Tower Semiconductor: Not At The Top Of The Tower Yet

| About: Tower Semiconductor (TSEM)
This article is now exclusive for PRO subscribers.


How did I get to own Tower Semiconductor?

Tower Semiconductor: Acquiring unprofitable foundries as a profitable business? It seems so when you know how to turn things around.

What to expect from Tower Semiconductor next and company valuation.

Investment thesis

Despite the run-up in recent months, Tower Semiconductor (NASDAQ: TSEM) shares have room to grow further. Sales of Tower's products have steadily increased in the past quarters, improving Tower's Fabs utilization rates. As a result, margins have expanded significantly, and the company has finally posted profits for FY2016. Management has demonstrated some track record of opportunistic M&A to fuel growth. Based on Tower's earnings power and relative undervaluation, I assign an "outperform" rating on the stock, with a new fair value range of $27-$34.

A recap of my last investing year

There is an article I am proud to have presented to my followers on Seeking Alpha last year: Summer Tech Picks Trading At Reasonable Prices. I am linking it here just for those who would like to get some basic knowledge about Tower before reading on.

The article was a rather unpretentious piece of research, but I think it had the merit to present some pretty actionable ideas in a simple fashion. Among the ideas, I disclosed my long position in TSEM, which was quite a fresh one at that time. As I write to organize investing ideas and strategies in my mind, most of the research I publish tends to be about stocks I own or follow because I am considering ownership.

2016 may not have been my best investing year: notably, my positions in Buckle (NYSE: BKE) and other retailers failed to live up to expectations. It has been a terrible time to be stuck in retail and apparel stocks, and I have to acknowledge it seems I also ended up punished for my distrust of the good old suggestion about the industry: never trust P/E and DCF but only sales comps and mall traffic. On a side note, the metrics are still horrible.

Fortunately, my overall investing results were fine, and this thanks to the good performance in sectors such as technology. In my August article, I introduced Syntel Inc. (NASDAQ: SYNT) Gentherm Incorporated (NASDAQ: THRM), and Tower, three companies I was following closely at the time.

Deeper research led me to shy away from Syntel. The stock performed poorly, and its peers were no better. I went long on THRM in October when it fell below $30, which was also the recommendation in the article. After re-checking the idea, I found it still valid. THRM is now trading above $38: it turned out to be a nice swing trade for me. The real deal on the long side, however, was Tower Semiconductor, which is still among my holdings.

Tower Semiconductor: a story of opportunistic M&A

Towerjazz, despite being a recent story of growth and success, largely remains an underfollowed foundry in the semiconductor space. While industry leader Taiwan Semiconductor Manufacturing (NYSE: TSM) enjoys quite a good coverage also here on Seeking Alpha, nobody seems to care too much about Tower. From an investor perspective, that's truly a pity because, while TSM shares reported a solid +14% since last August, TSEM is up 66%.

Fig.1: Comparison between TSM and TSEM performance, August 2016/March 2017, Google Finance.

It is indeed true that Tower is nowhere near TSM in size and technological edge. However, as the demand for semiconductors burst out in all directions during the last years, TSEM management was smart in taking the necessary steps to attain a leadership position in specialty analog applications (RF, HPA, CMOS/Imaging sensors) and specifically address the target market. With over 1 billion sales and 2 billion market cap in 2016, the company is arguably no small fish.

Fig.2: Foundry landscape, company presentation March 2017.

(A note to investors interested in international opportunities: These days I am following with great interest the upcoming quotation of X-FAB Semiconductor Foundries AG on the Euronext market).

How the company grew this much? After starting out in Israel, Tower managed to grow through M&A, taking advantage of industry consolidation around the world. It was not a smooth process, and Tower's initial acquisition, a Micron Technology's (NASDAQ: MU) facility in Nishiwaki, Japan, was closed in 2014 after just 3 years of marriage.

Nevertheless, the confidence of CEO Russell Ellwanger in the strategy led to similar experiments first with Panasonic (OTCPK:PCRFY) and then Maxim (NASDAQ: MXIM). Tower's BM was relatively simple: Acquire the unprofitable foundries and turn them around increasing their utilization rates. To achieve the improvement, Tower maintained tight relationships (fixed supply contracts) with the previous owners, while finding the necessary new customers.

