高塔半導體 (TSEM) 2020 Q4 法說會逐字稿

Ladies and gentlemen, thank you for standing by. Welcome to the Tower Semiconductor Fourth Quarter and Full Year 2020 Results Conference Call. (Operator Instructions) As a reminder, this conference is being recorded February 17, 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 fourth quarter and full year of 2020. 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 forms 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 fourth quarter and full year of 2020 financial results have been prepared in accordance with U.S. GAAP. The financial tables and data in today's earnings release and in this earnings call also include certain adjusted financial information that may be considered non-GAAP financial measures under Regulation G and related reporting requirement is established with the Securities and Exchange Commission. The financial tables include the 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. I welcome everybody to our 2020 fourth quarter and fiscal year business and financial results. Thank you for joining our call. We finished 2020 with revenues of $1.266 billion, representing year-over-year 3% growth and 5% organic growth. We entered 2021 having completed a very strong fourth quarter with revenues of $345 million, having exceeded our mid-range guidance, which represented an 11% fourth quarter year-over-year total growth and 20% organic growth with resulting EBITDA of $95.8 million, a net profit of $31 million. We continue to maintain a healthy balance sheet that fully supports and enable us to add value while capitalizing business opportunities to facilitate future growth. We are guiding the first quarter to a mid-range of $345 million counter to the typical Q1 seasonality, which will represent a year-to-year 15% total growth with an organic growth of 20%. In addition, from this $345 million revenue base, we expect sequential quarterly revenue growth throughout 2021.

Looking at the revenue breakdown, we'll first discuss the end markets and give numbers to the best of our ability and knowledge regarding the end markets our ICs are serving.

Infrastructure revenue, which predominantly is our RF optical with a certain amount of advanced discretes was at about $230 million. Wireless was approximately $425 million. Automotive was $180 million, which we serve with power ICs, power discretes, imaging and RF radar. Consumer, including computing, power management for home appliances and general accessories and home-use security cameras was approximately $220 million. Industrial was approximately $100 million. Image sensors for high-end photography and medical applications was at about $50 million. Aerospace and defense was also at about $50 million. And there is an additional of about $10 million of devices that we sold that will be divided among the end markets that I've spoken to, but for which we don't have the granularity to break down further.

Analyzing the revenue by technology platform for our 3 corporate focus being seamless connectivity, green everything, an interactive smart system, the breakdown is as per follows: seamless connectivity, which for us is RF infrastructure and RF signal for mobile platforms, total to approximately 36% of our corporate revenues, or about $450 million; mobile was $280 million, a 22% 2020 over 2019 growth; infrastructure, which is silicon germanium-based, including the initial silicon photonic ramp, was approximately $170 million, a 15% year-over-year growth; green everything, our contribution being energy-efficient power management ICs and power discretes was approximately $405 million or 32% of our corporate revenues; power management ICs were $195 million, flat year-over-year, however, 25% up organically; power discretes totaled to $210 million, a decrease of 12% year-over-year; interactive smart systems, which relates to our image sensors and nonimaging sensor offerings, represented about 18% of our corporate revenues at about $230 million, having realized a 7% growth in 2020 despite significant contraction in medical, predominantly stitched large die dental sensors. The rest of our business of about 14% of our corporate revenue served various mixed signal CMOS technologies, mainly computing and specialty memory applications.

Looking at our activities in our different business units. Within our analog business unit, our silicon germanium infrastructure business, which provides technology for advanced optical transceivers, where we enjoy greater than 60% market share, experienced double-digit growth in 2020 versus 2019. Customer forecasts continue to show this elevated level of demand to be sustained through the remainder of 2021. In the next several years, there's an expected industry growth of data traffic of about a 15% CAGR. Our opportunity is to mirror this growth and benefit from both the continued rollout of 5G infrastructure, which drives demand for our 25 gigabit per second transceivers in telecom networks, and by data center build-out, which drives demand for our 100 through 800 gigabit per second transceivers.

At 400 and 800 gigabit per second, we also anticipate increased adoption of our silicon photonics platform, further increasing our footprint in the optical market and providing new opportunities for growth. As previously mentioned, we are well positioned in this market having already announced a partnership with Inphi for which we began volume production, and we additionally have over 30 customers engaged with us at various stages of qualification and development.

Last month, we announced participation in a DARPA program, developing a SiPho platform with integrated lasers, further differentiating our capabilities in this emerging market. Our mobile business, which provides components for RF front ends and handsets grew, as mentioned, over 20% year-over-year due to a combination of increased market share, overall market recovery and the beginning of a transition to 5G handsets. Growth was broad-based and included ramps of our newest 200-millimeter and 300-millimeter technologies as well as very strong demand for existing offerings. Forecast for mobile are strong for 2021, and we expect that 5G handsets, which, as previously mentioned, require 30% to 50% more RF content will increasingly replace older models over the next few years, creating a sustained opportunity for growth in this market.

Our power IC organic business grew 25% in 2020 over 2019 through gains in market share, both at 200 millimeter and 300 millimeter across a wide range of voltages and applications. We see increasing demand for power management ICs in multiple applications, including hybrid and electric vehicles as well as consumer e-bikes, computing and industrial applications. In 2020, we released a breakthrough power IC, 200-millimeter technology, Gen6, which is now prototyping with multiple customers. This technology offers over 35% power efficiency improvement and/or equivalent amount of die-area reduction at 24-volt operation through an innovative transistor design.

This new technology complements our platform leadership positions at lower voltages with our previously announced 65-nanometer BCD or 300-millimeter process and at higher voltages with our recently announced 140-volt RESURF and 200-volt SOI technologies. As validation to the value of these power platforms, customers are approaching us for long-term volume contracts for which we have already signed one significant one.

Looking at power discretes, we see strong recovery from most customers, included, but not limited to, our Tier 1 MOSFET customers.

