高塔半導體 (TSEM) 2018 Q3 法說會逐字稿

Ladies and gentlemen, thank you for standing by.

Welcome to the TowerJazz Third Quarter 2018 Results Conference Call.

(Operator Instructions) As a reminder, this conference is being recorded, October 29, 2018.

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

I would now like to turn over the call to Ms. Noit Levi, Vice President of Investor Relations and Corporate Communications.

Ms. Levi, please go ahead.

Thank you, and welcome to TowerJazz' financial results conference call for the third quarter of 2018.

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.

TowerJazz assumes no obligation to update any such forward-looking statement.

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

Russell, please go ahead.

Thank you, Noit.

Welcome to our third quarter 2018 results conference call.

We ended the third quarter with revenues within our guidance range at $323 million, resulting in net profit of $34 million and free cash flow of $29 million.

We entered the year having announced in 2017 that we were walking away from lower-margin RFSOI business that had faced extreme 20-point price pressure in order to maintain the business and as well needed to be running Fab 2 Migdal Haemek or in Fab 3 Newport Beach for the short to midterm.

For the sake of higher-margin silicon germanium in Fab 3 and higher-margin RFSOI mix and CMOS image sensor in Fab 2, we did not accept the price reductions.

Some of the higher-margin replacement business did not hit customer forecast as discussed at the beginning of Q3 and left some holes for the third quarter, which continue in the fourth quarter for Fab 2 at Migdal Haemek.

Again, we referred to this at the Q2 call.

The silicon germanium has reached full ramp in Q4.

Although the TowerJazz silicon supply has been well managed, not missing a single customer demand start, some of our business relies upon highly specialized starting material supply directly from our customers.

Here, we faced shortages in Q3 and presently in Q4, resulting in some decrease of revenue.

There are plans in play to mitigate this specific supply challenge for 2019 and beyond.

Our revenue numbers are below our own expectations for 2018.

However, our activities, advancing our technical road maps with associated customer traction, is very strong.

It's that, that I want to focus on for the next little while to show our extreme confidence that what we have seen in Q3, what we're seeing now in Q4 being perturbated somewhat by a -- somewhat of a market pullback is not of a big concern for us as the underlying activities really promise to be continued growth engines for the company.

In the past quarter, we won a major bid for RFSOI switch for a 65-nanometer 300-millimeter platform.

This is a Tier 1 customer who is a significant existing RF industry player and with whom we currently have no RF business.

According to this bid, we will become their sole RFSOI foundry partner.

The customer is planning initial tapeouts by the end of 2018, which we expect will contribute to the growth of RF business in 2020 and beyond.

As per announced, silicon germanium plans, the present work in progress in the Newport Beach manufacturing line will enable an increase of about 70% in silicon germanium shipments in Q4 versus Q1 of 2018, resulting in our fourth quarter annualized silicon germanium revenue of over $200 million, fully utilizing the added silicon germanium capacity in Newport Beach.

The improved product mix resulting from the silicon germanium increase should contribute over $20 million annualized gross and EBITDA margin dollars.

In the past months, I've met with 3 of our Top 4 silicon germanium optical customer CEOs as well as the CEO for an additional new served market for TowerJazz that is presently in a high-volume ramp.

Without exception, we received praise for our support of their business and updates of continued market growth, including reaffirmation of their respective ongoing volume demands.

We continue to expand our silicon germanium capacity with the early 2019 qualification in San Antonio and a further increase of Newport Beach capacity which should result in an additional incremental 30% silicon germanium capacity during 2019.

Together with Dr. Marco Racanelli, General Manager of our RF high-precision analog business unit, I attended a milestone review meeting and a specific CEO/CLO business review with our 2 leading customers in the silicon photonics area.

The continued need for bandwidth and the transition to 100G and beyond for cloud applications and data centers is driving silicon germanium growth today and will enable silicon photonics growth in the future.

Silicon photonics helps integrate several components that today are assembled separately into optical modules and as cycle matures, it will win over these discrete solutions due to its inherently lower manufacturing costs and smaller form factor.

In the optical silicon germanium space, we have a commanding greater than 60% market share and SiPho should open up an additional several hundred millions in market opportunities for which we are working to gain a similar high level of market share.

In the past quarter, we also won a major Tier 1 bid for our 65-nanometer 300-millimeter power management BCD platform, which we only released in the second quarter of this year.

According to this customer, this platform provides tens of percentage points improvements to key figures of merit, driving needed power efficiency and provide as well a substantial cost reduction through efficient layer count against the foundry's, otherwise, industry best present offering.

In the image sensor area, I personally met with 3 very exciting customers in the past quarter.

Capitalized by our 300-millimeter Japanese factory, we won the leading sole position as a foundry supplier with a multi-generation road map for a Tier 1 DSLR customer.

The high level and working-level meeting reconfirmed both the appreciation for the joint activity milestone achievements and the time line for release of our first and second-generation products.

Another meeting was with an extremely exciting company in the area of augmented reality, resulting in a reconfirmation of our win for the next-generation time-of-flight product for our next-generation -- or for a time-of-flight product for the next-generation product expecting high-volume release in 2020 and continued further detailed operational and technical program alignments for subsequent generations.

Lastly, I met with one of the top revenue leaders in the image sensor market with whom we presently have no business.

With strong discussions and activities, including wafers that will surely be shipped from our 300-millimeter line with our most advanced processes, we are encouraged and confident to gain market share with this customer in this area.

Our image sensor management, the General Manager, Dr. Avi Strum, with additional leading technical experts, held a prestigious seminar in San Francisco this past month.

This unique GIS technical seminar showcased our advanced platform and cutting-edge technical solutions and advanced manufacturing capabilities for a wide range of imaging growing markets.

