高塔半導體 (TSEM) 2021 Q2 法說會逐字稿

Ladies and gentlemen, thank you for standing by. Welcome to the Tower Semiconductor Second Quarter 2021 Results Conference Call. (Operator Instructions) As a reminder, this conference is being recorded, August 2, 2021.

Joining us today are Mr. Russell Ellwanger, Tower's CEO; and Mr. Oren Shirazi, CFO.

I would now like to turn the conference over to Ms. Noit Levy, Senior Vice President of Investor Relations and Corporate Communications. Ms. Levy, please go ahead.

Thank you, and welcome to Tower Semiconductor Financial Results Conference Call for the Second Quarter of 2021.

Before we begin, I would like to remind you that some statements made during this call may be forward-looking and are subject to uncertainties and risk factors that could cause actual results to be different from those currently expected. These uncertainties and risk factors are fully disclosed in our Form 20-F, F-4, F-3 and 6-K filed with the Securities and Exchange Commission as well as filings with the Israeli Securities Authority. They are also available on our website. Tower assumes no obligation to update any such forward-looking statements.

Please note that the second quarter of 2021 results have been prepared in accordance with U.S. GAAP. The financial tables and data in today's earnings release and in the earnings call also include certain adjusted financial information that may be considered non-GAAP financial measures under Regulation G and related reporting requirement as established with the Securities and Exchange Commission. The financial tables include a full explanation of these measures and the reconciliation of these non-GAAP measures to the GAAP financial measures.

Now I'd like to turn the call to our CEO, Mr. Russell Ellwanger. Russell, please go ahead.

Thank you, Noit, a pleasure. We're quite excited to share with you our second quarter results and business activities.

Our revenue for the second quarter of the year was $362 million, a record for Tower, representing 17% year-over-year total and 26% organic growth. In the order of revenue dollars, the year-over-year organic growth was mainly driven by RFSOI at over 40% year-over-year organic increase; power IC at 35% year-over-year organic increase; image sensors, 30% year-over-year organic increase; and power discrete at 23%. To note, all business segments demonstrated growth in the second quarter.

We guide the third quarter of the year to increase to a mid-range of $385 million, representing year-over-year 24% total growth and a 38% organic growth and breaking a $1.5 billion annualized revenue run rate. This should be driven by a further and large year-over-year increase in RFSOI and image sensors, with all business segments expected to demonstrate growth. Our customer demand and served markets are strong, hence, we expect continued top and bottom line growth in the fourth quarter as tools become qualified for production, enabling further increases in high-value flows.

To support continued growth based on our customer demand, we continue to execute the previously announced capacity expansion plans as well are adding 200-millimeter new capacity, which will be addressed with more details by Oren. Towards specific large 300-millimeter growth, we signed an agreement with ST Microelectronics. In this partnership, we will join forces for an accelerated fab ramp-up, a key factor to speedily reach a high utilization level and, therefore, a competitive wafer cost.

The Agrate 300-millimeter manufacturing facility, which Tower will refer to as Fab 10, is currently being facilitized. Tower will install its own equipment in 1/3 of the total space, which should triple our current 300-millimeter foundry capacity. The fab is expected to be ready for equipment installation later this year and start prototype production in the second half of 2022. Our strong execution in advanced 65-nanometer, 300-millimeter-based analog RF, power platforms, displays and other technologies, will be significantly enhanced by this activity in Agrate.

The United States Senate passed the Innovation and Competition Act of 2021 and appropriated $52 billion towards U.S. semiconductor growth. The bill is now before the House of Representatives. Providing its passage, our target is to receive funding towards further growth initiatives in the United States, potentially at our beautiful 130-acre campus in our San Antonio, Texas facility or possibly elsewhere.

Towards this end, during the past quarter, we've been strongly involved in activities related to this act. I have attended 3 industry leader panels on the subjects of growing and/or securing the U.S. microelectronics supply chain where I have focused as well on the need for advanced analog semiconductor integrated circuit manufacturing. One such panel was a CEO roundtable hosted by Senator Cornyn and Congressman McCaul, both sponsors of the CHIPS Act, and their respective legislative bodies of visionary activity. And I attended a special session at the SelectUSA Conference hosted by U.S. Secretary of Commerce Raimondo. I spoke of 4 imperatives that we believe are essential for foreign entities to qualify to receive incentives from the United States for United States onshore semiconductor manufacturing growth.