Moving to our sensors and display business unit, first to discuss CMOS image sensors. In the past quarter with Opix, we introduced a state-of-the-art indirect time-of-flight, iTOF, imager with unparalleled performance and accuracy and sensitivity. Based on Opix measurements, the sensor [17] accuracy are better, meaning higher level, higher performance than the 2 otherwise industry-leading iTOF sensors in the market. This sensor will enter volume production in the second quarter of this year. It is planned to be embedded in smartphones and other devices for face recognition and 3D imaging applications, such as fast autofocus and artistic picture-focused blurring effects. The sensor is based upon our unique pixel level stacking, state-of-the-art platform with the best in industry, less than 2-micron electrical connection pitch.

During last year, we engaged in several programs of large x-ray sensors, some already having moved to production. Our differentiation in this market is in pixel performance, especially sensitivity and linearity, and in yields. For 300 millimeter, we are the only foundry to supply [such] sensors in mass production. 300-millimeter tooling enables very high yields, and very importantly, a design advantage to manufacture a full 21 centimeter by 21 centimeter detector from one wafer, and hence, eliminating the need for expensive wafer tiling.

Our next-generation industrial sensors on 300 millimeter, using our state-of-the-art global shutter pixels are also ramping into mass production. Our pixel size of 2.5 micron is the smallest in the world. We provide high-resolution sensors with current maximum resolutions of 288 megapixel with excellent sensitivity and shutter efficiency for the display screening market. We expect to see many of these new machine vision sensor products based on our 65-nanometer 300-millimeter platform ramp to production this year.

The industrial market continues to grow steadily. And we expect to see nice growth of these high-resolution sensors, tens of millions of dollars at a high margin. The lifetime of such products is long, 5 to 8 years, so we expect high margin, steady business, based on these products. If we look at, for example, one such industrial sensor market, manufacturing lines for TV, laptop, smartphone, among others, there is a need for display inspections, which drives a large demand for very high resolution, fast global shutter sensors that meet the tight form factor requirements of the optics perfectly matched to our high performance, smallest in the industry, 2.5 micron global shutter pixel.

Alongside the new 300-millimeter product adaption, we see notable increased 2021 forecast for our existing 200-millimeter advanced platforms. The imaging area, that for us was hit the hardest by COVID, was dental X-ray. We are encouraged to see an initial increase in purchase orders, and customer forecasts now show second half of the year fully recovering returning to pre-COVID run rates.

Moving to nonimaging sensors and displays, there are 3 end markets that we are focusing on. For each of these, we chose a customer development partner who has a differentiated capability. In the MEMS area, we are entering the MEMS microphone market. This is a fast-growing market with microphones being embedded not only in ear buds and cellular phones, but also in many command operated devices. Speech recognition AI is being used in such devices. For high-fidelity speech recognition, differentiated performance of high-dynamic range and low-noise microphones are needed. We entered this market with a partnership with GMEMS as press released in Q4 '20. We are in the initial production ramp and are moving forward on developments for the best-in-industry signal-to-noise figure of merit. MEMS microphone is a large, new serve market for us reported to have been a $1.2 billion market size in 2019 with analyst projections of $1.7 billion in 2024.

The display market is undergoing a dramatic change from LCD-based screens with LED backlighting into micro LED or micro OLED displays, allowing substantially higher dynamic range with true black and higher brightness. In entering this display area, we announced our partnership with Aledia. This partnership continues well with developments in preparation for mass production of their unique gallium nitride nanotube-based micro LEDs, which offer unique figure of merit superiority at substantially lower cost than existing volume manufacturing solutions. In addition, we continue forward with our technology development of CMOS back plane for stitched large die micro OLED array for the virtual reality market with a significant market leader.

Moving to utilization. Our customer base continues to grow. The demand of our customers, existing and new, continues to grow and grow strongly. This is good validation of the value of the previously described technology offerings. To meet this increased demand, we are investing $150 million to increase our capacity as well enable some existing capacity to serve new higher-margin offerings. We are investing in Tonami Fab 5 200-millimeter, Migdal Haemek Fab 2 200 millimeter, San Antonio Fab 9 200 millimeter, and an additional investment in our Uozu Fab 7 300-millimeter site. These expansions will have the potential of adding about $150 million of revenue on an annual basis once fully qualified and [utilized]. We will begin to see some incremental revenue benefit in the second half of 2021 targeting full revenue capacity during the first half of 2022.

Fourth quarter utilization levels were as follows: Migdal Haemek, Israel Fab 1 our 6-inch factory was at 64% utilization; Fab 2 is at 76%; Newport Beach Fab 3 was at 75% utilization; our San Antonio Factory Fab 9 was at 67% utilization. Looking at our TPSCo fabs in Japan, utilization for the 8-inch foundry business is at about 65% rate, and our 12-inch foundry business was at about a 90% rate.

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

Thank you, Russell. Welcome, everyone, to our call, and thank you for joining us today. We released fourth quarter 2020 results today demonstrating double-digit percentage quarter-over-quarter and year-over-year revenue growth as well as very strong margin growth and balance sheet financial indicators. We also announced today a new $150 million capacity expansion plan that we initiated in 4 of our 7 publications, focused on our 8-inch and 12-inch facility, then due to our customers' demand forecast that are exceeding our current capacity.

I will now move to our fourth quarter and full year P&L highlights and then discuss our balance sheet and cash flow financial statements.

Revenue for the fourth quarter of 2020 was $345 million, reflecting 11% revenue growth as compared to $310 million in the prior quarter and 13% revenue growth when compared to $306 million in the fourth quarter of 2019.

Looking at our organic revenues, which are defined as total revenue excluding revenues from Nuvoton Japan, previously Panasonic Semiconductor and excluding revenue from Maxim in our San Antonio fab, revenue in the fourth quarter reflects 20% quarter-over-quarter growth and 17% year-over-year growth.

