TowerJazz: The Tragedy Of Locked Value

Executive summary:

• TowerJazz has substantial growth ahead, with multiple opportunities.


• Unfortunately, dilution and debt are blocking TowerJazz's value.


• Until dilution and debt issues are solved, I'm staying on the sidelines.

_________________________________

TowerJazz (TSEM) is an Israeli-based specialty foundry. For those who aren't familiar with the specialty foundry market, I'll begin with a brief introduction.

Specialty Foundries

The specialty foundry market supplies chips to all the semiconductor design houses that don't require cutting-edge chip fabrication technologies. That includes analog chips, such as power management chips, radio chips, controllers, audio chips, etc. Here's an easy-to-understand slide that compares the analog and digital chip content of the iPhone:

(click to enlarge)

Source: TowerJazz investor presentation.

Now that the top four digital foundries in the world (TSM, UMC, SMI, and Global Foundries) have gone on to chase Moore's Law, the specialty foundries have much less need to focus on R&D.

(click to enlarge)

Source: TowerJazz investor presentation .

While digital foundries fabricate chips using advanced technology nodes (65nm-14nm), specialty foundries work with customers who can manage with 65nm+ chips. TowerJazz is the largest specialty foundry in the world.

Operational Brief

TowerJazz owns and operate four fabs:

• Fab 1 + Fab 2: Both located in Israel, possessing a combined annual production capacity of 720K wafers.


• Fab 3: Located in Newport Beach, California. Has a capacity of 290K wafers per year.


• Fab 4: Located in Nishiwaki, Japan. This is an old Micron (MU) fab, and it has a production capacity of 700K wafers per year.

Recently, the company agreed to enter a joint venture (JV) with Panasonic to acquire three old Panasonic fabs with a combined capacity of 800K wafers per year. The agreement also stated that, in addition to its other customers' products, TowerJazz will manufacture Panasonic products for the first five years of these fabs' operation. One of the three facilities is a 300mm fab, which will enhance the product offerings of TowerJazz.

TowerJazz also recently announced that it has won a contract in India. For this project, the Indian government chose a consortium comprising IBM (IBM), TowerJazz, and an Indian company called Jaiprakash. The contract is for the construction and operation of a fab, with a total estimated cost of $4.1B. TowerJazz will supply advising services, support, and know-how. The contract is expected to bring revenues to TowerJazz of $300M over six years. The first three years of the project will be the construction phase, and this will be followed by three years of facility operation.

Debt Burden

Ten years ago, TowerJazz embarked on a journey of growth. This started with the merger of Tower with Jazz.

(click to enlarge)

Source: TowerJazz investor presentation .

Achieving growth required capital. That capital was not available to TowerJazz in its 2005 financial state. On its way, TowerJazz raised substantial amounts of debt in various ways. This debt has accumulated to almost toxic levels, keeping the company in a distressed valuation for the most of the last decade.

TSEM data by YCharts .

Here is a summary of TowerJazz's current debt structure.

Year	2014	2015	2016
Bank Loans	$83.2M	$39.9M
Debentures	$33.1M	$238.1M	$131.3M

Source: 2012 20-F.

Unfortunately, most of this debt is convertible, so that the threat of dilution hangs over the company. If all of the convertible debt is indeed exercised and converted, the company could face a massive dilution that would increase its share count to ~83M from its current ~47M. This would represent a 72.3% increase in the number of outstanding shares. The largest chunk of debt (series F) has a conversion ratio of 38.21 NIS to one ordinary share. That means that series F alone could cause a 20M share dilution. 38.21 NIS equals about $10.9, so series F is not yet in-the-money, but that price doesn't seem far away. For TowerJazz to reach its potential, it has to overcome this hurdle and leave its massive debt issues in the past. There are a few possible scenarios:

1. The debt is converted. There is no cash impact on TowerJazz, but shareholders are massively diluted.


2. The debt is not converted. TowerJazz will almost entirely drain its cash reserves to pay it, and cash flow will be impacted in 2015-2016.


3. TowerJazz reaches an agreement with debt holders to postpone the principal payments.

Substantial Growth Ahead

Now, we will take what I think will be an interesting approach to valuing TowerJazz's business. Let's try to estimate TowerJazz's cash flow in the coming years. The reason for doing so is that depreciation can distort the underlying business. I, for one, would like to value this business via the rawest indicator of value; this, of course, is cash flow. There are three elements to discuss: the core business (current operations which include fabs 1-4), the Panasonic business, and the operations in India. The last one will contribute greatly to both the bottom line and cash flow.