Firstly, there needs to be U.S. onshore R&D, especially in a foundry model, enabling innovative entrepreneurial ideas to successfully move through concept and feasibility and transition into value-add business. Secondly, all companies need to show a history of strong IP security, both external and internal, to the company. All recipients should have a strong history of supporting U.S. government initiatives. And lastly, there should be a legislative focus on differentiated analog technology, which addresses 80% to 85% of all system electronics by volume. We continue to work with the Congress of the United States, the Department of Commerce, other U.S. government agencies as well as other partners.

Moving to our specific businesses. During the second quarter, our RF mobile business was 25% of our revenues and is expected to show strong growth throughout the year. Estimates remain that 5G handsets will double year-over-year 2021 over 2020 to about 550 million handsets out of a total of about 1.3 billion. This shift, combined with the high content increase of 5G at our strong position in this market, fuels our continued growth. Importantly, in addition to revenue growth, the shift to more advanced and higher-value 5G technology is helping increase the average selling prices in this segment. This trend is expected to continue for at least the next several quarters.

Demand is strong in both 200-millimeter and 300-millimeter and the partnership announced with ST will help meet the requirements of our strong 300-millimeter design win pipeline. The RF infrastructure business, serving telecom and datacom end markets with our industry-leading silicon germanium technology, maintains a high run rate due to data center strength and was about 12% of our corporate revenues. For datacom, where forecast remained quite strong, we built both high-speed optical transceivers and high-speed hard disk drive preamps for storage.

Our power IC business was about 15% of our total revenues with strength in automotive, industrial and consumer segments. The business is benefiting from a strong market cycle as well as our significant presence in automotive battery management, which is outgrowing the market and provides for long-term growth with the worldwide trend of vehicle electrification.

Our power discrete business has more than recovered, representing about 16% of our revenues. Like power ICs, growth is broad-based led by automotive applications. We expect this business to level off over the next few quarters while we focus our CapEx expansion on other higher-margin segments.

Our imaging business represented 15% of our revenues, with main growing markets being the medical, dental x-ray and industrial sensors. In addition, we continue to grow in the cinematography and broadcasting market segment, being among our highest-margin imaging applications.

For the medical market, we see a recovery of the dental segment to levels that are higher than the pre-COVID-19 levels with a present high-growth trajectory. The main long-term growth drivers stem predominantly from the transition from traditional amorphous silicon-based flat panel technology to CMOS. Our customers, who are either x-ray detector suppliers or X-ray equipment suppliers who design with us their own sensors, are gaining more market share. And we, as sole supplier in most cases, are growing our share accordingly.

In the industrial market, we see a steep growth in all of the segments, machine vision for factory automation, for traffic control as well as automatic data collection. On the display side, we continue substantial partnership developments for backplane silicon for micro OLED, mainly for the VR display, which is a fast-growing market; and on silicon wafer-based micro LED technology for large displays, TV, laptops, tablets and smartphones.

Referring to utilization, starting this quarter, we will refer to the number of photo layers processed during the quarter per wafer size. Given the fact that we continue to increase the capacity of our manufacturing facilities with changing mix of flows, pure utilization numbers do not represent the company's operational performance on a time comparative basis.

For the second quarter foundry layers, all numbers given in 8-inch equivalents, for 150-millimeter, 451,000 layers were processed as compared to 353,000 in Q2 2020, up 28% year-over-year. 411,000 layers were processed in the previous quarter, Q1 2021. For 200-millimeter, 5,921,000 layers were processed as compared to 5,115,000 in Q2 2020, up 16% year-over-year. 5,772,000 layers were processed in Q1 2021. For 300-millimeter, 1,404,000 layers were processed as compared to 903,000 in Q2 2020, up 55% year-over-year. 1,375,000 layers were processed in the previous quarter.