Gross and operating profit for the fourth quarter of 2020 were $70 million and $33 million, respectively; $17 million and $14 million higher than in the prior quarter, respectively; and $15 million and $14 million higher than in the fourth quarter of 2019, respectively. Net profit for the fourth quarter of 2020 was $31 million or $0.20 basic earnings per share or $0.28 diluted earnings per share, which is $16 million higher as compared to net profit of $15 million in the prior quarter.

EBITDA for the fourth quarter of 2020 was $96 million, $17 million higher as compared to $79 million in the prior quarter and $21 million higher as compared to $75 million in the fourth quarter of 2019.

For the full year of 2020, revenue was $1.266 billion, $32 million higher than in 2019, and gross and operating profits for 2020 were $233 million and $91 million, respectively, as compared to $230 million and $87 million in 2019, respectively. Net profit for 2020 was $82 million, representing $0.70 diluted earnings per share as compared to $90 million net profit or $0.84 per share diluted in 2019.

We also announced today a new capacity investment plan of $150 million in the majority of our facilities due to customer demand forecasts that are exceeding our current maximum capacity capabilities. The investment is expected to be made in the coming 12 months in our Fab 2 facility in Israel; Fab 9 in Texas, U.S.; Fab 5 in Tonami, Japan; and Fab 7 in Uozu, Japan. The equipment will begin to have incremental revenue impact during the second half of 2021 and targeted to be fully qualified during the first quarter of 2022. We believe the CapEx payments will be mostly made between mid-2021 and the first quarter of 2022. In addition, in relation to the buildings and facilities that are used by us for TPSCo manufacturing, the lease contract for such building and facility was extended through at least 2032. This lease contract extension resulted in an increase in fixed assets and liabilities of approximately $60 million recorded in our balance sheet under U.S. GAAP ASC 842 named leases.

I would like now to describe our currency hedging activities. In relation to the Japanese yen, since the majority portion of TPSCo's revenue is dominated in yen and the vast majority of the TPSCo's costs are in yen, we have a natural hedge over most of our Japanese business and operations. In order to mitigate part of the remaining yen exposure, we executed 0 cost cylinder hedging transactions. These transactions had hedged currency fluctuations to be contained in a narrow range as compared to the spot exchange rate. Hence, while the yen rate against the U.S. dollar may fluctuate, the impact on our margins is limited.

In addition, in relation to the Japanese yen impact on the balance sheet, we have a natural hedge on JPY cash and JPY loan balances due to the extent -- to the extent the loan amount does not exceed the cash amount. This helps to partially protect us from potential impact of yen fluctuations. Lastly, in relation to fluctuation in the Israeli shekel currency, we have no revenues in this currency. But since approximately 10% of our costs are denominated in Israeli currency and we have some liabilities denominated in NIS, we also hedge a large portion of this currency risk by: a, engaging 0 cost cylinder transactions to mitigate exposure resulting from our shekel-denominated cost; and b, investing a portion of our cash in Israeli marketable securities denominated in the Israeli currency to mitigate exposure resulting from the shekel-denominated payables and accruals.

Looking at the balance sheet, we present a strong and stable financial position. Property and equipment increased from $682 million as of December 31, 2019 to $839 million as of December 31, 2020. The increase is mainly due to: a, the 12-inch fab capacity expansion program we announced already in July 2019, which equipment mostly arrived during this year -- during 2020; and b, approximately $60 million recorded following the extension of the lease contract of TPSCo's building and facilities in Japan as assets and liabilities under U.S. GAAP, ASC 842, as explained beforehand.

Short-term and long-term debt balance in the balance sheet increased as well from $312 million as of December 31, 2019 to $390 million as of December 31, 2020, mostly due to the signing of the extension of the building and facility lease contract in Japan. As described before, this lease is treated as a capital lease, thereby, increasing fixed assets and liabilities by approximately $60 million net.

Our shareholders' equity reached a record of $1.45 billion. Our cap table consists of 108 million outstanding ordinary shares and an additional 2 million ESOP-related shares, resulting in a fully diluted share count of 110 million. Current assets ratio, defined as current assets divided by short-term liabilities was 4x.

And the last note on our cash flow report. In the fourth quarter of 2020, cash flow generated from operations was $73 million. Investments in fixed assets, mainly for manufacturing equipment, were $64 million, which included investments to increase our 12-inch fab capacity in Japan. In addition, we repaid $8 million of our debt during the fourth quarter of 2020.

For the year 2020, we generated $277 million cash from operations. We paid $64 million of debt, and we invested $257 million in fixed assets, mainly for the purchase of manufacturing equipment, including investments to increase our 12-inch fab capacity in Japan under our release from July 2019.

And now I wish to turn the call back to the operator. Operator?

(Operator Instructions) The first question is from Rajvindra Gill of Needham & Company.

Congratulations on all the great results exiting the year. Russell, on the $150 million CapEx incremental CapEx investment, maybe you could talk a little bit about the supply/demand environment. This earnings cycle across semis, capacity is extremely tight. Demand has come in a lot faster than expected. And so the industry is scrambling across the board to increase capacity. So obviously, you're putting $150 million to increase the capacity. I'm just wondering what your thoughts are on the current supply demand environment. What was kind of the basis of the $150 million? Any thoughts on -- and do you think that should be enough to meet the surge in demand that's happening across your business?

Well, certainly, the investment is triggered by an increase in demand. We have a model as many companies will or do to have a prescribed ROI off of investments that we make. $150 million investment would obviously mean that we have it pretty much spoken for. The environment, you are correct, is a very strong demand environment. That environment makes things sometimes challenging. It also gives opportunities as you're making decisions off of an increased demand. We have some opportunity and benefit in that, that increased demand is off of a good revenue base where the factories were not fully utilized. So there is open utilization right now that can be used to grow revenue. And as you're looking at that, you always have to have, I think, a loyalty to customers and their previous run rates.

