2

Hey, guys. Hey, Russell. Hey, congrats on a great quarter and a great year, it looks like, shaping up so far. Maybe if I could just start with, could you give a little more detail on what segments you expect to contribute and maybe rank order them for your sequential growth through the second half of the year?

First and foremost is infrastructure as far as quarter-over-quarter deltas and absolute numbers. The silicon-germanium and the silicon photonics. We had stated at the year's onset that on silicon photonics, we had, what, circa $105 million revenue in 2024, and we expected a doubling of that in 2025. That's definitely within the second half targets that we're given, maybe, you know, higher than a doubling. That's a very important segment for us. Maybe, in some summary comments, I might give some more color on what the capacity increases that we're doing there, what that could result in in increases against what we're targeting in the fourth quarter of this year. That's the biggest. Our power management, we're expecting still strong contributions.

I mentioned that in the imaging, there was an over 20% increase in the run rate of 2024 in the first two quarters of 2025 in the second half, at least from what we target right now from our customer forecast, and that's mainly due to an increase in machine vision. Those are the areas of growth. Very important on a rebound is the weakness that we had cited previously in RF-SOI for front-end module for mobile. That has seen a nice increase from Q1 to Q2, and we're guiding an even stronger increase in the third quarter. The biggest, again, being infrastructure and RF-SOI for mobile, power management continuing strong, and CIS adding to it as well.

It sounds like a good strength across the board then, Russell. Congrats. Would you then say, would you consider yourself now fully booked through the end of the year? If so, do you have any available capacity to support any near-term turns business that could provide any yet upside to your estimates?

We had mentioned a 60% utilization in Fab 2 and in San Antonio Fab 9. We certainly have room there for immediate upsides. The biggest portion of both of those fabs is focused on utilizing the additional 5G and silicon photonics capacity that we've built there, which is, to an extent, in our hands, to an extent rests on completion of customer qualifications. That's within the numbers of the expectations that we have. We see those utilizations increasing. As we're doing that, there is room to handle immediate upsides. One of the really, I think, unique features of Tower Semiconductor is that we have a very strong worldwide manufacturing footprint where we're in Japan, we're in multiple sites in the U.S., we're in Israel, and we're in Italy. By definition, we cross-qualify all of our flows.

As there continues to be questions geopolitically and people get concerned a little bit here or there about tariffs and impacts of tariffs, we have the ability to move customer demands from one factory to another. One of the strong reasons of moving silicon photonics and silicon photonics, both and especially silicon photonics being very, very high-margin value-add products, is firstly to have added capacity, but secondly to be able to serve customers depending on their needs of where to ship products from. I think we've seen some immediate-term business, especially in the Tonami Fab 5 in Japan, in the power, and we're open to get more. Fundamentally, where we're focused at this point is meeting a very strong, consistently growing demand for the infrastructure, the silicon photonics, and the 5G. Hopefully, that answered your question, Cody.

It does, Russell. Thank you very much, and congratulations on the progress.

Thank you.

Speaker 3

Thank you. Now we'll go and take our next question. The question comes to the line of Richard Shannon from Craig Hallum Capital Group. Your line is open. Please ask your question.

Hey, Richard.

Speaker 2

Hi, Russell. How are you? Congratulations on some nice numbers again.

I'm doing good, it depends on your question.

I think I'm going to ask one on silicon photonics. You made an interesting comment about in the past supporting transmit functions and now supporting or prototyping, I think you said, for some stuff on the receive side. I'd love to understand a little bit more about that, but also maybe just kind of look at the bigger picture here about what you're expecting in terms of content, both the maximum amount as well as kind of expected growth in content over the next couple of years in silicon photonics, please.

For the specific receive function that we're doing now, we think, and it's for a specific application, I don't want to get into the details here because some of it really, it's not just what we're doing with the customer, but sometimes speaking too much can give other people hints as to what you're doing and why. It's a specific application of receive function that we've been able to very, very innovatively address with silicon photonics in an extremely high-performance, cost-effective way. We think that it would add about 20% to our serve market, plus/minus for this specific application in receive. As far as the demand that we're seeing, in talking about our growth, which we have many times, the fourth quarter, we would expect very high amounts of silicon photonics shipments.

We're probably seeing from our Q4 expected shipment level in silicon photonics by customer demand at the end of 2026, a doubling of demand, a doubling of what we would have to be supplying in capacity.