28 handpicked sensor companies attended, with TowerJazz providing detailed technical reviews, analysis and notes on the customer and industry benefits for our customized pixel solutions for global shutter, near-infrared enhanced pixels, backside illumination and stacked wafer, 3D sensing and LIDAR, among others.

On November 5, we will hold additional seminar addressing the European sensor market.

For our TOPS business, most all long-term contractual power discrete business remains with demand beyond our allotted manufacturing capacity with considerations and discussions to further increase capacity beyond our previously announced and completed 2018 increases.

This group is strongly involved in several new technology initiatives, among which are innovative technologies for display required by virtual reality, augmented reality and wearables, among others.

In terms of utilization rates in the quarter, Migdal Haemek Fab 1, our 6-inch factory, was above 90%; Fab 2, our 8-inch factory in Migdal Haemek, was about 80%; and Newport Beach, California, Fab 3, our 8-inch factory, was about 85%, which includes a 5% increase of the capacity in photo layer that occurred during Q2.

Our San Antonio factory, Fab 9, was above 60% utilization with foundry utilization up 5 points.

The 3 TPSCo factories had an average of about 50% utilization.

I'd like to provide a brief update on Panasonic partnership and long-term supply agreement.

We are continuing and in advanced stages of negotiation.

From present discussions, it is obvious that the partnership will continue for multiple years with the next stage contract extending for at least 3 years at production level from Panasonic similar to the present run rate.

Our partnership with Panasonic was and still is a great enabler for our technical and business development and growth.

I previously mentioned 2 big wins, both at advanced platforms of 300 millimeter, either of which alone at volume revenue would be a Top 10 customer level.

The available capacity provides us with the capability to further increase our advanced offerings, as I previously described.

The most likely outcome of these negotiations will be some changes to the pricing table as Panasonic should not continue to cover all the fixed cost of these facilities, augmented by our increased third-party utilization of these factories.

We remain excited about the future activities to further enhance our growth in leading markets, some of which is enabled by this long-term valuable partnership.

In summary, we are confident with our activities, advancing our technical road maps with associated strong customer traction.

As mentioned, I visited lead customer partners for each of our business groups over the past few months, and we are convinced that our focus and engagements are industry leading and will provide strong engines of growth for the next several years.

With our strength in the balance sheet, having reduced debt by $100 million, the notable change to a higher-margin product revenue mix and considering $315 million net cash balance, we are well positioned to capture merger and acquisition opportunities for either capacity or capability or a combination of both.

With that, I turn the time over to Mr. Shirazi.

Thank you, Russell, and welcome, everyone.

Thank you for joining us today.

I will start by providing the P&L highlights for the third quarter of 2018 and then provide cash flow and balance sheet updates.

Revenues for the third quarter of 2018 were $323 million, which resulted in $89 million EBITDA, $69 million cash from operations, $34 million net profit and $0.34 per share in basic earnings per share.

Since July 2018, we have reduced our standing debt by $98 million, which will save $7 million of financing expenses on an annual basis, strengthening our financial position and balance sheet ratios.

This $98 million debt reduction was comprised of: one, early repayment of a $40 million loan that carried an interest of LIBOR plus 2%, which we borrowed in 2016 in relation to the acquisition of Maxim San Antonio Fab; and two, early conversion of $58 million notes that carried an annual coupon of 8%, of which $19 million were converted during the third quarter of 2018 and the remainder during October 2018.

Our shareholders' equity reached a record of $1.15 billion as of September 30, 2018, which is $124 million higher than at the beginning of this year.

Gross and operating profit for the third quarter of 2018 were $73 million and $39 million, respectively, as compared to $79 million and $44 million in the prior quarter, which, in accordance with our margin model, reflected an approximate 50% margin impact from the associated revenue increase of this year.

EBITDA for the third quarter of 2018 was $89 million, representing 28% EBITDA margin and is compared to $95 million in the prior quarter.

Net profit for the third quarter of 2018 was $34 million.

This amounts to $0.34 per share and $0.33 basic and diluted earnings per share, respectively, as compared to $38 million or $0.38 per share and $0.37 per share basic and diluted earnings per share in the prior quarter.

This was compared to $55 million or $0.56 basic earnings per share and $0.54 per share -- diluted earnings per share in the third quarter of 2017.

I would like to elaborate on 2 specific lines in our P&L report, the noncontrolling interest and tax expenses.

One, noncontrolling interest line.

While EPS cost of operations are profitable, the noncontrolling interest line in the P&L is immaterial at $28,000, which is mainly due to our royalty structure.

Under this royalty structure that we described and presented commencing the second half of 2017, the royalties from TPSCo to Panasonic and to TowerJazz are calculated as a percentage of TPSCo's profitability as calculated using TPSCo's Japanese GAAP P&Ls.

The Japanese GAAP net profit differ from the U.S. net profit as far as TPSCo's P&L mainly in the amount of fixed asset depreciation, which is higher under U.S. GAAP when compared to Japanese GAAP.

As a result, a valuation of the assets held a purchase price allocation used for the U.S. GAAP only at the time we acquired TPSCo.

The lower depreciation expense under the Japanese GAAP results in higher profitability of TPSCo under Japanese GAAP and in royalty payments to Panasonic and Tower as compared to U.S. GAAP under which the noncontrolling interest line is calculated from the lower profits.

As a reminder, this royalty structure results in a higher cost in the P&L, lower pretax income, lower tax expenses, lower noncontrolling interests and higher consolidated net profits and higher EPS.

Item 2 in the P&L.

I want to elaborate on the tax line in the P&L.

I want to describe our applicable and effective tax rate.

Our U.S. affiliates, Jazz and TowerJazz Texas, which own our Newport Beach Fab and San Antonio Fab, respectively, are taxed in 2018 at a 21% rate following the recent U.S. tax reform as compared to 35% prior to the U.S. tax reform.