But on the increase in demand, over time, you move that towards a different mix, like not the previous run rate, but the incremental against previous run rates to where you would favor those that are going to be giving a higher margin, and that's a good thing. So not just the fact of an increased demand, but as we continue to release new platforms, the increase of demand is against a nominally higher margins. We expect that in the -- any time that you have a ramp itself the ramp costs money. So -- and for many, many reasons, you're increasing headcount, you're increasing other incentives, you're increasing your supply. So you don't see necessarily an immediate margin increase. In fact, maybe to the opposite, you see a margin decrease as you start a ramp.

But come the third, fourth quarter through both increases in utilization as well as the ability to favor higher ASP mixes, we believe that we'll see a very, very nice margin increase in the fourth quarter that would maintain as we then continue the utilization -- not just utilization, but utilization capacity increases that will come through this $150 million investment. So hopefully, that answers your question. I think it's a...

Yes, it does. No, that's helpful. And so to the margin point, Oren. So as the utilization rates start to kind of move up, as you fill the fab, and to Russell's point, you will be kind of migrating higher margin process flow, so you'll be mix shifting to higher ASPs and higher process flows, how are you thinking about the gross profit fall through kind of metric? You talked about 50% to 55% on incremental revenue. Is that -- do you think that would be going up as you go to the third quarter and fourth quarter? Any thoughts on the kind of gross margin?

Yes. So I believe, as Russell indicated, I mean, firstly, yes, on the long-term plan, we should see the 50% incremental margin, maybe starting Q4. Maybe a little bit Q3, but for sure, starting Q4 because like Russell indicated, there should be time now that we are increasing our maybe employee base, our cost, our working capital in order to be ready for that significant ramp that we are facing. With the additional equipment, of course, comes a maintenance contract and things like that. So those tools, like mentioned by me and by Russell in the call will arrive in the coming 12 months. But until we'll be qualified, it will take a few quarters and will contribute to the revenue. So for sure, Q4 or Q1, you will see the 50% incremental margin. For the coming quarter or 2, you will still not see that, of course.

Okay. So for the first couple of quarters of this year, we should be thinking about below 50% kind of gross profit fall through until these things get qualified and then getting above 50% kind of Q3, Q4?

Yes. Yes. And certainly Q4, yes.

Certainly Q4. Okay. And then a question, Russell, on the end market, you're seeing a nice recovery in power discretes, which was down, as you said, 12% year-over-year in 2020. So any kind of specific color on what's driving kind of the rebound in that? Is it by region? By end market? And then another follow up, just on the automotive kind of shortages that are impacting auto production, any thoughts on how that might be affecting your business?

So as far as the discrete itself, we see Q1 an increase in discrete against what we had in Q4. We see continued and reasonable increase in Q2. And in Q3 and Q4, staying at a good rate. The -- my belief is that one of the big reasons for the discrete drop that was quite severe in the previous year was that a lot of discretes are served through distribution centers and distribution centers are by model, in good times, they have quite a bit of inventory because their only differentiator is being able to supply parts when they're asked for.

I believe that distribution centers react more quickly than anyone else whenever there's bad news, and maybe overly correct when there's bad news. So if you have our customers that have a big portion of business going to distis, that will always overcorrect. I think that, that's one reason that there was such a big drop in discrete. But other than that, I didn't necessarily have good reasoning for discretes' dropping as our power management itself did not drop. And I think that they more or less go hand-in-hand in end markets. So -- but that's what I believe was the reason for it. And it's coming back now. Will the distribution centers go to the inventory levels that they had before? I don't know. But we do see that snapping back. However, I am not saying that discretes are presently or in Q2 at the same levels as they had been in previous years. It's still not back to those levels. But it is up against what we were seeing in 2020.

The other areas where we had been down in 2020, and as I had stated previously, the dental sensors are coming back in full force. And that -- we see a very nice increase in POs and forecast that bring us back -- at least back to pre-COVID times and potentially maybe even beyond. On industrial sensors, we had seen an impact even before COVID, mainly driven I think through some trade wars and a decrease of some manufacturing lines being built for which obviously industrial sensors are used for, and industrial sensors are -- they're humming right now.

So I think maybe one of the beauties that I think of is that, we've -- in our own business, where we had growth in many areas last year, and those levels that we had achieved through the growth will maybe continue or at least stay at that same level as we have been. All areas that were down last year are either fully recovering or recovering to a reasonable extent. And that's a good thing. And when you have a fourth quarter that has a 20% organic growth year-over-year, and within that, you still have certain weaknesses. And then on top of it, to have that base entering into the new year, and have the weak portions that are returning and the forecast of growth. That's a very nice position.

Good combination, yes.

It ties into your first question as well about how confident are we on the demand? That's the reason for confidence in the demand.

The next question is from Achal Sultania of Crédit Suisse.

Maybe a couple of questions from my side. First, on the demand. Obviously, we are actually seeing demand recovery across all end markets, which is pretty obvious. So clearly, your outlook for the rest of the year seems pretty confident. What I'm trying to understand is, how much within that outlook can we assume for some kind of new projects or customer wins that you're probably working on? And could that be a meaningful contributor as we go into the second half of this year? And maybe if you can shed some light on those specific areas. I think you mentioned iTOF as one of the projects that you're working on, but I'm just trying to understand what are the new things that you're working behind the scenes. That's one.

And the second one is on CapEx. Obviously, you're spending this $150 million which is going to be spread over this year and probably next year. So just trying to understand how should we think about the CapEx number for this year. And if your old model of $45 million per quarter of maintenance CapEx is still valid and what does that mean for full year 2021 CapEx?

Okay. So on the last part of the question, Mr. Shirazi will address that. On the first part of the question, I don't know that we have anything behind the scenes that we're working on that I would mention right now. The things that I've mentioned are -- for a reason, they're may be too premature to mention. But those things that we have mentioned that are big growth drivers in the company, one for sure is silicon photonics.

