Sun Microsystems.

Sun Microsystems – an Rx for survival.

April 25, 2005.

In the heady days of the dot-com boom, Sun proclaimed that they were the “dot” in the dot-com. I used to quip “they’d better want to be the com in the dot-com”. Jokes in which Scott McNealy was compared to god were not uncommon, Indian programmers who could write a program to make a computer print “hello world” – could get a VW Beetle as a sign-on bonus to be the next great Web-Master. Blodget proclaimed that Amazon was a $400 stock – and it soon was! Yahoo had a market value in excess of the entire semiconductor industry’s [sans Intel and Motorola].

The boom also converted Sun from a purveyor to the tech pinheads who needed workstations to write programs and build chips – to a company that sold the infrastructure components of a network [web-server hardware and software]. Also, Sun was selling a lot of product that used Intel’s CPU’s – in addition to their proprietary SPARC-based boxes.

SPARC started out as an interesting competition to Intel’s x86 architecture, but by the mid-90’s every company that once tried to make SPARC chips – with the exception of Sun themselves, had ceased to do so [Cypress, Ross Tech, Fujitsu] – since Intel had won the CPU race so convincingly, and none of the names mentioned above profited significantly from SPARC sales.

Today, it is not sufficient for a CPU architecture to just be the “brains” of a computer box. A company that spends $$$$ designing a CPU must leverage it for use in embedded computing. Examples of embedded computing include game consoles, automotive [electronic ignition, ABS], and print engines. All CPU architectures - with the exception of SPARC, find their way into multiple embedded applications [which is exactly why Arm Holdings has been successful to date].

Getting to the point, my Rx for Sun’s survival is simple. Close the SPARC division, give up designing chips and use Intel’s CPU’s [much the same thing that Silicon Graphics did in order to survive] – or, spin off the SPARC division and see if they can survive [a-la MIPS]. This gets Sun out of the highly capital intensive and competitive chip-design market [which is not Sun’s core area of competence].

Sun can then focus on what they are good at – building software and hardware for web-services, network infrastructure, Solaris; and compete in the services sector [with IBM].

Copyright Bapcha, April 25, 2005.

SanDisk.

Qualitative:

The global market for semiconductors is around $260 Billion annually. In this market, memory chips account for about $40 Billion a year. The fastest growing segment of the memory market is non-volatile storage or flash chips. These are chips that retain the data stored in them even if the batteries are turned off. Of late, they have become ubiquitous and reside in everything – cameras, cell-phones, and MP3 players, “smart” credit/debit cards. Newer cell phones come with a memory slot [S/D, MMC], and camera pixel resolution is rising – as is the need for compact, low power, non-volatile storage. The largest player in the flash market is Sunnyvale based SanDisk.

So what is it that sets SanDisk apart from their peers? I know for a fact that SanDisk has great talent. They retain their talent by providing a challenging work-culture, and by addicting their employees to stock options. Of course this will impact their results when the new accounting rules are imposed come June of 2005, but I foresee an era in which earnings are reported with and without the effect of stock-option compensation.

Quantitative:

The company is profitable [they beat earnings estimated by a dime in the previous quarter] – while the cost per Megabit drops precipitously. Currently, their gross margins stand at 35% - good but needs to be better. Meanwhile, SanDisk is aggressively transitioning their designs to 70 nm wafers, and will probably have it in production by sometime later this year. This is absolutely critical – since they need to get their costs down before the next buying season.

SNDK has 180M outstanding shares – putting their market value at about $5 Billion. A full quarter of their valuation is in cash! The mean analyst estimate for earnings for Q1 is 31 cents a share and for revenues is $467 M. I expect SanDisk to handily trounce these numbers.

Possible negatives:

Due to the nature of the memory market, companies are routinely stuck with inventory that they cannot give away. SanDisk has had to write down their inventory in the past, and unless they transition to a lower cost structure in the near future, they will be stuck doing that again. While their receivables number is down [good], their inventories are not [not good/neutral]. This is one uncertainty that any investor in this sector has to live with – for there is no easy way to tell if something will or can be sold sometime in the future – and for how much.

In conclusion, SanDisk is a well run Semiconductor Company, with great products, a good brand, and effective employees and will reward risk-tolerant investors in the next two to three years.

Copyright Bapcha, April 20, 2005.