Moving now to corporate sustainability. We are about to issue our first formal environmental, social and governance, or ESG, report. However, the core of ESG has long been embedded in the DNA of the company. For Tower, ESG is much more than papers listing activities and targets, it is our focus in being a good, even excellent, company for our community and our world. It is how we serve our employees, customers, partners and stakeholders. Various elements of our corporate responsibility, sustainability and ESG efforts will be described in this report and are well aligned into our value vectors with the mindset of excellence in each of the described areas. We are continuously evaluating our activities in order to improve and ensure that our commitment and actions toward a company that betters society and betters the lives around us are achieved, valued and sustainable.

With that, I'd like to turn the call to our CFO, Mr. Oren Shirazi. Oren?

Hi, everyone. We released today our second quarter 2021 results, achieving record revenues of $362 million, reflecting a remarkable 17% year-over-year revenue increase and resulting in significant increases in gross profit, operating profit and net profit.

We are providing a revenue guidance of $385 million for the third quarter of 2021, representing an additional record revenue quarter. As discussed several times over the last few quarters, we continue to see significant customer demand and demand forecast. Hence, we announced in February 2021 a $150 million capacity expansion plan to increase our capacity in our 8-inch and 12-inch fabs. We now announce an additional capacity investment of $100 million to result in $250 million total CapEx purchases, which are expected to be paid over the coming 5 quarters, for equipment that begin to provide incremental capacity during the third quarter of 2021 and continuing into 2022.

As mentioned by Russell, for enhanced 12-inch capacity on top of our existing wafer fab, ST Microelectronics and us entered into a partnership to accelerate the ramp-up of the Agrate 12-inch factory. We expect minimal CapEx payments for this project in 2021 and will give details on future schedule and cost during the next quarter.

I will now move to our second quarter P&L highlights and then discuss our balance sheet and cash flow financial statements. Revenue for the second quarter of 2021 was $362 million, $52 million higher year-over-year, reflecting a 17% year-over-year revenue increase. Looking at our organic revenue, which is defined as total revenue excluding revenue from Nuvoton Japan, previously Panasonic Semiconductor, and revenue from Maxim in our San Antonio fab, revenue in the second quarter of 2021 reflects a 26% year-over-year revenue increase.

Gross and operating profit for the second quarter of 2021 were $74 million and $34 million, respectively. This gross profit is $16 million higher or 28% higher year-over-year. And this operating profit is $12 million higher or 54% higher year-over-year. Net profit for the second quarter of 2021 was $31 million or $0.29 basic earnings per share and $0.28 diluted earnings per share. This net profit is $12 million higher or 62% higher year-over-year.

Looking at the balance sheet, we demonstrated, again, a strong and stable financial position. A few points to note, our shareholders' equity reached a record of $1.52 billion as of June 30, 2021. Deferred revenue and customer advances balances under current liabilities and long-term liabilities in the balance sheet have increased by $9.6 million and $6.6 million, respectively, as compared to December 31, 2020, reflecting enhanced receipts from customers that asked to gain more capacity reservations to them to address their excess demand, offset by some scheduled repayment. Current asset ratio, defined as current assets divided by short-term liabilities, is strong at a value of 3.7x.

In regards to our cash and cash equivalents in the second quarter of 2021, cash flow generated from operations was $93 million. Investments in fixed assets, net mainly for manufacturing equipment, were $56 million. We repaid $20 million of our debt during the second quarter of 2021 and invested $17 million in short-term bank deposit and marketable securities. Our cap table consists of 108.2 million outstanding ordinary shares and an additional 2.3 million ESOP-related shares, resulting in fully diluted share count of 110.5 million.

Looking forward to the second half of the year, following the $385 million record revenue guidance for the third quarter of 2021, during this quarter, we expect to see increased margins, a step towards the fourth quarter incremental revenue impact over the margin, which we forecast should exceed the 50% incremental model.

And now I would like to turn the call back to the operator.

(Operator Instructions) The first question is from Mark Lipacis of Jefferies.

Next question is from Rajvindra Gill of Needham & Company.

Congratulations on the great momentum across all of the major business lines. That's great to hear. So Oren, just a question on the gross margins as we progress throughout the year. You had mentioned in your prepared remarks that in Q4 that you would expect the model to exceed the 50% gross profit fall-through in Q4 as some of those incremental costs related to capacity have been absorbed. Wondering how we think about that upside to the 50% gross profit fall-through without giving any numbers, but just are we seeing higher-margin process flows in Q4 that would give you the confidence to say that we're going to be above the 50% GP profit fall-through?