If we look at the 2019 to 2020, revenue increase was substantial by percentage, not substantial in the amount of millions. If we look at the 2021 to '20 forecast, it's remained substantial in percentage and becomes meaningful in millions as well. And then as you would look into 2022 and beyond, obviously, as I stated, maybe, obvious off of what I stated, in addition to Inphi where we press release going into volume production, we have 30 other customers. And that's quite a big customer base to be having -- working on the forward movement of a silicon photonics platform, that really complements an area where we already have a 60% market share. So I think that, that's definitely very strong.

If we look just at the silicon germanium itself, we have multiple generations of silicon germanium. I talked about the 2 different areas, one being areas that is being driven by 100 gigabit per second to 800 gigabit per second capabilities. And I think that, that's quite a big area with new platforms. So our -- what we call H5 platform, which is our highest speeds silicon germanium, many customers are taping out to it. Those that had taped out to it are now going into higher volumes that getting qualified. So as far as the growth, there's certainly growth in new platforms. And that, I think, is one of the strengths that we always look at, is having a very, very aggressive road map tied with our customers as to what our customers are going after and needing.

If we look at our RFSOI, the 65-nanometer platform, we've released a more advanced platform and worked with customers actually to have a very, very good LNA tied to that platform. So we're always doing more developments. Is that necessarily new? It's certainly a new platform that customers are taping out to and designing and working forward to where there's very, very high demand. And all of that becomes very exciting. If we look at the number of mask sets in Q4, it was probably the highest ever in the company history. So there's certainly many new tape-outs going on, but they are mainly off of the activities that we've been talking about.

iTOF should start growing in this year. Will it become an absolutely significant portion of revenue for the company this year? No, I mean, it will not. Is it in and of itself a significant advancement and a base of what can grow in the future? Definitely. So -- but across the board, we have many, many things that are happening.

If we look at the nonimaging activities that I spoke to, one area really that has grown already and will continue is what we're doing in MEMS, in the MEMS microphone. As I mentioned, very specific customers and one I didn't give the name, but it's -- I certainly won't speak to the forecast of a given customer. And when we're entering new markets and state that we have it, those specific 3 markets tied to certain customers and very differentiated platforms, it becomes a little difficult to talk about the degree of growth that we're expecting in it because it's information that obviously then goes to them, and it's really their volumes. It's not my volume. It becomes my volume, but it's their market.

So -- but certainly, what -- we think Aledia is a very, very strong opportunity. We think that the VR opportunity we have is extremely strong. And honestly, it can be very, very significant volumes. And what we're doing with the MEMS microphone, we think is already strong and will continue to grow.

So across the board, if I talk about power management, we have a huge amount of activity happening in power management. And this is, as I stated, with Gen6, you either have greatly improved performance, got the same die size or a person can have a much cheaper wafer because of the reduction in die size because of the performance of the Gen6 platform. So I -- again, if I state that -- and I didn't say it in the call, but we did have the highest number of mask sets in Q4. There's obviously then a lot of new activity that customers are doing, designing to existing advanced platforms and designing to new platforms that will be driving growth. So hopefully, that answered your question.

Yes. No, that's very helpful, Russell.

Yes. On the CapEx question. Yes, so I believe I said in the script, that we believe that the CapEx payment will be mostly made between middle of 2021 and the first quarter of '22. So you may assume that most of it, I don't know, $90 million or $100 million from that or even more will be paid in those 3 quarters, Q3, Q4 and Q1. So maybe you want to assume $30 million in each one of them. And there will be some that needed to be paid Q1, Q2 as payments to all facility and towards the first payment of the tools, but it will not be significant.

So if you ask me about the total forecast, I believe it should be -- Q1 should be back to the levels that we see -- we saw before this year at the $45 million per quarter, which we had before the [Uozu] expansion. Could be slightly up to $49 million or $48 million, not more than that. Same for Q2, $45 million to $49 million. And Q3 and beyond should have an additional maybe $30 million per quarter for CapEx. So eventually, until Q1 '22, maybe slightly some pushouts to Q2 '22, we will exhaust the $150 million.

The next question is from Cody Acree of Loop Capital.

Congrats on the progress. The capacity constraints that you are seeing do not quite work out in the numbers that you gave us for utilization rates. I would have expected those rates to be higher. Is there just a timing issue? Or are you strategically setting aside capacity for what you now are seeing in your visibility to be much higher demand coming? I guess, why is the utilization rates reasonable, but not seeing that from others?

I don't necessarily understand the question. The utilization rates were the utilization rates. The demand is very, very high. That goes beyond our 85% model. So one would expect and as I stated, that there is room for growth above the utilization rates that I had stated because they were not at 85%.

Now some of the CapEx that is being invested really is to address mixes and to be able to do higher-margin mixes. Some utilizations are a little bit low because we report utilization in a very, very objective way, and it's against photolithography capability. And I have stated that multiple times that every line because you're running different mixes in the line has other bottlenecks that are not necessarily photo capability. So as you're increasing demand, part of that CapEx is to enable a freeing of bottlenecks, so you can maximize the photolithography itself. And that's part of this CapEx investment is, as we're evolving as for the previous question, new platforms are coming out, you want to change the mix capability of the factory. So some of it is changing mix capability, but we are also adding full lithography within this $150 million investment and not an insignificant amount of photolithography either.

So the -- a lot of the CapEx that we're investing is to enable existing lithography to do different mixes and additionally adding more lithography to it. Does that answer your question, Cody? Hopefully, it does.

It does, Russell. Let's see, the strength that you're seeing across the board, could you give us any quantification, any color to stratify the end markets that you see as most in the highest demand and the most constraints? And then I guess just with automotive likely in that list, have you been able to massage capacity toward that end market demand, maybe outside of what you would -- how you would normally shift per customer need?

We have very few customers that we ship to that are 100% serving automotive. So it's in the demand that we have from our customers. Within whatever their run rate is, whatever their allocation is, they really have the decisions themselves to move it to automotive or not move it to automotive. We don't make very much decision on that if we're going to give more or less to automotive because we don't necessarily have, other than one customer that I can think of, we don't have any customers that I would say everything we ship to them is for automotive. So that really does, more or less, stand in their mind. It's not a decision we make.