Core Business

On the last earnings call, company CEO Russ Ellwanger mentioned the utilization rate for fab 1-3. In the Q2-13 call, he provided the utilization rate for fab 4. Running those numbers, we get the following:

Fab	Utilization Rate	Annualized Output
1+2	90%-60%	592,500 wafers
3	65%	188,500 wafers
4	40%	280,000 wafers
Total	60%	1,061,000 wafers

The quarterly rate is about 265,250 wafers per quarter. Given last quarter's revenues and cost of goods sold, we get the following numbers:

• ASP per wafer is $499.


• Average cost per wafer is $351.


• Average gross profit per wafer is $149.


• Average gross margin is 29.8%.

In the last earnings call, Russ also stated that the company has seen growth in its new masks. In the world of foundries, the term "mask" refers to a photomask, which is a pattern according to which the foundry manufactures a wafer. Each new product has its own photomask.

Russell C. Ellwanger - Chief Executive Officer, Chairman of TowerJazz

Our revenues in the quarter were $132.6 million, within our expectations for the quarter and representing a 6% quarter-over-quarter growth. While our revenues are below those of the third quarter of 2012, in the previous year, we had a high level of Micron contractual revenues. Excluding Micron, there's a year-over-year mid-single-digit growth. In Q3, we had record number of new masks entering our factories Q1 to Q3 '13 as compared to the same period in 2012, so the 34% increase or over 3,000 more new mask-sets entering into the factories.

As previously explained, the release of the photomask to the factory is the last formal step in the design development cycle, after which product manufacturing may begin. Each customer design has its own set of photolithography masks, the design patterns of which are transferred onto the silicon substrate. The time from mask tape-out to volume manufacturing is typically one year, with a net volume life of 2.5-3 years from the tape-out date. The 34% increase from Q1 to Q3 as compared to the previous year demonstrates effectiveness in realizing the customer projects and the sales revenue funnel.

This is very informative. In the first nine months of 2013, TowerJazz won 3000 more new products. It takes about one year for volume manufacturing to commence, and another 1.5-2 years until the customer stops ordering that product. I will assume that revenues from the current state of the business, or core business, will grow 5% from quarter to quarter, or 21% Y/Y. I think this is a fairly conservative assumption, given the expected market growth. I expect this growth rate to be maintained for the next several years.

(click to enlarge)

Source: TowerJazz investor presentation .

Panasonic Business

The Panasonic business will begin to operate under the JV in April 2014. The operation is expected to contribute $400M per year in revenues. TowerJazz has agreed to produce Panasonic products for the first five years, and given the cutthroat price that TowerJazz is paying for 51% of this JV ($8M in stocks), I think we can expect Panasonic to push TowerJazz's gross margin down. If the three Panasonic fabs operate at 85% utilization, they will produce 680,000 wafers per year. I will further assume that almost all of these fabs' output will be purchased by Panasonic. $400M divided by 680K wafers gives an ASP of $588 per wafer. Assuming a cost of $480 per wafer (old equipment), TowerJazz will then manage to get 18% of the gross margin from this business. S&M&G&A expenses are usually around 8% of revenues for the company. Applying the same ratio to the Panasonic business, we get 10% of operating income.

Operations in India

According to recent news, the final agreements should be signed around August 2014. Assuming a four-month margin of safety, I expect revenues to start coming in 2015. This project is expected to bring $50M in revenues per year for six years. The profitability of the venture should be very high, as TowerJazz will supply just IP. According to the Israeli press, involved parties have estimated the profitability of this venture at ~90%. That means that each year, TowerJazz will net $45M, which will go straight to the bottom line.

Putting It All Together

Next, we'll try to determine the non-GAAP net income, which should be about the same as the operating cash flow. (The D&A item is the main difference between GAAP net income and operating cash flow. Other non-cash items range from $10M-$15M a year). The reason to do so in this case is that TowerJazz has a big financial expense item every quarter. That item is a cash expense, so to get a closer estimation of operating cash flow, we need to look at Non-GAAP net income. Let's combine all of the above assumptions into one table.