TPSCo's profits from its Japanese operations are subject to an approximate 32% tax rate.

Our profits in Israel from Fab 1 and Fab 2 operations, while subject to a 7.5% statutory tax rate, are not expected to result in any tax payments in Israel for the foreseeable future since we have more than $1 billion of NOLs still to be utilized which can be carried forward indefinitely.

Considering this fact and since we have certain tax exemptions, discounts and credits in Israel mainly, our all-in worldwide weighted average effective tax rate is 7% for the third quarter of 2018, 7% also for the second quarter of 2018 and 6% for the first 9 months of 2018.

The 2017 all-in effective tax rates were even lower than that, mainly because we commenced incurring the 7.5% Israeli statutory noncash tax for the Israeli operations since the fourth quarter of 2017 due to the realization of the tax-deferred asset recorded in 2017.

I will now provide the cash flow and debt highlights for the third quarter and the balance sheet analysis as of the end of September.

Free cash flow for the quarter was $29 million comprised of $69 million cash from operations and $41 million investment in fixed assets net.

The other main cash activity during the third quarter of 2018 were $43 million of debt repayments net comprised mainly from the early repayment of the $40 million loan borrowed in 2016 in relation to the acquisition of the San Antonio Fab, for Maxim and N-trig; cash, including marketable securities net of gross debt as of September 30, 2018, totaled to $315 million as compared to net cash of $226 million as of December 31, 2017.

The gross debt outstanding amount as of September 30, 2018, was $293 million or $254 million post the Jazz notes conversion as compared to $334 million as of December 31, 2017.

The gross financial debt as of today, October 29, 2018, comprised mainly of approximately $100 million in bank loans in Japan and $122 million in debenture Series G.

Net current assets as presented on the balance sheet for September 30, 2018, increased to $709 million, resulting in a current ratio of 3.9x as compared to a ratio of 2.9x as of December 31, 2017.

Shareholders' equity as of September 30, 2018, has increased by $124 million from the beginning of this year and has reached a record of $1.15 billion.

Basic outstanding share count as of today, October 29, 2018, is 105 million ordinary shares, following the full conversion of Jazz notes, which I described earlier.

The fully diluted share count remains at 108 million ordinary shares on a fully diluted basis similar to that of the previous quarter.

I would like to describe now our currency hedging activities.

In relation to the euro currency, we have almost 0 business in euros.

Hence, no exposure to the euro in relation to the Japanese yen.

Since all Panasonic revenues are denominated in yen and the vast majority of TPSCo's costs are in yen, we have a natural hedge over most of our Japanese business and operations, excluding the portion in which the yen-denominated variable cost related to the third-party foundry business exceed the yen net gains from the Panasonic business.

In order to mitigate a large portion of this net yen exposure, we have executed 0-cost cylinder hedging transaction.

This 0-cost cylinder transaction hedge all 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, our margins are almost not impacted.

In addition, in relation to the Japanese yen impact on balance sheet, we have a natural hedge on cash and loans balance since the loans and the cash are both yen-denominated.

This helps to protect us from potential impact of yen fluctuations.

Lastly, in relation to fluctuations in the Israeli shekel currency, we have no revenues in this currency.

And while less than 10% of our consolidated costs are denominated in the Israeli currency, we also are hedged to a large portion of this currency risk using 0-cost cylinder transaction.

This ends my summary, and now I wish to turn the call back to Noit Levi.

Noit, please go ahead.

Thank you, Oren.

Before we open up the call to the Q&A session, I would like now to add a general and legal statement to our results in regard to statements made and to be made during this call.

Please note that the third quarter of 2018 financial results have been prepared in accordance with U.S. GAAP.

The financial tables in today earnings release also include certain adjusted financial information that may be considered non-GAAP financial measures under Regulation G and related reporting requirements 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.

Taking advantage of business and road map planning activities taking place in Migdal Haemek at this time, we have invited Dr. Marco Racanelli, our Senior Vice President of RF, HPA and A&D business unit as well as our Newport Beach Site Manager; and Dr. Avi Strum, our Senior Vice President of CMOS Image Sensor business unit to take part in the Q&A session.

Operator, please open up the call for Q&A.

(Operator Instructions) The first question is from Cody Acree of Loop Capital.

Russell, if we could go back to your prepared remarks and some of the statements from the press release, you said some of the higher-margin replacement business didn't meet customer forecast.

So I just want to make sure I'm reading that right.

So this is forecast that they gave you, they came in, debate order pushouts during the quarter.

I guess what accounted for that shortfall to their original forecast?

And you said this was primarily in Migdal Haemek, so if you could help us with the vertical or the products that, that was weak.

And then I have a follow-up.

The Q2 release, there are a series of questions to where it was stated, I believe by yourself, that all of the front-end module switch providers were forecasting their business increasing, and you didn't understand how it would be that someone would have pushed out.

I think that, without overstating anything from any specific customer, there was a customer that made some strong statements about their business and somewhat validated the statements we had made at that release call.

More than that, I really don't want to talk about any specific customer and their business.

The parts are advanced RFSOI parts using a copper back-end technology.

Okay.

Great.

And then if you could discuss a little more in detail the customer supply challenges, and maybe if you could just explain about shortfall and then maybe help to quantify the softness below the midpoint of your guidance and where you're guiding to today.

How do we get to kind of close those numbers or equalize those numbers?

So as far as you're referring to, on silicon supply, for most everything that we manufacture, we order the silicon ourselves and we have strong relationships with the silicon suppliers, either multiyear agreements that was press release with Soitec for the RFSOI or otherwise, we have a certain amount of our business that requires very special substrates on the incoming material.

A lot of these are various PFPE layers of different doping concentrations that we do not supply ourselves that are either manufactured by the customer or procured by the customer.

