Among the companies that we looked at in the part one, many can be eliminated with ease. So, let’s get the job done systematically.
- Applied Materials (AMAT). With pristine financials and the enviable position as the #1 purveyor of machines that the semiconductor industry needs to make their chips, Applied flies through the initial screen.
- nVidia (NVDA). With little debt, and already having survived a bush with bankruptcy, nVidia is flush with over a billion in cash, and a winning portfolio of graphics chips and cards. nVidia gets through the initial screen too.
- Silicon Storage Technologies (SSTI). This one is a little harder. This manufacturer of NOR flash is extremely tight fisted [and I like that]. Their off-balance sheet assets easily exceed the company’s current market value. Add to this, a steady stream of royalties [$30 to $40 million per year], SSTI squeaks through the initial screen.
- SanDisk (SNDK)- I have this eerie feeling that this is the only memory chip maker with the possible chance of making it to the final five.
- Lam Research (LRCX) - has done well in the last few years, and has decent financials, but they are looking at a very difficult years ahead of it. Lam does not make it through the initial screen.
- Novellus (NVLS) - similar to Lam.
- Kyocera (KYO) - this maker of ceramic packages for chips is flush with cash, and has survived Japan’s nuclear winter through the last two decades. They sure need a break.
- Intel (INTC) - the #1 chipmaker is still profitable. I am sure it will make the top five,
- Texas Instruments (TXI) - has a chance.
- Linear Tech (LLTC) - with gross margins in excess of 70% and a current yield of over 3.5%, Linear is sure of making it to the top five.
- Maxim (MXIM) - with a current yield of 6.5%, a billion in cash and no debt, and gross margins close to 60%, maximum gets through round one with ease.
- Marvell (MRVL) - This one gets a reject due to corporate governance issues.
- Broadcom (BRCM) - yes! These guys easily make it through round #1.
- Altera (ALTR) - with gross margins in excess of 60% and a history of profitability, ALTR makes it through round #1 with ease.
- Xilinx (XLNX) - numbers are similar to Altera’s and with excellent products and software
- Lattice Semi (LSCC) - CUT.
- Actel (ACTL) - CUT again.
- Micron (MU) - CUT. No need to explain things here.
So, let’s see. The survivors are:
Applied Materials, nVidia, Silicon Storage, SanDisk, Kyocera, Intel, Texas Instruments, Linear Tech, Maxim, Broadcom, Altera, and finally, Xilinx.
With gross margins in excess of 50% despite manufacturing Flash memories, Intel will obviously survive this economy, and thrive. Intel has a cash hoard of almost $12 Billion with debt of a little over $2 Billion. While INTC was barely profitable last quarter, the fact is that they were profitable! While AMD and Via technologies make clones of Intel’s CPU, even when AMD was gaining ground on Intel with their Opteron Processor a few years ago, Intel’s market share has never dipped below 80%. Added to this, the fact that Intel’s Atom has cornered the low-power net-book market, and found its way to becoming the processor of choice for Apple (AAPL) [who resisted for almost two decades], Intel is the undisputed leader of the semiconductor industry.
If gross margins were the only way to judge a semiconductor company, the easiest pick for this list of five would be Linear Tech. Linear Tech makes a multitude of analog and mixed signal chips, and has always been a favourite of mine. In fact, I discussed in a previous article the only super-profitable niche in the semiconductor industry - where Linear Tech is the obvious leader. In the three months that ended Dec 28, 2008, Linear Tech had revenues of $249 Million. For the first time in almost a decade, Linear tech had lower revenues in the current quarter as compared to the same time period a year prior. The reason for this is because automobile manufacturers buy a lot of LLTC’s products, and that market as we all know is in the tank.
Yet, Linear’s chips are cheap - with ASP of $1.56 in the current quarter, UP from $1.49 in the corresponding quarter in the previous year - due to decreased sales to cell-phone companies and increased sales to industrials. Gross profit as a percentage of revenues decreased to 75.8% and 76.5% in the second quarter and the first six month period of fiscal year 2009 as compared to 77.1% and 77.2% of revenues, respectively, for the same periods in the previous fiscal year. The decrease in gross profit as a percentage of revenues for the three and six months ended December 28, 2008 was primarily due to spreading fixed costs over a lower sales base.