I'll give an answer there. We have seen very, very strong increases in demand. And for a variety of the increases in demand, in order to be able to cover it, we have worked with customers for increases in ASPs that provide higher margins in order to have a quicker return on the investment itself. So a good portion of what's going on in Q4 is a result of Q3 starts that are now under POs that have a higher selling price than previous quarters, and it's driven off of customers that are working towards increasing demand versus previous run rates for the most part.

And yes, as mentioned during the script itself, as we move more and more into some of the, for example, 5G-driven activities, as we move into higher gigabit per second data center, as we move towards more large stitched die/wafers, all of that are higher-margin flows to begin with. So it's a combination of a richer mix and where we talked about the increase of another $100 million in 200-millimeter investment, that really is, for the main part, to increase capacity for a richer mix, for higher-margin mixes. So Q4 is a combination of both. It's as higher-margin mixes are growing and in instances where we are and have increased capacity that those are coming in at this point or will be coming off of purchase orders that had an increase in ASP.

Okay. Got it. That's super helpful in terms of the drivers for the margins. With respect to the ST Micro partnership that you announced a few weeks ago, a very, very interesting announcement. Wondering kind of the rationale for both parties, particularly ST Micro who essentially began construction of this facility in 2018 in Italy and is kind of bearing the bulk of the construction but is allowing you guys to leverage your expertise in 300-millimeter and 200-millimeter to really kind of maximize the foundry, I'm curious kind of what was ST Micro's thinking in terms of partnering with you and the value that you'll bring to the table. And maybe if you could just remind us in terms of how much expected capacity you're kind of targeting in 2024 and beyond so we get kind of a sense of the potential scale of this partnership.

So the fundamental drive was really stated by ST in the press release itself, and that was to drive a faster ramp to high utilizations and a faster ramp of capacity, which obviously then reduces the fixed cost per substrate. So the deal is a very strong, we believe and they believe, win-win for both companies. The specifics of the financial model has not been released. And I'm not sure that all the details will ever go out into the public. That's part of the confidentiality that ST and ourselves have. But the model itself is a very good model for both companies for many, many reasons. And I believe there's -- the major driver, again, from their side was what was stated in the PR, and that is a faster ramp to high utilizations, which is the key towards good margins in a manufacturing facility.

The next question is from Mark Lipacis of Jefferies.

Can you hear me?

Yes. Now we can.

You can? Okay. Great. All right. I don't know what happened there. Russell, on Agrate, I think you mentioned second half of '22 prototype. I just want to make sure I had that right. When do you think you'd start shipping production revenues?

Our target would be to start shipping production revenues in the first half of '23.

Got you. And when you talked -- you talked about, I think, several capacity expansion plans. I don't think I heard silicon germanium mentioned there. Are you -- what is the -- is that included in the capacity additions? Can you talk about what you're doing with silicon germanium from a capacity expansion standpoint?

It is definitely included in it. The infrastructure activities that we're doing, we had mentioned, are very strong. We have a growing demand consistently in data center. And yes, so the silicon germanium and silicon photonics are both parts of our expansion plan. That's where we said about the -- to invest in 200-millimeter differentiated platforms. So 5G platforms are certainly one of our differentiated platforms.

Got you. That makes sense. And then on the CHIPS Act and your view to kind of target San Antonio there, do you have a sense of when you would expect to get visibility on this? And to what extent would you expect this to be direct funding or versus tax credits?

Expectation doesn't necessarily tie into what will happen. Our expectation is direct upfront funding. And so that's where we'd like to be. I think city, county governments do a lot and have offered a lot in areas of tax abatement. And that, I think, is a very good thing for ongoing margins and reduction of ongoing running costs. But the big issue of any greenfield activity is the upfront investment and the amount of years that it takes to bring that upfront investment into a net profit positive business and upfront grants are what really enables that.