For that business, we have seen weakness as far as a supply of silicon not a weakness in the demand of the customer, just in the supply of silicon itself.

And to what degree is that, it's on the order of multiple millions of dollars.

So that's one of the deltas to the midrange guidance of Q3 to the $12 million performance in the first quartile, a portion of that sits within the silicon supply that we did not know upon going into the quarter that we'd had that shortage.

Looking at Q4, we see a continuation of that shortage to more or less the same level as we had seen in Q3.

Now the other part of a shortage deals that with any grouping of customers, and you can imagine that the $300 million to $350 million level, there's many, many customers that make that up, there will always be some customers that have asked for pushouts, there will be others that will ask for pull-ins, and there's a give or take of any given quarter, which is the reason why one gives their guidance range.

To give any single number is very, very difficult as customers, until the date of shipment, really have say over whether or not to ship, to a certain extent, depending on the relationship with the customer.

And certainly, on wafers that are started within the quarter, they have tremendous say over do you start them or not.

So you always rely on a certain amount of upsides to compensate for downsides that you know will happen in any quarter, and you try to balance that out.

In the Q3, we would have had or could have had a significant amount of upside had the silicon been available that was not available within that same frame of customer application.

Hopefully, that answered your question.

It does, Russell.

And you said so Q4, it continues but you mentioned some possible solutions for this.

When do you expect maybe to see some of that?

The solutions that I talked about would be ourselves taking more control over the manufacturing of the starting material, and there's discussions being done real time on that.

Okay.

And then also, I guess with companies like Texas Instruments, it sounds like the current shortfall, the softness in the guidance is pretty customer end market specific, supply constrained specific in certain segments, but not terribly broad-based.

When you go to TI's, Walmart's last week where they said that during Q3, they simply saw a mass slowdown in activity across all their major end markets, which really can only be explained as either distributors or OEMs just pushing more inventory risk down the supply chain.

I guess is it fair to say that right now, you're seeing very confined pockets of softness but not the kind of mass slowdown that we're seeing from companies like TI?

And if that's the case, do you -- are you considering the possibility that, that maybe the next leg to come?

We've seen a softness.

I don't think it's anything severe.

The $340 million guidance for us was a natural Q3 to a Q4 guidance.

That's a 5% increase.

So we do see a softness that's certainly below that which we had targeted.

It's not overly severe.

I mean, we had targeted being somewhere at the $360 million, $370 million range, and we saw pullbacks.

We don't necessarily have any customers that have given us any big warning sign that their business is down for next year, that they see anything bad happening in -- truly, in 2019.

And I'm not saying that, that may or may not come.

What we're seeing now I think is somewhat similar to what you stated, and I never want to put myself in your shoes of being a market analyst.

But there's -- can always be a variety of reasons for a aberration in the market or a slowdown or slowness.

You have typically Q4 to where people will build extra inventory with the knowledge that, that inventory if they don't sell it for the end of year, Western holidays or the beginning of the New Year or Eastern holidays, they believe that will go fairly shortly.

If there's geopolitical doubt and uncertainty, people are willing to take less risk on carrying an inventory.

So I think that the first step of any geopolitical uncertainty is not taking risk on holding inventory that may, at some point, become old and obsolete.

And in our industry, certainly, there's enough devices that if you're going to hold inventory for a 4- to 6-month period, you could be supplanted by another supplier or it could be that the next-generation part is taking off and replacing the part that you're holding inventory.

So if I was to say what I think is happening right now, I think it's more of that line of cautionary buy with the idea that if you left 2017, you had maybe the opposite happening that at the end of a very good year with all good signs in the market, people took on a great amount of inventory to not want to possibly be short of end-of-year buys from their end customers.

I think it's somewhat the opposite that people, at this point, don't want to be stuck with inventory that could go to be 0 need and obsolete.

Excellent.

And then lastly, either for you or for Oren.

Nicely improving balance sheet, that coming down stock price, obviously, not where you'd like it to be.

What are your thoughts on your use of cash or would you be considering a buybacks here?

Or what are your other plans for uses of cash?

Well, per our strategy, we said that we want to invest the cash for growth.

We are growth company.

We want to do some acquisition, capacity, capability like Russell talked about.

And also, we have the possibility to expand the San Antonio Fab and Uozu Fab, so this is where we believe the best value to the shareholders will come from growing the business, growing the potential of the company.

Yes.

So -- Noit had mentioned that both Avi and Marco are here because we're having internal planning meetings.

One of the things that is facing us and it's a wonderful thing to be facing us is that the adoption of our platforms in the 300-millimeter factory is happening fairly quickly.

By looking at the forecast and customer comments where we're judging on the market, it is very strongly possible that by the end of 2019, by the end I mean in the third quarter, that we'd have to trigger an additional capacity buy in order to meet the 2020, 2021 demand.

And it would -- as with any of the internal capacity build-outs that we do, we focus them on having somewhere from the time that the tools are up and running of a 12- to 18-month cash return on the investment.

So we make sure that we have that type of demand before we trigger it, but that is one need of cash.

Certainly, anytime that one enters into any situation, independent of how short term it is of a market slowdown, there's many, many M&A opportunities that make themselves available that otherwise wouldn't be available.

We've stated multiple times that we're engaged real-time on a variety of discussions.

We are engaged in a variety of discussions.

Nothing at the point that it should be or needs to be announced.

And in some cases, it's foolish to preannounce anything not just for setting expectations, but you also, with your competition, know exactly what you're looking at.

And it's either create competition and raise prices or possibly lose an opportunity because your mouth was too big.

But on that being said, there are certainly multiple avenues for a cash balance that we're developing.

I think the very, very nice thing is the fact that the cash balance is strong and growing and off of the refinancing that was done or refinancing the retirement of $100 million of debt and the lack of necessity to pay a coupon on that debt, the increased margin in the mix that we have -- I mean, there's -- we've released Newport Beach as a separate entity for the past years.