Linear’s current dividend rate of 4.1% [88c/share/year] is easily earned by the company. So LLTC is my #2 pick.
If I had to pick one stock in the memory market, it would be SanDisk. While SNDK has negative gross margins in the current [brutal] Flash pricing environment, they do have a royalty stream that amounts to anywhere from $350 to $400 Million dollars per year. This number will increase as more Flash companies are taken to task for violating SNDK’s strong patent portfolio [but they need to start making $$$ first]. I do not see a recovery in the memory market till at least calendar Q3 of 2009. In fact, I expect SNDK to have positive product gross margins only in calendar Q1 2010 - and that will be due to migration to tighter geometries and their ability to successfully push from X3 to X4 across their product line. [X3 is the ability to store eight (23) different threshold levels in every storage bit - while X4 will represent sixteen (24) "levels" in every storage bit].
But the real reason for my picking SanDisk is due to the fact that it is run by an “owner/founder” CEO - as opposed to employee CEOs. While they might have overpaid for M-Systems and Matrix Memory [for which they took an asset impairment write-down recently], they are still the company to beat.
Another “owner/CEO” that I respect a great deal is Jen-Hsun Huang of nVidia. NVDA is cheap on a valuation basis, and the next version of the Apple OS [snow leopard] will take advantage of the processing power of the multiplicity of cores in NVDA’s chips. I have written about NVDA’s technology, and processing power in prior articles. For the sake of brevity, I will not repeat myself. I expect NVDA to NOT be profitable in 2009, but this is an article about the ability to survive the nuclear winter that semiconductor issues are in the middle of. NVDA’s cash position is sufficient to let them innovate over the next two years - and the only way that they will lose is by directly challenging Intel. For starters, only AMD and VIA tried to challenge Intel in the processor wars, and they have both lost market share due to the success of Intel’s Atom. NVDA needs to launch a flanking attack - meaning, sell the computing power of a GPU and do not market it as a CPU.
NVDA has a cash hoard of $1.25 Billion, and very little debt of about $26 Million. They have been buying back their stock aggressively, and have spent in excess of $1.46 Billion so far, with $299 Million spent in the quarter ending Oct 28, 2008. I think that they have bought back enough shares, and while they are authorized to buy back up to $2.7 Billion in total, further buyback of their own stock will seriously affect nVidia’s competitive ability to thrive after the current lean spell. So, I am picking NVDA at slot #4, but will keep a close watch on their cash position, and competitive position [as AMD's ATI division has a line of excellent GPUs that compare favorably with NVDA's offerings].
My final pick was a toss-up between Marvell and Broadcom. While I expect BRCM to lose a nickel a share in Q1-2009, they should break even or make a little money in 2009. While Marvell might do better financially, I have more doubts about MRVL’s corporate governance than I do about BRCM’s. Moreover, since Dr. Henry Samueli was my graduate student advisor at UCLA, let’s call this my “sentimental”, yet well researched pick. BRCM has a cash hoard of $2 Billion, with $0 in debt, and a product offering that is better than everyone else’s in the enterprise networking sector. Plus, I know for a fact that BRCM has some of the most brilliant minds designing their chips.
So there we are. The five companies that will survive this downturn in the semiconductor market, and thrive afterwards, have a few things in common.
A strong balance sheet, with a large cash position [or in Linear's case, a manageable debt, and a fantastic competitive position as reflected by gross margins].
An excellent competitive position.
Strong management that is still “invested” in the company.
- Has survived prior attacks on its competitive position.
- Passes my rigorous standards .
If I had doubts about a single pick, it would be about SNDK, but people NEED memory. There is no way that computing can move forward without flash memory. Sure, rotating hard-disks have “out-Moore’s-Law’d” Flash memories, but heck, if you hate memories, pick Altera or XIlinx. They are rivals with excellent balance sheets, gross margins in excess of 60% and split the market for Programmable Logic [with small concessions to Cypress (CY), Actel, QuickLogic (QUIK), etc.], but I chose to not pick them.
The final list: Intel, Linear, nVidia, SanDisk and Broadcom.
Disclosures: I own a small long position in INTC, BRCM. No positions in other companies.