Four Semiconductor Stocks Worth Owning in 2009

There is a serious over-supply of chips. There is a tremendous over-supply of Silicon [wafers]. Yet, there are some bright spots. The spot market was hot the week of Christmas, but the up-trend was NOT confirmed by long-term pricing of Flash and DRAM memories. Solar has for sure hit the skids, and it does not make any sense to install solar - if you are a business or a homeowner - unless of course, you live in California or Hawaii - where rates can easily exceed 38c/KWH on a sunny weekday afternoon, and the state is still willingly paying $1.55/watt for your solar power generation system which should cost a lot less than what I got as quotes from four of the largest names in solar.

When I tell them that there is an over-supply of panels, these guys just want to stick to script and convince you that it really is worth your while for the state/feds to pick up the tab - for an overpriced system that will not supply electricity when the grid is down [even when there is ample power generated by your system]. So, the worst company to buy into now is MEMC Electronic Materials (WFR) - it sells silicon wafers - that need a lot of energy to manufacture. Which gets us to the best semiconductor company at the current quote - NVIDIA (NVDA).

My reasons for buying this graphics chip-maker is manyfold:

a. Excellent Management.

b. Decent gross margins.

c. One-time screw-up in the design/manufacturing chain that is now behind NVDA.

d. And this statement from their latest 10Q:

During the three months ended October 26, 2008, we entered into a structured share repurchase transaction to repurchase 23.1 million shares for $299.7 million, which we recorded on the trade date of the transaction.

Not too bad - a company that is doing what they promised to - at least a quarter of the way there.

Pick #2 is SanDisk (SNDK).

Sure, the stock has run up a little in the last week of December and the first two days of Jan, but this is one solid memory company that will survive the down-turn. Moreover, SanDisk, and its partner in crime - Toshiba (TOSBF.PK) have taken the necessary steps to shut down part of their manufacturing in order to squeeze supplies. Moreover, the semiconductor industry as a whole is shuttering 30% of its capacity before the end of Feb 2009.

Reasons for liking SanDisk:

a. Excellent management - who are still stake-holders in the business [though not as much as they held a few years ago].

b. Has access to capacity when the market turns.

c. WIll have gross margins back in the black in the next two quarters.

d. And most importantly, its $400 Million/year royalty stream - which the market is signaling that they renegotiated with Samsung (SSDIF.PK) in the last weeks of Dec 2008 - on which SNDK spends little if any $$$$ generating [except litigation or the threat thereof once every few years].

Pick #3 is Applied Materials (AMAT).

Applied is the #1 manufacturer of semiconductor manufacturing equipment, and diversified into making machines that are necessary for the manufacture of flat panels and solar cells. With a hand in every possible use of semiconductor technology, a pristine balance sheet, and a stock that has under-performed the market since the internet boom, it is high-time that AMAT led the semiconductor back to good health - like it used to in the 1980’s and 1990’s [it is still big enough, its products, pervasive enough, and the company's financial health is better than ever].

Reasons why I like Applied….

a. Pristine balance sheet.

b. Participates in all semiconductor manufacturing - including chips, solar and flat-panel manufacture. In fact, they even supply technology to thin-film solar companies.

c. Should lead the sector up - after this down-turn.

Pick #4 is SunPower (SPWRA) (SPWRB).

The reasons for my liking SunPower have not changed since a prior article of mine. I sincerely believe that SPWRB can survive with or without subsidies - as it makes the most efficient panels, and is an integrated solar supplier. I prefer SPWRB over SPWRA - since the B shares have eight votes per share, and in a realistic market, should be worth more than the A shares - which have a vote per share. The reason why the A’s are deemed more dear is because of the fact that the A’s are option-able, while the B’s are not.

This is silly, and I know hedge funds that are short the A’s and long the B’s - counting on the spread between the A’s and the B’s to shrink/disappear.

Anyway, the four semis worth owning for 2009 are: NVDA, SNDK, AMAT and SPWRB.

Disclosure: No positions in all stocks.

Comments

[anand:]

As I understand it, SPWRA (or more precisely its joint venture) buys polysilicon, produces ingots, converts ingots into wafers (SPWRA buys ingots from its joint venture), converts wafers into cells, converts cells into modules, and installs modules.

Do you have a good estimate of what SPWRA's cost of goods sold is per watt excluding polysilicon (which SPWRA buys from suppliers)? What are your projections for how quickly SPWRA's cost of goods sold per watt excluding poly will drop over time.

Jan 09 01:46 PM

[anand:]

I believe that SPWRA uses 6 grams of poly per watt. This means that the polysilicon related cost of goods sold measured in cents per watt = (cost of poly in dollars/kg)*(kg/167 watts)*(100 cents/dollar)

For example at $100/Kg, poly represents about 60 cents per watt. At $35/kg (which is the cost of goods sold for Poly inclusive of depreciation), poly represents 21 cents per watt.

Is there a way to verify that SPWRA requires 6 grams of poly per watt?

Jan 09 01:54 PM

[Bapcha's St...:]

SPWR's COGS is 50% higher than the leaders in the business FSLR and SunTech. FSLR and STP are at $2/watt or lower.

Bapcha


On Jan 09 01:46 PM anand wrote:

> As I understand it, SPWRA (or more precisely its joint venture) buys
> polysilicon, produces ingots, converts ingots into wafers (SPWRA
> buys ingots from its joint venture), converts wafers into cells,
> converts cells into modules, and installs modules.
>
> Do you have a good estimate of what SPWRA's cost of goods sold is
> per watt excluding polysilicon (which SPWRA buys from suppliers)?
> What are your projections for how quickly SPWRA's cost of goods sold
> per watt excluding poly will drop over time.

Jan 12 10:17 PM

[BigVinnie:]

SPWR use to be the joint venture. Cypress Semi Conductor sold it's "joint venture" half making spwr two separate stocks. One being spwrA and the other being spwrB.

I'm selling short until January 29th. The price will bottom to $18~$20 and will come up to about $25~$28 by feburuary.


On Jan 09 01:46 PM anand wrote:

> As I understand it, SPWRA (or more precisely its joint venture) buys
> polysilicon, produces ingots, converts ingots into wafers (SPWRA
> buys ingots from its joint venture), converts wafers into cells,
> converts cells into modules, and installs modules.
>
> Do you have a good estimate of what SPWRA's cost of goods sold is
> per watt excluding polysilicon (which SPWRA buys from suppliers)?
> What are your projections for how quickly SPWRA's cost of goods sold
> per watt excluding poly will drop over time.

Jan 21 10:17 AM