Okay. Fair enough. That's one way to look at it. I appreciate those comments. My follow-on question, Russell, is in the RF mobile space here. I guess the way I'd like to think about this is to what degree are we seeing this improvement both in the second quarter? I think you're talking about the third quarter at least or potentially second half here about this improvement here. To what degree is this cyclical like inventory replenishment and/or share gains here? To what degree does that play into your comment from last quarter about seeing some strong growth in RF mobile in both 2026 and 2027? Thank you.

I don't see the 20% growth that we had Q2 over Q1 and the 30% that we're targeting in Q3 over Q2. Cyclicality, it's hard to speak about cyclicality, but certainly, it's related to inventories having been consumed that had been a function of the very strong 2024. We also have multiple customers that themselves are growing their market share. That's one way that one can always outdo a market trend, by serving customers that they themselves are growing their market share against others that you might not be serving. The Q3, Q4 numbers that we're looking at, I think that they're just very basic share gains that we have and our customers have.

Some of what we would be expecting in the latter half of Q3 and Q4 deals with an existing customer that we've had for a long time that has just recently increased their forecasts and their POs. What has driven that, I honestly don't know at the moment. There are POs that have come in just recently.

Okay. Thank you, Russell.

Speaker 3

Thank you. Now we are going to take our next question. The question comes to the line of Matthew Hosseini from Susquehanna Financial Group. Your line is open. Please ask your question.

Thanks for taking my question. I have two. First one for Russell. I want a better understanding of the big picture, especially when it comes to data center infrastructure. We are migrating from copper to pluggable and at some point migrating to co-package optics. What I want to better understand, Russell, is how do you see your customers evolving? In that context, how are you actually planning for capacity in a sense that to avoid a double, in a sense to avoid too much capacity coming online because there is also a convergence among your customers? The transition in technology from copper to pluggable and eventually to co-package and how the mix of customers is evolving and how that impacts your capacity planning. I have a follow-up.

The first part of your question, I don't really fully understand. I think that plugables have been the mainstay for multiple years. I don't see it being in any competition with copper. Plugables have been the mainstay. They're maintaining the mainstay. The major thing for silicon photonics is the movement from an ENL solution that's been the mainstay solution to a silicon photonics solution. That's the difference there. I don't really follow the copper statement. Certainly, a lead customer that we have at 1.6T has decided to do everything with silicon photonics versus ENL because of two functions of the silicon photonics. One being cost, which, you know, the ability to not have to do separate indium phosphide modulator laser, that's a very, very complicated process and very expensive to have the modulator being made in the silicon photonics chip as well as integration of all the passes.

It has a strong cost benefit, but they also noted a very strong performance benefit at 1.6T, and that's because of integrating everything onto the PIC. I would, and I had said that in the script, we see this 1.6T where there's both a cost and apparently a performance benefit as being a big drive that will converge more into silicon photonics from ENL. Presently, what is the percentage of silicon photonics versus ENL? I couldn't tell you exactly what that is. I know that our demand for silicon photonics is strong and continues to grow and to grow very strongly. At the 1.6T, it seems to be even stronger, well, it doesn't just seem, that is. As we move to 3.2T, stated we are working with lead customers on producing the capabilities there, which requires to do a 400 gigabit per second modulator.

It requires different materials, and we're very advanced on that, pursuing two different material types. I see the plugables staying very strong for a good period of time. There are those in the industry that question if CPO will ever really come into the market. I'm not saying it won't. We have our own strategy for CPO, and I think we're pursuing it fairly well. CPO, if it comes in, should it come in, probably is not till 2029 or 2030. Everything we do with 5G and everything we do with silicon photonics will still be in demand. I don't see it being something that would discount that.

Great. Thank you for clarifying my question. Actually, as a follow-up to it, I want to understand how integrating the modulator into your process technology is perhaps helping you with incremental revenue per wafer. Is that how we should think about it?

For silicon photonics, the modulator is always part of the PIC. We're going from using our capabilities for the photodetector and the modulator, where we have strong capabilities in germanium and silicon, obviously, to being able to provide that as part of the PIC. That is one of the benefits of silicon photonics itself. That's one of the big drives of it. Otherwise, you're doing things in indium phosphide outside. As I stated previously, the process to make the indium phosphide modulator, in addition to the laser, is a very complicated process because it's different layers for both that are done together. It's many photo steps, passing steps, and growth. If that answers your question, I'm not sure.

Yes, absolutely. Actually, I want to highlight, I want to make sure I understand. The indium phosphide alternative that one or two of your competitors are highlighting has a cost disadvantage, and your solution is more cost-effective, right?

Yes, I believe so. Most of our customers still use a legacy solution, right? It's a question of that they also turn into silicon photonics for certain speeds, and it appears that the silicon photonics adoption is only growing at 1.6T, and I believe it will be stronger beyond.