Year	2014	2015	2016	2017
Revenues	$930M	$1.2B	$1.38B	$1.58B
Gross Profit	$243M	$347M	$396M	$456M
Operating Profit	$102.6M	$172.5M	$198.9M	$231M
Financial Expenses	$34M	$30M	$20M	$12M
Net Income	$68.6M	$140.5M	$175.4M	$219M

All numbers refer to non-GAAP measures, which exclude depreciation expenses.

As of September 2013, TowerJazz had $141.4M in cash. Let's examine three scenarios:

1. All of the convertible debt is exercised (no cash expense).


2. None of the convertible debt is exercised (TowerJazz has to pay 100% of its convertible debt).


3. Principal payments of debentures are postponed (again, there is no cash expense).

Year End	2014	2015	2016	2017
Scenario 1 Cash Reserves	$60M	$170.5M	$266M	$485M
Scenario 2 Cash Reserves	$30M	($49.4M)	$146M	$285M
Scenario 3 Cash Reserves	$60M	$170.5M	$266M	$485M

Assumptions: CapEx of $80M per year. Scheduled payments of bank loans' principal. Scenarios 1 and 3 result in the same cash reserve balance.

Because the probability of each scenario is unknown, I'll assume a 33.33% probability for each. Doing so brings us to these results:

Year	2014	2015	2016	2017
Revenues	$930M	$1.2B	$1.38B	$1.58B
Year-end Cash Reserves	$50M	$97M	$225.7M	$417.8M
Yearly Operating Cash Flow	$68.6M	$140.5M	$175.4M	$219M

Comparison to Peers

Let's take a look at the valuation of two peers, UMC and SMI.

Company	UMC	SMI
Sales	$4.15B	$1.96B
Market Cap	$4.93B	$2.71B
Operating Cash Flow	$1.4B	$738M
Price-to-sales	1.18	1.38
Price-to-cash-flow	3.5	3.7

Numbers are from UMC's latest report, and SMI's latest report. The numbers are 2013 yearly results.

Valuation

If we apply those ratios (taking the low end) to TowerJazz numbers, we get:

• 3.5 times the 2017 cash flow is $766.5M.


• 1.18 times 2017 sales is $1.86B.

Why the difference? It exists because SMI is far more profitable than TowerJazz. While SMI's cash flow is about 37% of sales, TowerJazz's cash flow in 2017 would be just under 14%. So using the cash flow multiplier, we get to a market cap of $766.5M. I lean toward the $766.5M valuation, while giving consideration to the many uncertainties about what the company's capital structure will look like in 2017.

With the dilution 'shadow' hanging over TowerJazz, purchasing shares of the company is currently an action that involves far more risk than reward. Let's say you were to purchase one percent of the company at today's market cap of $342M. You pay $3.42M, and you get about 478,500 shares. Next, let's say that you hold them until 2017, the company's market cap rises to $766.5M, and that 33% of the potential dilution actually occurred. The company would then have almost 60M shares. Your piece of the company will have shrunken to 0.79%, which will be worth $6M. This is a 77% increase in value over three years or an annualized return of 20.9%. The worst case scenario here is a full exercise of convertibles, which will bring the share count up to 82M. In this case, your piece of the cake will have shrunken even further, to 0.58%. That will be worth $4.44M. This is a 30% increase over three years, or a mere 9.1% of annualized return.

IBM Rumors

IBM is looking for a buyer for its foundry business. Given that the company has worked closely with TowerJazz on the Indian fab, many people believe that this is a perfect match. Frankly, I believe that this is out of reach for TowerJazz. IBM's two fabs are big, and one of them fabricates 22nm chips: a technology node that is not relevant in the market that TowerJazz participates in. If any kind of transaction happens, it will change the basic definition of TowerJazz's business, but I still think that this is a low-probability event.

Conclusion

I think that given its current debt and dilution risks, TowerJazz's great growth story is not accessible to long-term investors through conventional investing. Not, at least, until the company clears its balance sheets and works out its dilution issues. Once these issues are a thing of the past, TowerJazz's growth story could be accessible to investors and offer great value. Until that day comes, I will stay on the sidelines on this one.

Source: TowerJazz: The Tragedy Of Locked Value

Disclosure: I have no positions in any stocks mentioned, and no plans to initiate any positions within the next 72 hours. 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. (More...)

This entry passed through the Full-Text RSS service — if this is your content and you're reading it on someone else's site, please read the FAQ at http://ift.tt/jcXqJW.