Newport Beach itself is capable of somewhere of $260 million to $300 million revenue off of that base to increase the margin by $20 million.

That's very, very substantial.

So the activity that we did with the silicon germanium Newport Beach, it's a 7-point increase of that factory.

That's a wonderful thing to do that puts us in a very strong position to be able to use that cash for growth.

And that's how we're going forward.

We stated that we're continuing to invest in Newport Beach silicon germanium capacity.

So it's a very interesting thing because that capacity is funded by that capacity.

The growth that we've done in 2018 on the silicon germanium is creating the cash that's funding the growth that we'll be doing in '19, which is a very, very interesting thing to say and 100% true.

So hopefully, that gives more flavor to the -- to your question, Cody.

It does.

Appreciate all the time.

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

Russell, I wonder if you could kind of elaborate a little bit further on the Panasonic long-term agreement.

You mentioned that it will continue over multiple years.

There might be some changes to the pricing table because Panasonic won't cover some of the fixed expenses, but that will be offset by higher third-party facilities.

I was wondering, when that agreement -- a renewal is going to be kind of finalized.

And how should we think about the cash flow impact of the changes to this new agreement?

So the first part of your question is very easy to answer.

The present contract ends the end of the first calendar quarter.

So normally, the extension and/or new agreement would be in place for the second quarter and targeted to be finalized either at the end of this year or the beginning of next year.

As far as the actual changes in a margin profile coming out of TPSCo, very sincerely, it probably isn't a smart thing to talk about it because if we were to say something that is on the lower end, it certainly is the target now for Panasonic.

If we say something on the upper end, it might not be something that's achievable.

So I'd really rather not give details until the point that the contract is completed.

And the second question is on Tacoma.

I was wondering if you could give any kind of update on that.

You had mentioned on the last earnings call, you had mentioned that the Nanjing Economic Development Zone (sic) [Nanjing Economic and Technological Development Zone] became majority shareholder and that there was a customer that had committed to 50% of the total capacity of the fab.

I was wondering if there's any details -- additional details on that.

Nothing of any market difference from that statement.

Indeed, Nanjing Development Zone is the general partner and the majority shareholder.

I think they have 57%.

They're putting together an investment consortium to fund the ongoing project.

I believe that there's been progress made there, and we're looking forward to the next step when that's finalized and moves along.

As far as the customer, what we stated was that there is a customer with whom we have an agreement that would use the majority of the volume we have from that factory, which is 50% of the factory.

But of course, the factory has to be funded and built out before that agreement can be triggered and then finalized, right.

So the build-out of the factory is still on track.

And when is that again if you can remind us?

I don't believe that the build-out of that factory is on track.

It's behind schedule.

The investment has not happened yet.

So that was a big part of Nanjing Development Zone becoming the general partner.

The previous general partner wasn't effective at putting together the investment consortium and getting it off the ground.

So I think there's good activity happening being driven through the NDZ and what it is targeted to be.

From our desires, it would have already happened.

But we're sitting there as a -- really, a minority partner and not someone who is investing a large amount of net cash.

So we are giving support as needed.

We're attending meetings, discussions as needed.

And the only formal update that we'll give is upon the construction -- well, the construction of the building, I think is complete.

But upon the ordering of the equipment and the movements at that point, I think that, that would be a significant milestone that we should announce and would be a public interest.

But as it sits now, there is -- in a material fashion, what we said at the last call was really the material changes that had been made since pending previous update.

And that material change is simply that the Nanjing Development Zone, mainly the Nanjing government, is now the general partner and the majority owner, which we see as a very good thing.

And on the -- a follow-up on the previous question with respect to the customer supply shortages of this highly specialized material.

I'm just trying to understand why that particular -- those particular customers did not have enough material.

Was it an understatement of demand?

Was it their own manufacturing issues?

And how much kind of visibility do they give you so that you would have some sense that this supply shortage would affect your high-margin business?

We have, from these customers, very long-term contracts.

And within pluses and minuses, they're very well met from both sides and they're very strong relationships.

If you look at the power discrete business as a whole, it has had a huge, huge growth, probably one of the biggest growth that there is within power discretes over the past few years.

Why would they have had a supply constraint?

That is really not for me necessarily to speak to.

That would be something to speak with individual customers about a -- I don't like overly speaking about customer business.

I think it's okay for me to speak about customer business as it impacts our revenue or as it impacts deltas to our revenue.

But as far as -- the fact is that had there been more starting material, there would have been significant different revenue levels in Q3 and Q4.

I mean, significant material, different revenue.

And last question from me.

So the decision to kind of walk away from low-margin RFSOI business, it had a 20-point impact negative or it's a 20-point margin business, I think you had said.

I didn't say that.

I said that the price reduction that was being requested was a 20-point reduction in price.

I didn't say anything about what the margin was.

Very well.

So a 20-point reduction in price as you walked away.

Do you feel that you have the mix of the RF business to where you want it now?

And going forward, we won't anticipate this kind of drastic changes in the mix of RF business from the past?

Seeing as Marco is here and this is specific on his business, I'd like to ask Marco to answer that.

And then I'll maybe add a few words of flavor afterwards.

Sure.

So I think yes, you're right in the sense that the strategy that we talked about at the beginning of 2017 of foregoing some of the low-end RFSOI business so that we can make room for silicon germanium, specifically in Fab 3, and some higher-end RFSOI in Fab 2 in Migdal Haemek -- sorry, Fab 3 being Newport Beach, Fab 2 being Migdal Haemek and the IS business, that has largely happened.

So the reductions have happened, you can see Newport Beach is back up to 85% utilization, which is close to being fully utilized or right around the target utilization despite the fact that we increased capacity overall as we stated in Newport Beach with silicon germanium high-end business.