Okay, great. I'll go back into the queue.

Thank you. Now we're going to take our last question for today. It comes to the line of Lisa Thompson from ZACKS Investment Research. Your line is open. Please ask your question.

Hi, thank you for taking this. I'm looking at the numbers, and it really seems like RF Mobile has come roaring back. Do you think between RF Mobile and RF infrastructure, they're both going to come in pretty much the same for the year, or can Mobile even beat that?

Let me take a look. No, infrastructure will be substantially bigger, but the mobile is a good size. Understand for infrastructure, it's two different products that we're serving that both serve it, right? It's the 5G for drivers and TIAs, and in some cases, still for CDRs, and it's the silicon photonics for the PIC. It's two different products that are both needed. For the RF-SOI, which is, and we're also doing some just plain RF CMOS for controllers, but the bulk of it is RF-SOI. No, the infrastructure is bigger. The mobile numbers, if we look at Q3, Q4, they're very high run rates, similar to past years.

No, that's great. Now, if you had to describe why it's come back so much, is it just specific customers, or is it the market?

I think a combination of both. As stated, we have customers that they themselves are going market share against their competitors that, in some cases, we do not serve. That's always a big thing when you're increasing your share of market. I stated it at the beginning of the year. We had seen, I think, not just us, but many people saw a pullback in mobile for the fact that 2024 had been a very big growth year in mobile. Nobody wants to short-ship their end customer for the fear of losing SKUs. They build up more inventory than is needed, not knowing when there would be a glut, so to speak. That's how you have inventory corrections. The beginning of the year certainly saw an inventory correction. From everything we're seeing now, it appears that most of that inventory has been eaten up.

Okay. Thank you. Oren, one question for you. I mean, on other income, you keep saying that the number is going to fluctuate, but yet it's steadily increased for the last, you know, year and a half. Should we expect it to continue to increase?

Speaker 1

No, I think we had a good gain here, like I mentioned in my prepared remarks, that caused the majority of the $3.8 million increase, which is good in financing and other income net. From the $10.6 million we presented last quarter to $14.4 million this quarter, I think the number we should expect in the future is the same like the baseline of the $10.6 million. This was a gain from movement in this quarter, but one cannot predict the future. I would use the same baseline, which includes all the numbers without any exceptional items or one-time $10.6 million.

Okay. Great. Thank you so much. That's all my questions.

Speaker 3

Thank you. Now we're going to take our next question. The question comes to the line of Richard Shannon from Craig Hallum Capital Group. Your line is open. Please ask your question.

Great. Thanks for letting me ask you a couple of follow-on questions here. Russell, one for you here. Back in the third quarter call of 2023, you set out this framework from a revenue and margin or profitability perspective. I guess I have two questions to this. How do you think you're tracking so far, almost two years into this, in terms of the margin profiles here, both at the gross and EBIT line here? Maybe I'll try to stretch it here and see if you'll respond in any way here. How do you think about a timeframe for hitting this revenue goal? Any thoughts you would give us? You're referring to the model of the $2.7 billion?

Speaker 2

That's correct.

On a margin perspective, I think we're probably outperforming from everything that we're doing. On a timeline to the $2.7 billion, you know, candidly, we were always looking at somewhere 2028, 2029 to be there. I think that's still what our targets are.

Okay. Great. That's helpful. My follow-on question is probably for Oren here. Oren, I'd love to get a sense of any way you'd quantify this as well, if possible, thinking about how we should see depreciation grow here as we're adding CapEx here. As kind of the corollary of the question here, thinking about free cash flow, obviously, you're going to be negative here with a strong CapEx investment here. How do we think about next year? Is this going to be positive, nicely positive? Just any way you can characterize that would be great. That's all for me.

Speaker 1

Okay. So depreciation, actually, you can see that in one of the tables that we attached to the press release, there is a line which is attached to the cash flow called depreciation amortization, which includes RSU amortization. The amounts are listed there as about $65 million, $70 million a quarter. We expect those levels to remain pretty much the same, maybe slight increase because, like you mentioned, the incremental CapEx, on the one hand, increasing depreciation. On the other hand, there is depreciation that goes out of the books because of investments that were in the past. Your point is valid that CapEx currently is higher than in the past. Depreciation should go up slightly, but most of that is fixed costs, and I don't believe it should go more than the $70 million, $75 million per quarter that we present.

