TranSwitch Corp. TXCC
April 21, 2004 - 12:42pm EST by
2004 2005
Price: 2.12 EPS
Shares Out. (in M): 0 P/E
Market Cap (in $M): 195 P/FCF
Net Debt (in $M): 0 EBIT 0 0

Sign up for free guest access to view investment idea with a 45 days delay.


I have written this stock up twice in the past, so the interested reader should also reference my previous write-ups.

The first time I wrote this stock up was near the bottom of the NASDAQ tech bust, and it was $0.74. The second time I wrote it up, the stock was $1.10. It is now $2.10, although the stock has been over $4 since that time.

I am the company's largest single shareholder and I have devoted a disproportionate amount of time and energy to this name because I think that it is an absolute home run idea, which has now been combined with an excellent entry point.

The stock originated as a speculative idea for several reasons:

1.) While there was a sizeable net cash balance, there was also a large convertible debt maturity in September 2005. At my suggestion (and you can reference my 13D filing one year ago to this day), I urged the company to exchange their convertible debt for one with a longer maturity so as to avoid any type of crisis at maturity issue. Somewhat to my surprise, they actually did this. Therefore, the company continues to have a sizeable cash and net cash balance, but the debt maturity has been deferred until 2007.

2.) Revenues had been literally decimated, leading to large losses and cash burn. What's more, it was unclear then when and if revenues would ever recover. I bought the stock heavily because I was convinced that the new products would hit the sweet spot of the market, but a year ago, my primary margin of safety resided in the low price and cash balance. The anticipated time frame for revenue recovery has always been in the latter half of 2004. Well, we are almost there now, and the propsects for a significant recovery in revenues is more certain than ever.

Therefore, I believe that the stock is substantially less speculative that it seems, and carries a good margin of safety at these price levels.


TranSwitch is a fabless telecomm IC manufacturer. They deal primarily in the intellectual capital of designing complex, richly featured IC's that enable the telecomm infrastructure. Actual production is outsourced to foundries, so the capital requirements are quite low. If chip designs are successful, then this becomes a very good business with significant visibility and free cash flow.

TranSwitch sells their chips to all leading international telecomm OEM vendors such as Cisco, Nortel, Lucent, Alcatel, Siemens, Huawei. Their chips are highly integrated, extremely complex and represent the core functionality and core intellectual property of these OEM devices. While this is not a break-up value play, I have no doubt there is a very considerable value in their intellectual propety (100+ patents).

The primary end customer of these OEM devices would be worldwide carriers and service providers.


If you have been on the planet Earth over the past several years, you should know that the telecomm industry has been through a virtual depression. From the top, telecomm spending has declined almost 70%. As this decline in spending fed its way through the food chain, almost every single company suffered dramatically, with many going bankrupt.

If you are familiar with the history of industry busts, you will know that at the bottom no one thinks the industry will ever recover. Yet, the industry does invariably recover, ofetn quicker than anyone ever imagines. And when the industry does recover, the weak companies have been eliminated, oversupply has given way to undersupply, the survivors are lean, mean and hungy, and the remaining stocks often do better than anyone dared to imagine. Just think real estate in the early nineties. Or the paging companies more recently.

In the particular case of the telecomm industry, the bottom was made in late 2002. 2003 was a year of recovery off the bottom, and 2004 wil be a year of modest growth. 2005/2006 could be years of very significant growth. But here is the important point: aggregate service provider spending will be up modestly, but the composition of that spending will be dramatically different from the pre-bubble days. A very important generational shift in network architecture is occurring, and those who are well positioned will reap significant rewards.


I would urge every interested investor to read the letter form TranSwitch's chairman in the recently released annual report.

Things can rapidly get complicated here, so I will try to keep it simple.

Prior to the bubble, voice was the dominant traffic type. Post the bubble, data is the dominant traffic type. Voice traffic is more or less stagnant, growing perhaps at 1% per year. Data traffic continues to bacially double every year. During the bubble, when money flowed like champagne, it was assumed that all-new data-optimized networks would be built to accomodate the ever increasing amounts of data traffic. Post the bubble, it became clear that voice traffic still paid the bills. While data traffic was exploding, it wasn't necessarily generating a lot of revenue yet. What's more, in the post bubble environment, there was no money to build these wonderful new all-data networks.

The macro them, then, is the convergence of voice and data networks--retooling the established voice networks to more efficiently transport data. And as service providers beginn spending again, all the money is going to go in this direction, creating very signficant opportunities for those who have the right products.


During the bust, most pure telecomm IC companies sharply reduced R&D spending or diversified into other, more promising businesses such as storage. TranSwitch did the opposite. They made a very gutsy bet on the future direction of the telecomm industry, and they signficantly increased their R&D spending to develop new products. The strategy didn't have to work, but it did, and they are now about to reap a very significant payoff.