Now as far as the end markets, in the near term, we have very, very strong demand for 200-millimeter RFSOI, what we call our QT9 platform. We have very strong demand for 300-millimeter RFSOI, and we have very strong demand for -- through power management, both really extremely strong demand at 200 millimeter and strong demand as well at 300 millimeter. So in the short term, those are very, very high demand.

In the mid to long term, we have very big demand within image sensors and the others maintain the -- when I say short term, I don't mean that it ends. But the biggest short-term demand is really in the area of RFSOI. As far as demand increases, it's RFSOI and power management. And then in the mid to long term, I would add the image sensor into that. At least the image sensor business unit, so image sensor and displays.

The next question is from Natalia Winkler of Jefferies.

So I was wondering if you can give us some color about the split of the new capacity that's being added between 200 and 300 millimeters. Like how do you think it would roughly fall between these 2? And then as we think about the total new revenue max for TSEM, I think our estimates before this were about $1.6 billion. And now you're saying that this new capacity would add about $150 million per year. Is it fair to just kind of sum up these numbers and then think of the total as a sum of these?

The second part of the question, I'll have to think about for a minute. I -- the $1.6 billion was before we had a reduction in the Panasonic contract that we had released last year, right? So I don't know that you're dealing off of the correct base at this point. But -- and I'm sorry, the first part of the question, was it a 300-millimeter to 200-millimeter split? The bulk of it is 200 millimeter. As discussed, previously, we added last year about or we invested last year about $100 million for 300-millimeter capacity increase. And we're giving an additional multiple tens of millions of 300 millimeter, but the bulk of the investment is 200 millimeter.

Understood. That's very helpful. And then as we think about kind of longer term, longer term beyond 2021 in terms of how you would see to expand your capacity further. I was just trying to understand if there's kind of room for you to continue with organic expansion. Or if at a certain point, you may be reaching kind of the max that you can add organically at 300 or 200-millimeter fabs?

It's a very, very good question. 200 millimeter, we have a reasonable amount of leeway. That's a question of working out whatever peripheral agreements one would wish to have or could have to increase the manufacturing footprint in our San Antonio factory. San Antonio sits on over 100 acres of land that we own. And it's a beautiful facility, a very well-run facility. So we have, if you will, an infinite amount of build-out capability there. It's simply a question of striking the right agreements to grow. And I'm not trying to speak out of turn here or whatever, but it's pretty obvious and some of the questions about automotive.

There's a lot going on right now in the U.S. legislation with regards to bills, infrastructure bills, CHIPS Act, et cetera. And there becomes questions then about how proactive the government wishes to be in order to build capabilities onshore, and if we would be the right partner for them to start doing that with. So that's -- as far as 200-millimeter footprint, honestly, we have huge opportunities to grow with adding additional clean room space. Some of that would be converting maybe gray area into a [white] area, others is really adding additional buildings.

So San Antonio is a prime spot that I would say we have, if you will, infinite expansion capability, providing that the investments can get the right ROI. And that becomes -- to a degree, what and how much the government will wish to partner on that. I think one thing for sure is that the automotive needs become very, very big in the U.S. and very specifically, San Antonio does a lot of automotive manufacturing. So it sits really in the core of what the U.S. wants to get done. San Antonio is also a place that we have put our advanced silicon germanium flows as well, which for RF communication is a very, very big deal for the United States. So we'll see.

But to answer the 200 millimeter, we really have, more or less, potential for unlimited growth depending that the ROI becomes proper to do it, and ROI is always a good decision. On a business model and a financial model, growing something organically is very, very good on a margin basis because you have a lot of fixed cost that's already absorbed in that factory or a set of factories. So that's always a good thing on the long term. The question is the upfront investment and how much do you have to do and then depreciate and if there is abilities to partner with the government that becomes obviously advantageous maybe for both people.

Okay, on the other side, the 300 millimeter, we still have capability to grow in our existing facility in Japan. However, any growth at this point beyond this incremental new investment that we're doing, will necessitate facilities work as well and facilities work becomes expensive. So within 300 millimeter, there is ability to grow. The growth is not unlimited. And it becomes a question of the ROI of doing facilities work for some finite additional growth versus the potential of doing a deal outside of organic growth.

If I may add, Russell, organically, we can also add not much, but we can add in Fab 2 and in Tonami.

Yes. And we are. That was part of the...

Yes, but, in addition even.

In addition, correct. I'm sorry, that is correct. And we have plans beyond this investment we're talking about to grow Tonami as well without needing to do facilities work, that's correct. But not unlimited.

Yes. Yes, not infinite.

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

I guess a couple of questions on a similar theme here. You gave us some numbers about growth in a few different application markets. And if I got the numbers right, you said mobile was 22% growth last year, power ICs, at least organically, it was 25%, and then you had silicon germanium/infrastructure at 15%. You seem to be talking fairly positively about all those markets here. How would you see growth in each of those markets relative to those -- that performance you had last year? Are they similar, better or worse? How would you characterize those?

I specifically didn't state. I stated that we'll have growth in the company throughout the year. And if I look at those markets itself and was going to speak to it, let's see. I think we'll maintain growth in silicon germanium. As I stated earlier, a very, very good growth in FIFO. We'll continue growth in power, and we'll see certainly year-over-year growth within discrete. So I -- but -- and the RFC model also continue to see, I believe, strong growth. But I didn't get to and didn't really want to -- I didn't wish to give numbers.

Okay. That's fair enough. Maybe in one of those areas in RF, you talked about gaining share. Do you still expect to gain share in 2021?

Yes, sir.

Okay. And where do we sit in share right now? Is this -- do you have a majority of the market, do you think? Are you getting close to that?

I'm sorry, the question one more time, please, Richard?

Again, on RFSOI, you said you're gaining share, where do you sit with share today? Are you at or close to a majority there?

