Emerging Growth Conference
C.E. Unterberg, Towbin
07.10.2007 New York
Tower Semiconductor
Russell Ellwanger, CEO
To begin with, the company was founded, Tower Semiconductor, in 1993. Its located in Migdal Haemek in Israel. There’s two factories. There’s a six inch factory that was actually built in 1984 for National Semiconductor Israel. Then again in 1993 became Tower Semiconductor.
The six inch factory makes technologies from .35 microns to 1 micron. Then there is an 8 inch factory that started shipping wafers in 2003 and presently has the capacity of about 24,000 wafer starts per month. That is a .18, .16, and .13 micron factory with a 90 nanometer roadmap. So according to the roadmap we should have 90 nm in the second half of 2008.
I had joined the company in the second quarter of 2005. Oren Shirazi had shortly before that been appointed CFO and we instituted several differences in the company shortly thereafter. At the time that I came to the company, the revenue level was about $19M per quarter and the EBITDA (earnings before interest, taxes, depreciation and amortization) was -$10M. So we established two major financial goals. We announced them to the public as our targets and that was to be positive EBITDA in Q4 2005, positive cash flow from operations in Q4 2006.
We restructured into product lines making us more customer-centric and over this period time we did grow the revenue by close to a factor of 3 and we did achieve the goals that we set of the positive EBITDA which we achieved as $3M in Q4 2005 and we achieved the cash from operations of $7M in Q4 2006.
Growth was very positive for us, that we look back on with good pleasure in that the incremental revenue vs the incremental EBITDA was at about a 74% EBITDA margin. Even in the foundry business is you are selling dollars for $.99 its not necessarily easy to grow the business, but to grow at that kind of incremental margin, the only way to do that is if you are adding value to your customers. We are very pleased with our growth at that point.
The GMs have also improved nicely during this period of time, where we started at negative 26% and finished at a positive 42%. This is non-GAAP so this is without depreciation.
For 2006 Tower YoY revenue growth out-grew the industry by about a factor of 5. Was at the top of the class of foundries as far as growth with a 100% YoY growth. What drove this growth was a major drive to bring on a different taxonomy of customers to where we grew large customers into the company.
So in Q1, 2005 you can see here we have only 2 customers that we were shipping more than a 1000 wafers a month to. In Q4 2006 we had increase that to 12 customers that were active and that we were shipping more than a 1000 wafers a month to. The plan at the end of this year is to have increased that to 16 customers that we are shipping more than a 1000 wafers a month to.
The different products that we make are the CMOS image sensors. Cameras for cell phones where we ship VGA 1.3 up to 3 megapixel cell phone cameras. High end film digital cameras. These are 14 megapixel type cameras. Sensors for automotive as well as, [pause] We are very strong, in fact I think that we are the market leader in supplying digital sensors CMOS for dental xray applications.
We have grown quite nicely within the mixed signal CMOS area. In the wireless LAN, Atheros is a very big customer for Tower. Atheros is the leader within the wireless LAN segment. RFID, we are the sole supplier of the gen 2 chip for the leader of gen1 RFID, Alien technologies. And growing strong within the power management portfolio of high voltage devices.
And then our core CMOS products as well have done very nicely for the company. We are the sole suppliers to SanDisk for all the secure storage controllers. Consumer products, image processing, the demodular chips for example for Zoran for DTV. Very strong in PC accessories and peripherals and home electronics. So those are the cores for the products that we make that has driven the revenue growth.
In addition if you look at foundries. The foundry model initially was fabless companies. These are several of Tower’s customers. Foundries such as tower or others. We make the wafers off of their designs, process them and then they are shipped for assembly and test.
However over the past half decade, very strongly and it gets stronger continually, the IDMs themselves, the integrated device makers, that have their own design centers and their own designs and also their own internal fabs, are driving very strong towards asset-lite models.
In fact, week by week there are more announcements of more and more of the IDMs divesting themselves of assets. (missed word) just made a big press announcement about that as well.
