The big news in semiconductors was the possible takeover of SanDisk by Samsung on Friday, 5th Sep 2008. Let’s for a second assume that Samsung [which already has close to a 50% market share in NAND flash] can indeed acquire SNDK [the #2 player with a little over 10% market share] - and keep the US and EU anti-trust gods happy……
The question of the day is, “how much will it cost Samsung to acquire SNDK?”. There is a short answer and a long answer. The short answer is - not just SNDK’s enterprise value. The long answer is SNDK’s enterprise value plus off-balance sheet obligations [but, paying for these obligations buys SNDK the wafers they need to make chips]. Barrons’s published an article a month and a half ago on how SNDK’s off-balance sheet obligations will affect everything. I think that it is a non-issue I am sure that there is a dozen lawyers looking over SNDK’s obligations that total a whopping $6.5 Billion, of which $1.8 is due within twelve months. Now, why does SNDK have these huge “obligations”?
While SNDK claims to be fabless - and they are in the purest of definitions of fabless, fabless; they have partnerships with various fabs - and, on a rolling nine month projection basis, guarantee to buy a certain number of wafers at preset prices. Sure, SNDK needs these wafers to make their chips, but SNDK also guarantees leases on property, plant and equipment. Out of the $1.8 Billion due in the next twelve months,
a. $1.05 Billion was for fab capacity expansion - including p.p.e and depreciation
b. $225 M was for operating expenses
c. $474 M was for the purchase of wafers [that SNDK needs to make chips].
Toshiba [which benefits from SNDK's off-balance sheet obligations] is betting huge sums of money on future SanDisk technologies - including a hair-brained 3D cell, which I guarantee will not work. In fact, SNDK hedges in their own 10Q filing and says “For example, our competitors are developing new technologies such as charge-trap flash and three-dimensional technology which if successful and if we are unable to scale our technology on an equivalent basis, could provide an advantage to these competitors“. I think that SNDK has more to fear from Seagate than from 3D or charge-trap based technologies.
Also, SNDK will generate royalty and licensing revenues of almost a half a billion dollars in this fiscal year. 70% of this is from Samsung, and this agreement needs to be renegotiated in August of 2009 failing which, SNDK’s financial results will be adversely impacted. Samsung could easily get rid of this $350M/year in royalties and licensing fees by acquiring SNDK.
SNDK’s management is tenacious, always ahead of the curve, and the company makes sure that their engineering and marketing groups are always alert and firing on all cylinders. Since I know too many people who work at SNDK, I will not be too specific or share stories in this note - but will gladly share my thoughts with you over e-mail.
1. SNDK is uniquely positioned to win the NAND Flash wars.
2. I think that SNDK has one of the most aggressive and successful management teams in the semiconductor industry. Their entrepreneurial spirit will be snuffed out as a Samsung subsidiary [as opposed to them staying independent].
3. SNDK receives over $500M/year in licensing and royalty revenues, and should be able to renew the agreement with Samsung - without having to litigate.
4. SNDK has a unique fabless [yet with fab obligations] business model - that other companies may want to look at in order to emulate.
5. SNDK’s finances are excellent [once you admit the fact that without the fab agreements in place, they cannot stay in business].
6. SNDK currently trades at the low end of its PS ratio [which I think of as a useful evaluation criterion for cyclical semiconductor companies].
7. Historically, SNDK has come out of semiconductor cycles - roaring.
Bottom-line: I do not want Samsung to acquire SNDK.
1. Bapcha is long SNDK. No positions in other companies mentioned.
2. SNDK’s latest SEC 10Q filing.
I was talking to a couple of people I know, and e-mailing a few more regarding Samsung’s $26/share offer for SanDisk - and SNDK’s prompt rejection of the [as it is now] - friendly overture from Samsung. So, the question was - is a hostile takeover is SNDK possible at all?
For starters, Eli Harari is an excellent CEO, and SNDK is his baby - a symbol of all he has accomplished successfully as a technologist and a businessman. He is proud, and has a modicum of ego (actually, a lot more than that), and cannot imagine himself playing second fiddle to someone in charge of building ships and chips at Samsung. He has built a fantastic business out of making non-volatile memory, and portable music devices (SNDK’s Sansa - though a distant second to the iPod, has sales that are at least thrice that of MSFT’s Zune).
Secondly, there are two anti-takeover provisions in the 8K that I pulled from the SEC Website. Quoting verbatim from the filing referenced in the previous sentence…..
“We have taken a number of actions that could have the effect of discouraging a takeover attempt. For example, we have adopted a stockholder rights plan that would cause substantial dilution to a stockholder, and substantially increase the cost paid by a stockholder, who attempts to acquire us on terms not approved by our board of directors. This could prevent us from being acquired. In addition, our certificate of incorporation grants the board of directors the authority to fix the rights, preferences and privileges of and issue up to 4,000,000 shares of preferred stock without stockholder action”.
In short, they have not one, but TWO ways to stave off a hostile takeover. In addition to these anti-takeover provisions, Toshiba, that benefits from SNDK’s off-balance sheet obligations that I outlined in a previous article, has already said publicly that a Samsung takeover of SanDisk will not go uncontested. Plus, it will corner SNDK between a rock and a DMZ south of North Korea - especially as it applies to Flash Partners I & II - since Samsung has enough Flash capacity to meet all of their, and SNDK’s needs.
Yet, a deal can be consummated if Eli wants one……
Published on Sep 16th 2008 at http://www.bapcha.com
Disclosures: Long SNDK.