So for us, the important thing is to -- whatever works out or whatever might be worked out with us is then to make sure that it really pencils well for the company and for our shareholders. And that's what our target is, is to have a deal. We'd love to increase capacity in the United States. Again, I think the San Antonio facility is a very beautiful campus to target it within. But certainly, it has to be something that pencils. And so that would be part of the discussions that we would have at a point that whatever committee would be ultimately used to determine and grant awards would be granting these awards.

The next question is from Richard Shannon of Craig-Hallum.

I want to follow up on the gross margin question here. You talked about investment in differentiated flows. And at the same time, you talked about some of your power discretes maybe leveling off here. I want to jump to that one in a second. But how do we think about the longer-term gross margin model of the company as you make all these investments, obviously, expecting to get a run rate utilization on those? And it'd be interesting to hear that answer in context to your past history in gross margins where you've, 3, 4 years ago, got into the mid-20s level here and you're at the low 20s today. I'm curious to the degree to what you think you're going to approach those prior high levels.

Yes. So Richard, I believe I related to that a little bit in the prepared remarks. So for Q3, we will see probably gross margin increase and all the margins will increase, however, not to the full 50% incremental model. And from Q4, we expect it to exceed the 50% incremental model also, as Russell explained previously, because of some very good selling price increase trend and the improved mix. For the next years, if you ask going forward, I think you should go from the baseline of the Q4 betterment on the margins and then, from that point in time, to assume incremental model slightly above the 50%. If you want to be conservative or not, slightly above but better than the 50%, if the trend in the market, which you know better than me, will continue or [SME] continue like now because now it's a very good trend for us.

Okay. And Oren, just a follow up on this since I don't have details on what kind of total dollar capacity you're expecting here, I'm not sure how far to take that incremental 50% gross margin and get to -- hope for a level down the road here. Again, I'll ask if it's -- if we can get to near prior levels here as you fill up this new capacity. Is that possible? Or can we approach that? Just any comments relative to your history would be great.

What is your question, the maximum revenue capacity? What was your question?

No. Where we can go with gross margins over time given all these investments here, again, relative to your history here where you've gotten up to the 25%, even close to the 26% level a few years ago and now we're at the 20% level. And obviously, we've got some investments that will increase depreciation here. But over time, as you fill that up, can we get back up close or to that level that you see?

Yes, of course, it depends what outlook in the future you're looking into. But for the next year, yes, definitely, we can achieve the 25% or even higher -- slightly higher.

Okay. Perfect. That's helpful. Kind of a multipart question here on the power market here. First of all, you said -- I think, Russell, you said discrete, you're expecting to level off here over the next few quarters. I'd love to understand why there. And if you can make some broader comments also on your power IC space where I think you've been gaining some share. It would be great to know what voltage levels you're seeing the benefit and to what degree that's happening in EVs, that would be great to know.

Yes. So our power discrete market has been a little bit different than others in that for power discretes, we don't offer a foundry offering flow for the power discretes, we have customer flows that we bring into the company. And we do, in many instances, co-developments on additional flows or incremental flows or capabilities. The power discrete business itself is not the highest-margin business in the company, but it's a very stable business for us by virtue of the relationships we have with those customers.

The fact that it's very dependent upon each of these customers, we've reached levels of revenue that we think is very adequate in the company, and the company is not presently taking the capital expenditure to put into those flows as there's others such as -- was the follow-up question previously, such as our silicon germanium, such as some imaging capabilities, certainly SIFO capabilities, that both have a higher margin impact for us. And secondly, we own the flow, so we have much more freedom in the market as to who we work with. So that's the issue there. It's not that we don't have customers that would not wish or do not wish to grow with us in that area. It's, at this point, not our preferred area that we wish to grow our capacity in the company.

The next question is from Patrick Ho of Stifel.

Congrats on the nice quarter. Russell, maybe first off, in terms of -- I think you guys were very prescient in terms of the capacity expansion plans to meet the market demand environment. I know you have a lot of customers today. And obviously, some of the capacity expansion is dedicated to getting more share with the existing customers. But can you qualitatively maybe discuss how potential new customers can be brought onboard as you expand this capacity given the current demand environment?