So these companies themselves are driving continually less assets and driving only internally their core products and using more and more foundries. This is a major focus of Tower to pick up this IDM business and one that we have been successful at.
As far as IDM transfers, in many instances the IP that is owned is the process flow itself. It is not the design, but the process flow to get to the end application.
Israel is a low-cost manufacturing region. Most of the IDMs have been manufacturing within the USA and Europe. And it gives you a substantially lower cost. So transferring to Israel allows us to make the wafer and sell it for reasonable margins for Tower and at less than some of the internal manufacturing costs for some of the IDMs.
For them one of the big benefits of coming to Israel is that it is an area that has very strong IP security. It is a strong security culture. There are several regions of the world where to transfer a process flow it will be very difficult not to create an industry along with transferring your process.
This is an area that we have grown in and done well in. Right now in fab 1 which is a 24 year old fab, we are in a 10 year contract with Vishay Siliconics on an IDM transfer. We are less than two years into the contract on 45% of the capacity of that fab has been spoken for for the 10 years.
As well we entered into a multi-year contract with international Rectifier, which by contract will start revenue in Q3, Q4 of this year.
As well with SanDisk, we have a multi-year contract not an IDM (?) contract.
Moving off the product line what we did in 2006, what does 2007 look like? If we look here in a non-GAAP presentation, Q1 2005 we were at -52% operating margins. Q1 2006 we saw a slight improvement to 14% operating margins. If we look at the Q1 2007 we increased that to 21%. Q1 2007 is very significant as well.
We showed that we were a leader in foundry growth for 2006. If we look at the top 4 foundries in the world that make up somewhere close to 80% of the worldwide foundry volume. The Q1 YoY weighted average is down about 10%. Tower was up 55% YoY. So a very good start as far as our growth in 2007.
Our guidance for Q2. We had guided to be up slightly more than 32% YoY where the weighted foundry average was down about 13% on guidance.
What are we focussed on now? What do we look like? Where are we going? In Q4 2006, you look here and you see the dark blue of fab 1, the larger than 180 nm products, the .35 and above. Fab 1 at about 16000 wafer starts. Fab 2, 180 nm, was at 15,000 wafer starts.
Our high utilization, the capability here was to generate about $220M revenue. If you look at the revenue that we have shown then at this point, the $51M, you would see that we are running at a very high utilization. Being very effective against our model. That would drive $30M of EBITDA, $5M cash from operations.
However the investment for this 15,000 at an annual deprecation of about $140M and the delta cash from that investment was a negative $15M because of the interest paid for the investment. So you can see that as far as the RLI, it wasn’t sitting there off of the 15000 investment.
Off of the customers that we have developed, and the growth, we raised the money in 2006 and recently have just completed a ramp for 24,000 wafer starts in fab 2. 20,000 .18m, 4000 .13. The 4000, .13 is taken up in a multiyear contract by SanDisk.
At this level, the potential of the company is to hit about $320M, the EBITDA to be about $100M with about $75M cash from operations. This capital expenditure per 1000 wafers was 16M. So substantially better being built upon the infrastructure that was initially set. The incremental depreciation per year, $30M. The incremental potential cash generation is $70M. So the RLI was here for this investment for anyone that invested in it. I think that we are on track to fully utilizing the capacity.
Certainly we are evaluating an option to increase the capacity of fab 2 from 24,000 to 36,000. And to do this in a very very cost effective manner where we would be buying used equipment that is focussed at 130 and 90 nm. We have several customers that are with us very strongly on the 90 nm roadmap. Additional customers that we are working on very closely with to win major IDM deals for 130 nm..
This investment that we would be targeting right now to get about 12000 wafer starts at about $9M per thousand, for about a $100M investment. The incremental depreciation would be on the order of about $24M per year. The incremental potential cash $80M. So the depreciation vs the incremental cash, we have here a factor of about 3.5. The duration of payback of the investment would be about 18 months.
The company then moves from being a company of a $100M EBITDA to a company of $185M, but very significantly moves from being a company of somewhere a few million shy of $20M per quarter cash flow from operations to being a company that is on the order of $40M per quarter cash flow from operations upon filling the capacity.