So the reduction of that low-end SOI business has indeed occurred and the backfill with silicon germanium has occurred in terms of the wafers running in the fab and will generate the revenue in Q4 as the outs occur.

So to a large extent, yes, the answer is that we have gone through that transition.

Now still in terms of the ups and downs of the RF business, we're not immune completely from ups and downs in the market.

We talked in -- and Russell talked about what has happened with one customer, for example, in the second half of this year where we were faced with downside from this customer given whatever market conditions this customer was facing.

So we're not completely devoid of markets going up or down or particular customers going up or down, but the overall transformation of, let's say, the Newport Beach factory from the slow-end RFSOI business into silicon germanium is largely behind us.

The next question is from Mark Lipacis of Jefferies.

I apologize if I did not take the notes fast enough.

But Russell, did you quantify like the percentage of weakness that you had was from the tight input materials versus a lower demand or walking away from business?

No, I actually did not.

We stated that we did walk away from a substantial amount of business through not accepting a 20-point reduction.

Now I would like to add a clarification on that.

The reason of having walked away from it wasn't that it was necessarily a bad margin business to maintain.

It was that -- it was a bad margin business for the 2 factories that we were constrained to take that business in because we had higher margin parts that it was competing with.

And it's not that we aren't working with that customer or similar customers for similar business in other factories to where that margin is a very good margin.

So we did walk away from the business, and it did have a meaningful impact coming into this year.

That's why we had stated in 2017 that our RF growth compared to in the 30% to 40% range in previous years would be single digit.

Now it turned out to be substantially higher in 2017, but a lot of that was driven because of the big uptick within the infrastructure business and the actions that we had taken to start growing the capacity there.

So no, I didn't state that.

And I also didn't give the exact number for the supply constraint, the substrate constraint revenue in Q3 and Q4 other than it is within the multiple millions for each of the 2 quarters.

Okay.

Fair enough.

And then on the SiGE capacity in Newport Beach in San Antonio, the -- there is addition, there is an -- capacity addition in Newport Beach, there was -- I thought there was a planned capacity addition in San Antonio.

And then as you also plan to shift from some capacity of non-SiGE from Newport Beach into San Antonio just to free up more SiGE capacity.

So are all those finished and you're now utilized on all those dimensions for SiGE?

Or is there still more to come?

The San Antonio silicon germanium first platforms will be qualified in Q1.

The qual runs, if I'm not mistaken, Marco, correct me please, are planned to be processed during the first weeks of December, and then the qualification, the live testing burdens done by the customers, end of December, January.

So I think that those first platforms for the SiGE number of customers will be qualified in the first quarter, and the capacity in San Antonio will be ramping through Q1, Q2, Q3.

Nominally, reaching a full level of capacity in the Q3, Q4 time range.

The additional capacity growth in Newport Beach are tools, some of which have already been installed, some that is going to be installed in the beginning of next year, and that should target with the San Antonio an incremental 30% increase over the present level in Newport Beach.

But Marco is -- more or less, correct?

Yes.

Okay.

Okay.

Fair enough.

And then...

The other RF products going into San Antonio, that is there, and that is ramping nicely.

That's RFSOI in RF controllers.

Okay.

That's helpful.

And then it sounds like you -- it sounds like you're expecting the capacity utilization at Uozu to increase as early as Q1.

Is that a fair takeaway from your comments?

And if so, what -- could you provide some color on the kinds of products that you expect to ramp there?

Certainly.

We'll be, by customer forecast, having a strong ramp in RFSOI, so that should be taking off very strongly.

We'll be having continued ramp within large area image sensors, and that has already taken off.

That's where we're serving studio and medical applications right now.

And coming into the end of '19, 2020 time frame, we will be ramping the DSLR first-generation products of a very, very big customer.

We have mixed-signal capabilities are at 65-nanometer that customers have been excited about and starting to ramp the -- up to 16-volt BCD process where we mentioned a very, very big win.

That very big win will not be ramping in the 2019 time frame, although Protos will be going through at that time.

But it will be ramping in the '20, '21, '22 time frame and reaching full production there.

So I believe that, that's a pretty good summary of what's going on there for the 2019 year ramp.

I don't know, Avi, did you want to add any color to the image sensor activities there?

Yes.

And we have many projects for high-end image sensors running right now in Uozu.

It is well-known that especially the high-end image sensor projects takes time to ramp.

But once they are ramping, the longevity of such project is about -- between 8 to 12 years.

So what Russell is saying is absolutely correct.

We have started this activity.

For example, we said DSLR Tier 1 customer a year ago, but it takes time for that product to get into production.

Once it's in production, it would take it for many, many years.

That's helpful.

And I understand that it can take a while, but because it sounded like you had a big 1 or 2 customers ramping on image sensors for production.

Is that a fair takeaway?

Or did I misunderstand that?

It's more than that.

I talked only about the high-end cameras, but we have multiple projects, especially with a Tier 1 medical customer for 1 die per wafer for x-ray on 300-millimeter wafer.

It is still in design phase.

But again, once it's in production, it would last for many, many years because it's a medical product.

And of course, we have other products in the machine vision market where we have the best technology today in the world, and we expect this to ramp with multiple customers that are running prototypes now and would ramp into production next year.

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

Russell, you mentioned a couple of different engagements and wins that I want to ask you about.

First one, you talked about becoming a sole foundry source for an RF switch device.

A couple of questions in that.

Is there any way you can characterize what generation of RF we're talking about, 3G, 4G or could it be 5G?

And then any sense of how we can think about the potential size of it?

On the latter part, the size, as I had mentioned, in and of itself, that volume production, whether or not we do have other business with this customer, that opportunity itself would qualify it to be a Top 10 customer.

And I think that, that gives more or less a range of what the revenue is of that opportunity.