There's multiple players there. I think we're probably sitting somewhere in the mid-20s.

Okay. All right. That's helpful. And then in silicon germanium, do you expect 400 gig to be a material contributor to sales this year? Or is it more in 2021?

2021 is this year.

I'm sorry, 2022, sorry.

Yes. We're starting to see a growth in 400, which obviously, it's not a 400-gig chip. It's a 4 by 100 or an 8 by 50. But the -- it will grow stronger. What -- I think the big deal is that we don't see a decrease in the 100. And it then figures that the 400 growth is not cannibalizing the 100-gig baseline.

Okay. That's helpful. One last question for me. I suppose we could address this to all of your kind of new areas of growth. And you identified a few of them in the nonimaging space like micro LED. When do you see that opportunity kind of blossoming here? And as you said, you've kind of selected a first partner to help work in that area. To what degree is the work done there? Or customer engagement there more broadening to see that pipeline grow, where we see micro LED becoming a really big deal and I don't know, 1 to 3 or 4 years, how would you look at that?

Well, we know that customer, as I mentioned again today, our customer there is Aledia. And as I stated, I really don't want to talk about Aledia's guidance to us. Number one, it's not proper for me to. And as far as micro LED, I think it's the right way to go. I think nanowire is a very, very exciting technology. I think that Aledia has an incredible technology and very strong differentiation. But will micro LED grow? Certainly. Is it on 2021? I don't think so. Is it 2022? I think that, that will start. And I think 2023, 2024 will be very big or will become very big during those years. But I don't think 2021 is going to be huge at all for that, but I think it will start to grow in '22 and will become significant in '23, '24 and '25. What is your thought?

I'm hearing similar time frames, but just want to get your sense as well. And that's helpful. I'm glad if it's -- fits roughly speaking there, so.

Next question is from David Duley of Steelhead Securities.

Nice results. Could you help just frame what you think -- or help us understand what the size of the photonics business is now or perhaps in 2020 and what your goals or aspirations are for that business?

Okay. I know exactly what it was in 2020 for us. In 2020, it was just shy of $8 million. If I look at 2021 by forecast, it's a substantial increase of that, maybe close to 3x increase in what we're seeing. Where do I think it will get to? I think it will get to, for us, I mean, our targets would be to be in the several hundred millions. So I think that it's very possible for it to get there. But yes, that would be -- our target is for the photonics business to be sitting within some small amount of years, certainly upwards of $100 million. And I think, with the target of getting between $200 million and $300 million.

And does this -- will the photonics revenue have above-average gross margin similar to silicon germanium?

Presently, it certainly has well above-average gross margin. It's very, very high margins at present. As anything goes into high volume, margins come down, but I believe it will stay very high margin. But it will -- yes, I think it will be among the highest margins we have in the company.

Excellent. And then in your prepared remarks, you mentioned something about a long-term contract, and I had some audio difficulties. And so I just don't -- could you just elaborate about what you were talking about there?

If I recall the exact statements, it was -- I'm now paraphrasing. It said that as a validation to the strength of the power platforms that we have, we have multiple customers asking us for long-term supply agreements. Now when they're asking for a long-term supply agreement, that takes on some term or some type of a take-or-pay agreement to where there's a certain volume that they're committed to buy. And there's a certain volume that we commit to give them. Now customer really won't give you that unless your platform is very powerful because they're committed to use that platform. So that was the statement that I made. And if to add color to it, that is color now that it would be involving, to some level, a take-or-pay agreement.

So we would expect a customer deposit liability number to be going up over time?

Not necessarily. A take-or-pay agreement doesn't necessarily mean upfront money. The take-or-pay agreement means that they have x amount of wafers that they have to take per month. It could be against an upfront payment as well, but that's not necessarily how it is. I mean it's -- it could be that they give an upfront payment to enable higher capacity, of which they have -- which are committing a degree of that capacity to them. But there's multiple different contracts that can be done. Sometimes you would ask for it depending on what's being driven. You'd ask for upfront cash to help on an investment, and other times, you wouldn't. But regardless, a take-or-pay agreement really means that they're committed to a certain amount of wafers per month or per time period.

Okay. The final question from me is you mentioned, I think, the MEMS microphone business. I don't recall you talking about that in the past. I remember you've talked about MEMS, but not specifically that type of product. Could you just elaborate as far as what the market opportunity is for the company in this area?

Yes. We actually press released it, David. So we certainly talked about -- it's also in press release. But the market opportunity, I think I had stated that in 2019, the market itself, MEMS microphone was reported to be $1.2 billion. And some analysts believe it will be $1.7 billion in the 2024 time frame, something like that. Now that's the MEMS microphone itself. The silicon content, I don't know, if you take it, 30%, 25%, 35%, but it would be somewhere in that, the silicon content itself.

So for us, the overall MEMS microphone silicon business, I suppose that the served market is somewhere about $0.5 billion. And I think the beauty seriously is that anytime you enter a new served market, independent of how much market share that you grow, it's incremental revenue. And it sounds funny, and I don't mean this to sound really ridiculous, but it's infinite growth in that market because [shorten] serving it before. So if you go from a 0 to a 10% or a 15% market share, it's all incremental growth, not cannibalizing anything that you did before. And so it's -- I think the opportunity is very, very big for us.

As stated, we have a good partner there, and we're working on extreme figures of merit, maybe the best signal-to-noise ratios in the industry. So did that answer your question? I hope. But there is a PR on it. It's -- I believe the PR was maybe December, but I'm not positive. Noit will get back on the exact date of that. We could even prepare...

I'm sorry, I forgot. I'm sure I read it, I just forgot about it. As far as the $500 million TAM that you just kind of highlighted, would a reasonable expectation for you guys be 10% or 20% market share here in a year or 2? Or is that the kind of -- can you get $50 million to $100 million out of this business?