So I think that the company has done quite well, in its movement in utilizing the 15 and driving upwards to show the utilization of the 24 and the 36 is now what we are looking at, what we are focussing on.
[86] We recently last month raised $40M in long term bonds in the public market in Israel towards this investment. And we’ll be targeting completing the investment at also a longer term debt vehicle so not off of an equity vehicle.
But this is right now the direction that we have within the company. To continue to increase capacity, but to increase it at the advanced technology nodes and I think again the potential here is very strong. Moving from a $300M revenue company on an annual basis to the potential of being upwards of 450 to 460 and generating on generating on the order of $40M cash per quarter.
To summarize, 2006 was a really a very very strong breakout year for Tower. We achieved consecutive record sales (?). We achieve positive EBITDA in all quarters and did achieve a positive cash from operations in Q4, 2006, our announced target.
2007 is poised to be of a similar order to maintain that momentum. We forecast positive cash from operations throughout the year. We guided Q2 to be the 8th consecutive revenue growth quarter for the company.
We are in the final stage of a 60% capacity expansion in fab 2, that’s going from the 15,000 to the 24,000 wafer starts which obviously has the incremental revenue and cash impact that I have showed in the previous slide.
We have multiple very strong customers, really some great customers in the pipeline with various new products and designs. Within that we have up to $90M yearly revenue from several signed multi-year supply agreements. What is commonly referred to as take-or-pay type agreements.
We are right now pursuing several opportunities for cost effective 130nm/ 90 nm capacity expansion. Thank you very much. I wanted to end the presentation just to highlight to you the different customers that we have. Those that have been press released I can talk about.
If we look at the Tower customer’s within fab1, I have mentioned Vishay Siliconics and ON semi. So these are the two large customers for fab1.
In fab2 SanDisk, a very large customer for us.
Zoran a very customer. Atheros a very large customer. Winbond, a very strong customer and you will see International Rectifier also coming into the top five type customers as we ramp according to the contract that we have. Alien is a customer that is very interesting as far as the RFID space. The gen2 RFID is something that is really forecasted within several years to get upwards into the billion of dollars as far as the market. Alien has a very unique technology for the packaging of the chip that gives them a very definitive advantage against their competition. We are the sole suppliers to them for the gen2.
Another very interesting customer that we have is Schick Technologies. They are the leader in dental xray. We are the sole suppliers to them for the dental xray chip. Others here too. I don’t want to understate any of them.
Micronics (?) too is a top 5 customer for fab2. Thank you very much for your time and I would be very happy to try to answer any questions you might have. Thank you.
[tape 110] Q&A
Q: There have been a lot of press articles, in fact more so than company prs per se would you care to comment on some of them, or all of them?
Ellwanger: Can you be more specific? [laughs]
Q: In regards to fab equipment purchases from some sources.
Ellwanger: I mentioned that we are looking at growing our capacity in this area. To increase our very cost effective basics (?) 130 nm and enter into the 90nm processing space. AMD Dresden.
What has been published in the press is that we will be acquiring tools from AMD Dresden. That I will not confirm or deny. The AMD factory is 90 nm /130 nm microprocessors. So it is advanced logic tools. It is basically converting to 300mm and has bid out all of the tools in that factory that were 200mm. So that was what you were referring to there, but we have not press released nor confirmed or denied that we have something going on with AMD Dresden.
What we do say as per the bullet in the conclusion, we are pursuing several opportunities for cost effective 130 nm/ 90nm capacity expansion. That we are going after very strongly.
Q: Why stock is weak or stock is strong?
A: It is very difficult to comment on that whether the stock is weak or the stock is strong. There have been a few time that we have had announcements that I thought the stock would respond very nicely. Sometimes is does. Sometimes it hasn’t.