And it wouldn't even be #10, it would be higher up in the list.

But as far as the specific end market application and longevity, Marco?

Yes.

So the expectation is that there will be a family of products, of course.

So it will expand starting with 4G and continue onward to 5G and new generations as we go forward.

So it will be a family of products, not a single product or is -- even a single standard most likely.

Okay.

Perfect.

And as a follow-up to that and I think Russell, you'd mentioned this in your prepared remarks as well.

How big do you have to be to be a Top 10 customer in percentage of sales?

Actually, I just looked to that a few minutes ago.

Top 10 was...

$20 million.

I'd say on the order of about $25 million or so.

$25 million?

Okay.

That's helpful.

Perfect.

Maybe just one more question since Marco is on the line with us.

You mentioned both in your press release and your prepared remarks about some excitement in the silicon photonics market.

I guess a couple of questions there.

How wide is the customer funnel for that one?

And is that a product portfolio that should ramp before the end of next year?

Or is it more of a 2020 story?

So to answer the first question, we have 2 major projects right now, so 2 lead customers.

But in terms of the funnel, we put out a PDK for our initial technology.

And we have about 10 to 12 customers that we're working with on that initial PDK in addition to the 2 lead customers.

And in terms of the ramp, there is some very low volume production even happening today on the first of the 2 projects that I mentioned.

But most of the ramp, we would expect to slowly ramp in 2019, not be particularly significant in 2019, but start to become more significant in 2020 and beyond.

Okay.

And Marco, can you say whether these applications are more datacom or telecom focused?

So they're datacom.

So it's mainly data-center-driven business.

Okay.

Fair enough.

One last question from me, I'll jump out of line.

I know you typically don't talk more than a quarter in advance.

But is there any reason not to expect a "seasonally normal first quarter of the year?" Obviously, it's hard to predict with inventory changes in the industry.

But any comments you want to make, Russell, Oren, on thought process there?

As we see things right now, we would expect normal seasonality, nothing beyond normal seasonality.

And I think that's the proper expectation.

The next question is from David Duley of Steelhead Securities.

I had a couple.

Just going forward, as far as CapEx intensity goes when you're adding capacity at any of your factories.

How much do you think you need to spend to add $1 of revenue?

What's the CapEx to revenue ratio?

Usually, for -- averagely, for 8-inch costs us, today, I believe, like $8 million to increase each 1,000 wafer per month of capacity.

So if you go on the average prices of such wafer, if it's a standard process, not too long flow, not complicated like 5G, just normal, it's maybe 1,000 and maybe $700,000 revenue a year, so maybe $350,000 a month, so maybe $4.2 million a year.

So that's from an $8 million investment.

So this will -- our why at -- it's a 2 year, right, because you invested $8 million and you got $4.2 million net.

Now this is a...

But wait a second.

So $8 million of CapEx gets you $4 million of revenue annually?

Excuse me?

Could you just run -- I didn't quite understand what you just said.

$8 million of CapEx gets you $4 million of revenue?

No.

$8 million gets me $8.4 million of revenue, which gets me $4.2 million net margins per year.

So it appears to be almost 2-year ROI, but this because I used is just a number, $700 per wafer, for a simple wafer, 0.18, not 5G, not anything special, not image sensor.

So if you use 5G image sensor obviously, you'll get a quicker return, like 1 year or maybe even lower.

Okay.

And then you -- in the past, you'd mentioned the growth rate of the non-Panasonic, non-Maxim business on a year-over-year basis or year-to-date basis.

Could you share any numbers with us on that?

For this year?

Year-to-date, let's see.

I'm truly not prepared for the question.

I'll be very happy to follow up with you afterwards and give you the answer.

I wouldn't -- I mean, we could do it in a few minutes, but I won't be able to get it at the moment.

Okay.

And then final one for me is also along the same line.

In the past, you've given us kind of a breakout of revenue amongst the 4 big buckets.

Could you share that information with us, perhaps year-to-date?

Or if you meet your guidance for Q4, how it might look for the year?

We give that, you're correct, once a year at the end of year.

And we'll do the same thing this year.

We break it down into end applications, and that does take a bit of work because you have customers that are installing into multiple applications.

So we go through and dig into that what the different parts are going into whatever end sector.

And you can expect that when we release Q4, we'll give a breakdown of all of the yearly revenue into each of the 4 major buckets.

Okay.

Final one from me is you talked about, I think, an assortment of things that might help you fill up your 300-millimeter fab in Japan.

If your initiatives are successful as you expect, what would you think the utilization rate of your Japanese factory network will be exiting 2019?

And I think you mentioned it was 55% now.

And I think that utilization rate has been pretty steady for some time.

So I'm just kind of trying to figure out what the new opportunities are going to do to the utilization rate next year.

So as we stated, we should be on the upper end of utilization at the 300-millimeter factory, the Uozu factory.

We should be running a relatively high utilization rate in the Tonami factory.

The Arai factory is a smaller copper back-end factory that we have not built a lot of utilization in nor do we necessarily foresee that we'll be building a lot of utilization in the short term or midterm.

But both the Tonami, the high-volume 0.18 factory and the Uozu factory should both be running at high utilization rates.

And the Uozu factory, we would expect, again, that by the second half of the year to need to make decisions, providing that forecast play out the way that we believe that they will to increase capacity there.

And we'll make everyone abreast of that.

I do want to add one clarification though.

On any given quarter, you cannot take the utilization rate that we give as an indication of the third-party utilization for the reason that on any given quarter, Panasonic could be up or down off of the previous quarter.

And on a steady state, there's a contract with Panasonic and their demand has been somewhat stable, but there's a pricing table with them.

So the difficulty, and we really don't know the best way to go about it other than give a blended utilization, is that to separate what the third-party utilization is, we would be giving the exact utilizations of Panasonic in those factories.