The key to generating revenues is to get your chips "designed into" a manufacturer's platform. These are very highly functioning chips, and they are often the guts of the platform. Once designed in, they are in for the life of the product, and in this business that can be 5-8 years. If the product is successful, then there is a lot of money to be made.


The ET-3 will be the company's most important chip. It has received 87 design wins thus far, and at least 12 of these are with "Tier 1" firms like Cisco, Nortel and Fujitsu. TranSwitch was the only company to anticipate the need for this type of chip, and have been first and really only to market here. (Other companies have comparable chips but at higher speeds and for slightly different uses.)

Cisco is a sponsoring customer of the chip, and the ET-3 has achieved at least two very important design wins in their ONS series of transport platforms. (This is one of Cisco's core platforms.)

The boxes containing the ET-3 are just starting to ramp, and this is what will begin driving revenues in the latter half of '04. TranSwitch says that they would not even have attempted to design this type of chip if they did not feel it could generate $30 million in annual revenues of $30 million for some good number of years.

Let me be blunt: this product has been extraordinarily successful in racking up design wins. The company has basically a 100% market share for this type of device at the access level of the network. There is an unprecedented level of interest in this chip from OEM's large and small around the world. My feeling is that this chip could do somewhere between $50 and $100 million in revenues in 2005/2006.

What's more, the ET-3 is paving the way for other, less interesting but equally profitable chips that the company produces. By the company's estimate, each $1 of ET-3 revenue can pull in $ (or well more) of revenue from complementary chips.


I do not want to do a chip by chip analysis, but the company has numerous other chips that could be very important. The Ethermap-12, which is a higher speed version of the ET-3 is due to sample shortly. This has a higher ASP (average selling price) so while unit volumes will not be as great at the ET-3, the revenue potential comparable. As I said, the Ethermap-3 will be a company maker, so anything that is remotely comparable to the ET-3 will have a very very meaningful impact on the company's fortunes.

The Ethermap-48, an even higher speed version of the ET-3 and ET-12 should have been out by now but is frankly late. My analysis leads me to believe that this will be the best featured chip on the market, so I am confident it will gain its share of business. Unfortunately, the Vitesse chip has been shipping for some time so they are a little late to the game here. (Execution is a big risk here. These are complex devices.)

Another chip that will be impactful will the PacketTrunk-4. This chip has been sampled and the company is reporting extraordinary interest from manufacturers. Instead of discussing this chip, I will link you to a site which reviews it:

Incidentaly, Analogzone awarded this chip one of their Product of the Year awards. These awards are given to products that have the potential "to make a difference" to their company's. In other words, it is not an award for most elegant design. It is award for meeting the perceived needs of the market place.


Revenues have doubled over the past year, but still remain highly cyclically depressed. My expectation is for legacy revenues to continue to trend higher, but the real difference will be as the new products kick in over the 2005/2006 time frame. The beginnings of this acceleration should be seen in the latter half of this year.

As one considers the revenue outlook, it is important to realize that these are not "one-off" types of products. The end products are purchased by the large global and regional telephone companies. The initial products are subject to intense and length trial and qualification periods. After these are successfully completed, the revenue stream begins to ramp in a very serious and signficant way.

For example, the company has stated that OEM roducts containing the ET-3 are currently undergoing field qualifications. Orders are beginning to flow in in such a way as to believe that these products are going to be going into production.


I believe the following to be a possible, if not conservative, assessment of revenue possibilities:


ET-3 $50+ million
ET-12 $20+ million
Other $70+ million
Legacy $50 million


Personally, I believe all numbers above could be very conservative. The potential is there, the timing is unclear.

I believe 2006 will represent a year of very considerable growth as the ET-12 begins to achieve revenues comparable to the ET-3.

While these numbers may seem ludicrously high, I would remind you that this company has been spending about $50 million in R&D in the past year. In a normalized business model, R&D would be about 20-25% of revenues. Therefore, a more appropriate way of looking at it is that if this company does NOT produce these types of revenues, something will have gone very wrong in the game plan.

Given the company's very prescient read of trends in the telecomm market, I do not believe this is happening.

The company itself is very conservative in their projections. They have discussed what they call a "path to profitabiliity" in 2005. Breakeven would likely occur at slightly under $100 million in annualized revenues.

I believe the stock will trade at $6-8 under a breakeven scenario, and that would be a triple or better from here. I believe that breakeven will occur in the first half of 2005, with significant probability in the back half of 2005.


A very dramatic acceleration in revenue beginning in the second half of 2004 and continuing through 2005 and 2006.
    show   sort by    
      Back to top