Certainly. We are actually, in some instances, taking new customers onboard. But those customers are coming in under different types of arrangements where, potentially, they'd be paying for a certain capacity itself as it's all pure incremental capacity. When a customer would give money towards capacity, the model is one of 2. One model is that whatever the prepayment is, it would be amortized over time against wafers typically nominally with end time date so that the capacity is more or less are committed or money is forfeited. The other is that the money is just given with no amortization, which is more or less a guarantee for the company that there's a return on that incremental capacity and has one impact, and that is it really does increase the specific margin dollars for that customer on top of whatever the selling price is on the wafer. So those are our 2 models.

For the most part, at this point, we're working to support our present customers that all have increasing demands. And we think that that's very important that those people, who really have made commitments to us and are relying on us, that we're doing all that we can to enable their growth within the market. And I think one thing that is very, very indicative that we have a good truly taxonomy of customers is the percentages of organic growth that we're seeing that, I believe, really are much higher than the market growth itself. And hence, it shows that the customers that we're working with are good growth customers that themselves are increasing market share and, hence, we're increasing market share in the respective end segments. Hopefully, that answers your question.

No, it does. And maybe as a quick follow-up, I know 200-millimeter tools are different from 300-millimeter tools, but we're hearing more from the equipment industry that they're facing supply constraints. And again, I know 200-millimeter, there are probably a little more -- or there are probably more options out there to procure those types of tools. But have you seen any constraints on your end trying to get deliveries of some tools? Or is everything still kind of on track on your end?

For the most part, all the POs that have already gone out for equipment is on-time on deliveries. What we're seeing now is just that the lead time on deliveries is extending for new POs. And that's particularly the case for 300-millimeter, for 200-millimeter as well. But we've not seen a gating of any of our suppliers refusing to take a purchase order.

The next question is from Lisa Thompson of Zacks Investment Research.

So you spent about $56 million...

Lisa, we're having a very hard time hearing you.

Yes.

Could you hear me any better now?

Much better.

Yes. Much better.

All right. So this quarter, you spent about $56 million CapEx, and you said it was going to be like $45 million to $49 million. So do you have any different thinking on Q3, Q4 and Q1 from what you had previously said?

We have a little bit different thinking because of the -- what we said in the press release that we have announced an additional $100 million. So actually, this is why I said in my script that I expect now -- we announced in January -- in February $150 million, and we announced today additional $100 million, so total is $250 million, as I said in my prepared remarks, will be paid in the coming 5 quarters. So if you assume it linearly, it's additional $50 million to the baseline. Since we just announced the new $100 million recently, you may assume, obviously, that it will be below $50 million extra from Q3.

So I would assume lower than $50 million for Q3, like something like $40 million. And then $50 million a quarter and maybe some time $60 million to capture for the full $250 million. And this is on top of the baseline that we had without those CapEx plan, which is, like you mentioned, $45 million. So if you want to know Q3, Q3 will be about $45 million base plus additional $40 million from the -- towards the $250 million and pretty similar for the upcoming quarters. So between $85 million to $95 million total third quarter.

The next question is from David Duley of Steelhead Securities.

Just out of curiosity, as far as the Panasonic business goes, is that capacity fungible to other customers? Just kind of curious if it's unique to them. Or if, let's say, theoretically, Panasonic didn't exist, could that capacity be transferred to other customers easily?

Yes. So I understand the question. I just want to clarify so I don't offend anybody. Nuvoton bought the Panasonic Semiconductor, right? So it's now -- the group is called Nuvoton Japan technology. But -- so again, I just don't want to offend anybody there. I think, to a good extent, the capacity is fungible. To the extent that there's some specific flows that are used that are nonfungible, that's the case as well. So there are some flows that really is pretty much dedicated to end markets that are served with certain products, had previously been Panasonic Semiconductor, now Nuvoton Japan, serving Nuvoton as a whole. But a good amount of that capacity is fungible.

Okay. And then as far as the -- you mentioned the utilization rates, and you just gave us layer counts that were processed at different wafer sizes. Did you give -- I guess there's still a utilization rate number available, right? Because you gave us the total layer count at 100- and 200- and 300-millimeter, we could calculate the utilization rates. So could you help us understand, you gave us the number of process, what's the total number available so we can back into the utilization rate?