In a year or 2? No, I don't think so. To go from a 0 entry in a market to a customer partner that you're going with, I don't think that you would go to $100 million in 2 years. But I think that our target would be to get to those levels. But it wouldn't be within 2 years. I think that's not realistic. So the press release was December 21, and the title is Tower Semiconductor and GMEMS announced the ramp to mass production of MEMS microphone products.

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

So I'm kind of getting interested in this photonics thing you're talking about. You mentioned something about a DARPA contract where you try to integrate lasers. I know that's a really big deal because it's really labor-intensive. Can you talk a little bit about what's going on with that?

Yes. It's a project, it's called Luminus (sic) [LUMOS]. It actually was released by DARPA itself. And that we were the front runner of everyone working, I believe, to my memory, that's how it was listed. So yes, it's a project with DARPA. It has DARPA funding, and it's to integrate lasers on a SiPho platform. So basically, in addition to all the passes that you put on the SiPho, you're now putting an indium phosphide laser on it and making the connections on the -- to the N and the P and doing different metalization structures so that you have, not just a good-function laser, but signals from a laser going right into the photonics and doing it on board.

So it's to actually create an optical engine by putting the 2 together?

Creating more of a single optical engine. I mean there's other active components that you would maybe think of adding as well.

Interesting. Okay. And then just I'm wondering with this new $150 million investment, does that change your thinking at all about tax rate for this year or the minority income?

So first for this year, it will not have a significant impact because, like I mentioned before, the manufacturing equipment tools will arrive and installed and qualified mostly this year, but until it will bring additional incremental margin and revenue will only be a little bit Q3, but mostly Q4. And secondly, this equipment is actually bought also for Israel, also for the U.S. fabs and also for the Japanese fab. The mix is pretty much the same, like the profitability base of us. So I don't think it will impact the tax. I don't think it will change the -- what we saw this year was very stable at about 6%. All in effective tax rate, I believe it will be about the same.

Oh, 6%.

Yes.

All right. And so no real change in minority income either?

No. It will not impact the change in minority. The investments that we do in Japan will be like the previous model that we explained last year on the July 19 announcement that it will be -- goes to -- it's a foundry customers. This is the target of the acquisition of the tools, and it will be taxable in Israel. So it's a good tax environment.

The next question is from Rajvindra Gill of Needham & Company.

Just a quick follow up, some housekeeping. So Oren, what was the Panasonic revenue and the Maxim revenue in Q4 and for 2020? Just to make sure we're modeling it correctly.

Okay. So Panasonic is actually today Nuvoton Japan, previously Panasonic Semiconductor. So in the past, it was supposed to be between $90 million to $105 million in the previous contract. The new contract is like we said, a reduction of about $20 million. So it's between $70 million to $85 million every quarter. That's the commitment. And indeed, this is the run rate that they were in Q4 and also for the year, so very stable. And Maxim, like we announced before, is gradually decreasing contract. I don't think we mentioned the number of that in the past. But you can assume that Q4 was in line with the previous quarters and which is a little bit lower than 2019 levels.

And just what was it in 2019? Just...

We -- because of the contract with Maxim, we agreed to their request, not to give you an exact number.

Okay. Got it. Got it. Right. And -- so just so I heard it correctly. So the Panasonic Nuvoton is now $70 million to $75 million a quarter and has been stable or $85 million?

$70 million to $85 million per quarter.

So. Okay. Got it. All right. So that makes sense. And you expect that to be kind of stable in 2021 this year?

Yes.

Okay. And just on the CapEx. So what is the -- just so I have it again, it's a $150 million incremental CapEx to last year. So last year was, I believe, $256 million in 2020?

No, no, no. So let's -- it's incremental to the baseline, excluding that Uozu $100 million that we had last year. So the baseline, which let me remind you, the baseline for -- and you can see 2018 and 2019 CapEx was $180 million -- almost $180 million, so $40 million to $45 million a quarter. That's the baseline. 2020 was higher than that because we had this Uozu $100 million investment, which is done behind us. So for -- from mid of this year going forward to the baseline of $45 million a quarter, which is $180 million annual run rate, you should add the $150 million, which is approximately $35 million a quarter, right, if it's spread over 4 quarter.

But Q1, Q2 -- like I mentioned before, Q1, Q2 should be pretty low because payments will be starting to be made in the middle of the year. So Q1, Q2 is back to the baseline of $45 million, $49 million a quarter, which was the baseline in 2019, and for middle of the year, will go up by about $30 million, $35 million a quarter for the $150 million. Clear?

Okay.

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

I would. Thank you. So really, firstly, my great appreciation to our worldwide employee base at all levels and in all functions. Against a very challenging, difficult COVID-driven work condition, they did not miss a beat, finished the year with a very strong quarter and put us in a position to guide our first quarter revenue being the highest first quarter in a revenue base in the company's history. We're excited to execute on the opportunities before us to realize quarterly sequential growth throughout the year. We really have amazing customer partners. We look forward to grow with you. My many thanks to our investors. We greatly appreciate your support and really enjoy our interactions.

With that being said, we invite you to any and all of the following upcoming events. Next Tuesday, February 23, we will celebrate our 20th year of trading in the Tel Aviv Stock Exchange with a special virtual trade opening ceremony. We invite you to join. Information will be available on our website and social media platforms. On the same day, we will also provide a virtual presentation for the Israeli capital market hosted by Oppenheimer Israel. On March 9, 2021, we'll participate in the Susquehanna 10th Annual Technology Virtual Conference. I suppose it's the 10th annual technology conference, probably not the 10th annual virtual one. And hopefully, it will be the last virtual one. On March 11, 2021, we'll participate in the Loop Capital virtual spring conference. Please sign up for these. We would look forward to the interactions and otherwise, we always look forward to interactions. And please contact Noit anytime you wish to speak. Thank you very, very much. Bye-bye. Thank you.

Thank you. This concludes the Tower Semiconductor Fourth Quarter and Full Year 2020 Results Conference Call. Thank you for your participation. You may go ahead and disconnect.