Fig.3: Tower's actual global presence, company presentation March 2017.

Things seem to have started to turn out quite well. TSEM revenues kept increasing steadily over the previous quarters, and last year's total sales ended up 30% higher than 2015. Together with record revenues, the achieved efficiencies led the company to post record operating and net profits.

Fig.4: Comparison of TSEM Revenues and Operating Margin in 2015 vs. 2016, author's work.

The company's guideline is to achieve a Fab run-rate of 85%. With operations in Israel already running at 78% and 87% utilization and Newport Beach at 84%, there is some room for growth left in Japan (TPSCo running at ~50% utilization rate) and in Fab8 (acquisition from MXIM in San Antonio) which is, however, ramping up fast.

Management is expecting a revenue increase in Fab8 of about 30% in 2017 over the Maxim base contract. To increase production capacity, Tower seems to be getting ready for future surprises. Here is an excerpt of what CEO Ellwanger replied to a question about company strategies in the last earnings call Q&A session.

"IDM consolidation does allow available factories due to the consolidation and under-utilization that some factories become available within the market. And at this point, we have reached the level of customer range that most of the consolidations being done, we have a relationship with either both of the IDMs or with one or the other of the IDMs. So the ability to get into such a deal at a very, very early stage is very real."

Another option already on the table is a Greenfield site in China with a local partner, possibly the State itself. The Chinese government is building up Fab capacity and to do so is luring to the country outside expertise. TSEM would be one of the last large players to move in, provided that of the top 5 foundries, TSMC, GlobalFoundries, UMC (NYSE: UMC) and Powerchip already made investments of some kind in the country, and SMIC (NYSE: SMI) is a Chinese company.

The costs for a Greenfield site are huge, and the Fab 2 experience probably left some scars in the management, together with a large NOL that is now providing an interesting tax shield asset to the company. So, it shouldn't come as a surprise that Mr. Ellwanger clarified that Tower has no interest in a Greenfield site by itself but rather a contribution in a Chinese JV that would be largely an in-kind one, providing IP, equipment, and technology.

Outlook for 2017 and beyond looks good

As described above, management seems to have clear ideas on how to expand further its production capacity. The company has good relationships with all major IDM fabless manufacturers, and it is forecasting strong demand for its products throughout the current year.

[In 2017] according to customer forecast, we expect growth throughout the year with each quarter being measurably higher than the corresponding quarter in 2016.

The growth trajectory will probably continue into the first half of 2018 on the condition that, since current Fabs capacity will max out in the current year, Tower will add the required extra capacity. As I said, new capacity can come either from the Chinese JV or through acquiring another used and depreciated Fab, purchased at a low price and with some revenue stream already attached to it.

Management is not-so-subtly guiding a YoY revenue growth of 19% which, netted the effects of Fab8 acquisition from Maxim, is more or less in-line with last year's organic growth. Below management's comments from last earnings call:

We have 19% forecasted growth year-over-year as compared to 30% in the last year, but last year you should remember that we acquired the San Antonio Fab, so it's the Maxim incremental. […] we're actually keeping the same momentum and the same percentage of growth this year as well.

Elements of risk involved in the assessment of Tower performance and fair value

I have identified two elements of uncertainty in TSEM FY2016 numbers.

1. Considerable financial gains (non-operating profits).

The company posted significant other operating costs in 2015 (over $100 million), which were in part related to the San Antonio Fab acquisition. The situation reversed in the current year, and Tower posted significant gains which boosted its EBIT to $220 million. Operating profits in 2016 came in at "just" $174 million. The company has well-detailed everything in the earnings call, so nothing's hidden here, but these gains are likely to be one-off and therefore normalized results should be adjusted accordingly.

2. Major shares dilution.

The company suffered major losses in its Fab 2 Greenfield project that, as I briefly mentioned, also resulted in the significant NOL Tower reports in its balance sheet. The company's survival came, however, at the cost of significant shares dilution. Most debt recently disappeared converted into shares. The company used shares as a means of payment as late as 2015 for its acquisition of the Maxim facility, in spite of a pretty good cash position.

It is hard to come up with a precise shares count as of December 2016 using available information, but outstanding shares were below 90 million at the end of 2015. They could be near 100 million at the end of 2016 based on the latest data ($48 million net income against $0.52 diluted EPS reported for Q4 2016).