No, that's exactly why I'm using the process layers. The exact amount of photo layers is not necessarily a relevant number as there's other constraints and bottlenecks that are not necessarily photo-related. And we're adding photo as time goes on. So that's not an area that we'll be presenting at all anymore. We will just be presenting photo layers used. And on a comparative basis, it's a good operational metric.

And so then how do we tie that into utilization rates and efficiency of using your equipment since that's basically one of the most important metrics of a foundry business is the utilization of your equipment. So is there another metric that we can look at? Because when you just give us the foundry -- the number of layers processed, it's not necessarily utilization rate-driven. So is there any other metrics you'll be giving us to understand about how efficient you're using your equipment?

You can certainly assume that presently, we are fully utilized at every layer that we can ship or shipping before these new tools are arriving.

Okay. And then, Oren, what would we expect for operating expenses in the Q3 and Q4? I know you've been adding people and capacity at various locations. What should we kind of expect on a dollar basis for operating expenses in Q3 and Q4?

Yes. Operating expenses should remain flat in the future quarters because, you're right, that we are adding people but they are in COGS. It's mainly technicians and the operators, engineers to support the fab production. So that's in the cost. That's within the 50% incremental model. OpEx, R&D and SG&A, we are keeping them flat.

So we should see some nice leverage in the back half of the calendar year?

What, nice leverage?

Some strong earnings leverage from operating expenses being flat in the back half?

Yes, in percentages, yes.

Yes. So just to follow up on your question, in the prepared script, I mentioned the Q3 guidance and then I said that we expect top and bottom line continued growth in the fourth quarter as more tools become qualified for production, enabling further increases in high-value flows. So I think somewhat intrinsic to the statement, if we're depending on new tools to be qualified, what we're running right now is running at full utilization.

The next question is from Richard Shannon of Craig-Hallum.

Just a couple of quick follow-ups from me. Just to be clear, I'm pretty sure the answer is no, but I just want to confirm, Russell. The new $100 million CapEx addition here. That doesn't include any equipment that would be installed with the -- in the ST Micro facility, is that correct?

You are correct. It does not include.

Okay. That's what I thought. A quick question on another topic, your silicon germanium. Are you seeing much, if any, benefit from 400-gig devices or, I guess, devices that go into 400-gig datacom modules yet? Or is that still a pretty small piece of business?

It's a growing piece. I wouldn't say -- it's certainly not the biggest portion, but it is a growing piece. And yes, we've press released a revenue customer, Inphi, for silicon photonics. You know that Inphi does 400-gig, right?

Yes. Okay. Are you seeing 100-gig -- components for 100-gig, are those still growing at a decent rate? Or is that starting to plateau?

So in the prepared remarks, I had mentioned that what's really up and moving right now is the datacom, not the telecom. The telecom is where we do predominantly the 100-gig, right? So telecom is not -- the 100-gig is not increasing usually in volume at this point, the datacom is. And those are typically 25-gig chips.

There are no further questions at this time. Mr. Ellwanger, would you like to make your concluding statement?

Certainly. Thank you very much. Thank you for your interest and for the good questions. For your reference, posted on our website, on the quarterly release page on the Investor Relations section, is our second quarter 2021 financial results slide deck. Please access it as per desire and interest.

To summarize the call and where we're at, we really were excited -- are excited with second quarter record revenue results and particularly the activities that have led to a third quarter guidance of substantial continued growth and breaking a $1.5 billion annualized run rate. We believe that the $385 million Q3 guidance, which represents 38% year-over-year organic growth, is really extremely strong evidence that we're serving the right customers, in the right markets. That type of a growth, I think, is quite substantial. We are really executing well on our expansion plans. Most everything is in place and our target. And hence, as stated in the script, we are looking forward and expect fourth quarter growth in both top and bottom lines. And we're very excited with the partnership with ST at the Agrate factory and, over the next years, expect and believe that we'll see very important and significant movement in 300-millimeter growth and 300-millimeter capabilities as we qualify our tools and install more tools in that facility.

Other activities happening in the company are still exciting. And look forward to meeting with you again for the third quarter release and possibly other interactions in the interim. So thank you very much for your interest and best wishes for health and happiness. Thank you.

Thank you. This concludes the Tower Semiconductor Second Quarter 2021 Results Conference Call. Thank you for your participation. You may go ahead and disconnect.