The fact as a whole though, I think has actually done quite well over this period of time. The entry price when I came to the company was somewhere about $1.15, in that range. We had got up somewhere about $2.16, $2.18. During this period of time, we did raise a lot of money in the company. So although on a shares outstanding basis, we have 120M shares, on a fully diluted basis there’s more shares out than the 120. One of the movements in the stock was the fact that we had converted $160M of bank debt into $80M of equity. With the two banks that we deal with in Israel, they went on a 2 for 1 conversion ratio. Again from a debt of about $500M, $160M was reduced at $80M of equity. That was done in the capital markets. So that is a pure equity vehicle, but it doesn’t show up as shares outstanding. I think it is the dilution itself that took the stock down. At this point as I mentioned, we are looking at funding our growth over some longer term debt vehicles that will be in a good position to serve than to do further dilution. What do I think? I think the stock has actually done very well if you consider the fully diluted basis. At this point our drive is not to dilute further.
Q: As an overall industry, the growth rate has been down, why has ours gone up, what is different about the company? Why do we see that happening?
A: To start with we have certain specialized technologies that we have been involved with with some customers in some areas. For example if you look at the dental xray, that is not cyclic at all. So we have a very broad width of product offerings. We’re not overly leveraged in the PC space and the PC space is the most cyclic space that there is. So that’s number one.
Number 2 As we have brought on these large customers and I think this is maybe a more significant part of it. As we have brought on these major customers we are not the major market share owner from those customers. As we have shown our capabilities as we have delivered, very on time performance as we’ve shown very good time to market of new products we have been able to grow the market share with those customers and to outpace the market growth itself by being able to go from a 5% to a 20, 30, 40, 50% market share with a large customer. That’s really been the way that we have been able to outpace the market. It comes into an economy of a smaller number. If you own 45% of the market, and drive 100% growth, the market itself has to grow immensely. If you are coming in and you have less than a 1% share of the market, but you are showing very good technologies, good capabilities, your platform is very strong, your service is very strong, your ability to grow your market share can strongly out pace the market. So that has been, I think, the major fuel of our growth, and, I believe, it will stay the fuel of our growth for the next years. Does that answer your question?
Q: [Inaudible]
A: Many reasons I would say. In some cases, its not that we are taking business away from competitors, its that we are developing businesses. That comes within the pure specialty area. As far as the large customers, though, that we’ve developed, that’s obviously taking business away from competition. You don’t develop new markets that are more than 1000 wafers per month.
In summary, many customers really do want to have dual sources. Very few people want to be captive to any one supplier. Tower’s ability to produce very quickly, to port other people’s designs, to make them work quickly within our flow, is very very strong. So it always becomes a question of cycle time. Time to market. If someone is going to invest in you, which it always is an investment to start a new foundry, how quickly will they see a return on that.
Tower has an extremely strong design center. As I mentioned in my presentation, our two factories are in Migdal Haemek which is in the north of Israel by the Galilee. We have a strong design center in Netanya which is by Tel Aviv. Our design capabilities being very strong allow us to port a customer’s designs and make it work with our IPs and libraries and the Tower flow with transparent output for the customer. So I think that that’s been the way that we have been able to show value to customers and gain their confidence in Tower.
[inaudible question?]
There are some foundries that by virtue of their name and who they are, they will command a higher price. I’d say that for the Fab 1 type area we are probably the same as most anybody within those markets with the exception in fab1 a lot of our products are very high specialty products. The xray, all the stitched image sensor type activities that we have which command a very very high prices. So if you look at [pause] At any time that you have a technology node there will always be a very rapid decrease in selling price. Then it gets flat. And that asymptote that it reaches is really the manufacturing cost (?). Now at any given time off of that technology node you can see an increase in the ASP if you start to put building blocks on top of the node that’s specialized. Again that comes into xray sensors. It comes into some mixed signal applications. Very strong analog power management devices. So in fab one we have been able to do that very strongly.
In fab two we have as well certain specialty products: stitched products within the image sensor that are demanding very very high ASPs. Now if you were to talk about a very large volume commodity product that doesn’t have anything special in it, I’d say that Tower is very similar to most of the foundries and probably lower pricing than the top foundry.
SanDisk etc 