Regardless of the dilution issue, the company posted diluted EPS of $2.09 for FY2016 compared to a net loss in FY2015. I will now assess in detail how the spectacular growth of 2016 translates into value for Tower shareholders.

Company valuation

Last year I wrote that the price target for TSEM could have been within the $20-$25 range. While the stock moved within this range and is now trading around $23, considering how the story continued to play out, an upward revision seems appropriate. To assign a new target for TSEM, I worked on an EPV valuation but also checked on TSEM's relative valuation to sustain my assumptions.

For the relative appraisal, I used the same peers and multiples considered by analyst Lisa Thompson in her SA article: TowerJazz Is Growing While The Rest Of The Foundry Industry Stagnates. I agree they seem to be the most appropriate candidates for the purpose. My work in this regards is just an updated comparison to understand what Mr. Market thinks today about TSEM.

Valuation Based on the EV/EBITDA multiple

Fig.5: Valuation using EV/EBITDA multiple, author's work.

Valuation Based on P/E multiple

Fig.6: Valuation using P/E multiple, author's work.

The range provided for TSEM using such multiple methods is, therefore, $27.3-$36.7. The Market is still assigning Tower lower multiples than peers, and it's therefore relatively cheap. I argue such a discount is at least partially unwarranted and related mostly to the company's recent history of poor profitability and shareholders' disclosure.

Surely some of the P/E understatement could be reduced after addressing the concerns I reported. A way to account for such issues is to normalize earnings adjusting for one-time items and full share dilution. The results are reported in the reconciliation below, assuming a conservative 100 million total shares outstanding.

Fig.7: 2016 Net income adjusted for one-off items, resulting adjusted P/E and valuation, author's work

EPV Valuation

To derive Tower's EPV, I started out with its 2016 operating income. I did not use average margins being the company pretty much of a turnaround story. I also did not proceed with any adjustment except for a 20% add-back of SG&A expenses which I feel like could be the right measure of Tower's advantage over a new entrant due to established customer relationships and captive supply contracts. For completeness, I clarify that this add-back augments the final EPV result of around ~$2.8 per share.

To be conservative, I still subtracted the average of the five previous years' "other expenses/revenues" net figure, and I normalized the net income applying a regular tax rate of 25% disregarding the NOL asset, which is fair since I did not use historical margins. While I did not add back anything for D&A, I did find the level annually set aside by TSEM to be at least as conservative as its peers. I applied a diluted share count of 100 million resulting in a current EPV value of $13.5 per share.

When considering Greenwald's original EPV methodology, TSEM enjoys, however, a very high growth multiple of 2.5x given the high return on capital (ROC/WACC > 2) and the significant growth prospects for the industry and the company in the medium to long term (GROWTH/WACC > 0.75).

The resulting full EPV value accounting for growth is $33.7 per share. While I believe, considering the above assumptions, that 2.5x is the right multiple to apply to TSEM, it is important to note that the choice of the growth multiple deeply affects the fair value of EPV valuations. A more conservative 2x multiple, if applied, would have reduced Tower's EPV to just $27 per share.

Fig.8: TSEM EPV valuation summary, author's work


It is not easy to provide an exact estimate of Tower's fair value. While it is interesting to note that various estimates seem to point at a fair price of $27 per share, I believe this target is still quite conservative and not fully appreciating the high growth potential of the company. My opinion is that shares could easily move higher depending on how well the company will be able to expand its operations in the next months, and I maintain a $34 target as the upper end of my fair value range.

The company also still trades at a discount to peers while enjoying higher growth rates. I argue such a discount is no more justified. Though until 2015 the company was able to throw out cash flow but not earnings and suffered from continuous share dilution; this is not the case anymore after the solid diluted EPS posted for FY2016. I believe the stock has quite a long way to go and therefore remain bullish on TSEM for 2017.

Disclosure: I am/we are long TSEM.

I wrote this article myself, and it expresses my own opinions. I am not receiving compensation for it (other than from Seeking Alpha). I have no business relationship with any company whose stock is mentioned in this article.

Editor's Note: This article discusses one or more securities that do not trade on a major U.S. exchange. Please be aware of the risks associated with these stocks.