As far as free cash flow, in the last few quarters, we see the same pictures that the cash flow operations, which is very good, is a similar number to the CapEx. This is because of the total of $1.15 billion in CapEx that we announced, the $500 million grateful, the $300 million 11X, and the $350 million SiPho 5G CapEx. This $1.15 billion supposed to be, I refer to that in my prepared remarks, supposed to be still paid in the coming two, two and a half years. More than $500 million of that was not paid yet. This will be part of the CapEx in the rest of 2025 and surely 2026, and some of that in 2027.

I would expect the CapEx total level per quarter to remain like now, maybe slightly go up, to be between $100 million to $120 million a quarter, which is like now. However, the cash flow operation is supposed to continue its positive trend and improve as we go up on the revenue, which Russell mentioned, the mid-range guidance, which is $395 million, and the target for Q4 even higher. Obviously, cash flow operations will be higher when CapEx is supposed to remain in those levels that already reflect the new plans of $1.15 billion.

Speaker 2

Okay. Great. Thanks for all that detail, Oren. That's all for me.

Speaker 3

Thank you. Now we're going to take our last question for today. This comes to the line of Matthew Hosseini from Susquehanna Financial Group. Your line is open. Please ask your question.

Yes. I have two follow-ups. One on OpEx. Should I assume that OpEx in 2025 would trend flat to up on a year-over-year basis? Is that a fair assumption? I have a follow-up.

Speaker 1

Yes. Yes. We consider that fixed cost and the OpEx should remain flat at the current run rate of about $40 million a quarter.

Okay. Going back to the previous question regarding cash flow, you actually have done a good job of funding the CapEx and actually maintaining a stable net cash per share. Is there any plans for the cash, or should I just assume that you would rather be conservative and just accumulate cash beyond funding the CapEx? Any additional thoughts you can share with us?

Yeah. Like you mentioned, the destination that we believe for our cash is for CapEx growth, which is why we approved in the last few years the $1.15 billion CapEx spend for the Intel fab, for the Agrate fab, and for the SiPho and 5G, total of $1.15 billion. We believe that's the best returns for the shareholders that we invest in CapEx and see the revenue growing like Russell described. It's amazing growth this year, quarter over quarter. That's the purpose that we plan for our cash.

Speaker 2

Okay, thank you.

Thank you.

Speaker 3

Thank you. The speakers are enough for the questions for today. I would now like to hand the conference over to Russell Ellwanger for any closing remarks.

Thank you. Really, to everyone on the call, thank you for your interest in the company. Thank you for the good questions that were asked. We did deliver a good second quarter. We guide a strong third quarter with a target of an even stronger best-ever revenue fourth quarter. Spoke to some length, and there were some questions about increases in our silicon-germanium and silicon photonics capacity. The third and fourth quarter are expected to begin to see the benefits of these activities. I wanted to kind of frame what can be expected as all of this capacity that's coming online is qualified and shipped.

The end-state capacity, which should be realized in the second half of 2026, will be a capacity that is 33% higher in silicon-germanium and 2.2 times larger in silicon photonics than the fourth quarter 2025 targeted shipments, which total revenue is targeted to exceed $435 million. Very importantly, these increases in capacity track well with our customers' forecasted demands. It's, I think, for us, a very strong story. We believe that the markets that we chose to work in were, are, and continue to be the right markets. Our models, I think, are strong, and the incremental revenue to margin net profit ratios are good and will continue to grow. I thank everybody for their interest in the company. It's been extremely exciting. There are multiple activities that we're involved in in the next quarters. I wanted to just mention some of them.

In the short term, we would hope to see you at some of these conferences. We're excited to host our 2025 Technical Goals Symposium in China this September and in the U.S. in November. TGS, this global symposium, is our flagship technology event, bringing together customers and ecosystem partners, serving as a critical platform for showcasing our specialty analog platforms from SiPho, 5G, RF CMOS to power management and imaging, while aligning on future roadmaps and enabling stronger co-development partnerships. We'll be participating in the following conferences and truly look forward to meeting and engaging with all of you at these events. On August 20, we'll attend the sixth annual Needham Virtual Semiconductor SEMICAP 101 Conference. August 26, we'll participate in the Jefferies SEMI IT Hardware Comms Technology Conference in Chicago. On August 27, we'll attend the 2025 Evercore ISI Semiconductor IT Hardware and Networking Conference in Chicago.

On September 3, we'll participate at the Benchmark 2025 Tech Media Telecom Conference in New York. On September 10, we will attend the Jefferies TechTrack 2025 Conference in Israel. Again, very much appreciate your interest in our company. Look forward to providing you with updates on our progress towards achieving our long-term goals in the coming quarters and in the very short term, updating our achievements in Q3 and Q4. Thank you very, very much.

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