And we don't think that that's a fair thing to do to any given customer.

The next question is from Quang Le of Credit Suisse.

Russell, you said that, obviously, you would provide us later on the organic growth for -- excluding Maxim and Panasonic.

But my question is for 2019, are you still targeting the mid-single-digit organic growth for next year?

We've not given any target.

But nominally, our organic growth target is really higher than mid-single digit.

So as far as the target, that would not be our target.

I'm not sure honestly what the number is for 2018, and we'll go through that and provide that number.

But our target and our performance in past years on organic growth has been in the 20s and cannot be maintained every year or year-over-year, I'm not sure.

But I think it's a reasonable target that we would have and have had.

And in terms of the margin improvement in Q4, I'm not sure if you talked about it already, but can you quantify the margin improvement?

And is it better than your model of 50% to 55% flow from revenue to profit?

It should be slightly better than 50% or not slightly because Q4, too, is in -- first, we have the $7 million reduction in financing expenses that was dealt in Q4, so it's $2 million a quarter that we disclosed.

And also, we have the effect that Russell pointed out, the SiGe improved mix as compared to other lower margin.

So Russell mentioned $20 million -- more than $20 million net margin a year improvement, which is $5 million a quarter.

So together, it should be $7 million just on that.

And of course, if we get to that, the effect that we are forecasting per the guidance, $17 million higher revenue, right, from, as we said, $323 million to $340 million, which also will bring some incremental.

So I -- of course, we are not entering into exact calculation and guidance for net profit, but you can -- you should consider those 3 factors.

So very positive.

Got it.

And my final question would be -- apologies again if you already answered it.

But regarding your softness across your business units, is that softness that you see more in RF or power?

And if it's RF, do you see it more in mobile or infrastructure?

And do you see this improve by Q1 next year?

So against what we would have targeted for Q4, stated in release -- hopefully, stated in a clear enough fashion, that we see it somewhat equally distributed among the total demand, meaning that there's no single area that we see great softness.

We just see a general reduction.

And again, it's not a huge reduction, but we do see a general reduction against that which we had targeted.

As far as the one area that we don't see any softness in, presently, it's infrastructure.

The silicon germanium is fully loaded with a remaining queue, so that, I don't see any softness in Q4.

As stated, the -- I had mentioned that the Q4 silicon germanium revenue will hit at $200 million, so it's above a $50 million quarterly run rate.

And that's a pretty good achievement to have had within SiGe with a wafer shipment increase of about 70% from the first quarter through the fourth quarter according to the whip in line.

Now nominally, everything that is whipped and planned to get out by the end of the quarter should get out bar any singularity event in the fab.

So I think that's very strong.

But if you look at it, then we'd say that the other products that we're doing have seen somewhere in the order of $20 million, $25 million softness, and that's across everything.

It's not that I can pick out any single area where it's softer or stronger than the others.

It's pretty much peanut buttered across the different units so no single unit is hit very hard.

But that's what the composite is.

I see.

And can I just clarify, did you say that Q4 silicon germanium revenue is over $200 million?

Am I correct?

No.

Run rate.

I said Q4 is a run rate.

Run rate.

Sorry.

Yes.

Got it.

Got it.

Got it.

But it's basically that the business is going to be a very, very solid nice business at very good margins, adding value to customers that they have continued.

And I think that's one of the -- really, the exciting things about the silicon germanium, the activities going on is that, as we mentioned, we enjoy above a 60% market share within the optical market that we're serving.

And those customers, to a good extent, are already experimenting with or have done Protos on the newest version that we have, the -- it's called H5 plus, giving figures of merit well beyond what was, otherwise, the industry best.

And so there, the continuation, the activities, it's -- that's quite special.

And can I just clarify, the 70% that you mentioned, what this number refers to?

Is it the...

Wafer shipments.

Wafer shipments.

It's -- that's the increase in wafer shipments?

Or...

Yes.

And this is only for silicon germanium, right?

This is the increase versus Q1?

Q1 versus Q4 due to the capacity increase that we put in place.

It should from the whip in line, again, they haven't all shipped yet.

So barring any singular event in the line, we would expect by the work in progress, everything that's going to ship in Q4 for silicon germanium has already been started.

It's already well on its way.

So providing that there's no -- nothing of any abruption, which we don't expect, we should have shipped by the end of Q4 in silicon wafers, an increase of 70% versus what we shipped in Q1.

And the only reason to say that is we had talked about the increasing capacity and that capacity is fully utilized.

There are no further questions at this time.

Mr. Ellwanger, would you like to make your concluding statement?

Yes.

Thank you very much.

Truly, I thank everyone for your interest.

That's a very good Q&A, enjoyed it very much, of good questions.

I thank you for that.

I thank everyone that's been involved in the company, for your involvement.

I truly thank our customers for their trust in the company.

I thank our investors for their faith in us.

And as stated, the activities in the company are, I believe, stronger than they've ever been, creating incredible growth engines for the future.

Whatever small market pullback or extended market pullback there might be that we're getting into, the underlying fundamentals of the company are extremely strong.

The balance sheet is strong.

The margin increases that we're looking forward to are real, and the customer funnel based on really high-value capabilities is very real.

So as with everything, time is the teller of all truth.

We thank those that have been with us over the past years, and we look forward to strong relationships over the years to come.

In the very short term, we will be presenting at the Credit Suisse 22nd Annual Technology Media and Telecom Conference in Scottsdale, Arizona, November 26 through 29.

Dr. Racanelli will attend and be available for one-on-one meetings on November 28.

Thank you very, very much.

Bye-bye.

Thank you.

Thank you.

This concludes the TowerJazz Third Quarter 2018 Results Conference Call.

Thank you for your participation.

You may go ahead and disconnect.