Micron, SanDisk Struggle with Memory

Kris Tuttle submits: It’s no secret that memory supply/demand forecasts are not positive for the DRAM and NAND players like Micron (MU) and SanDisk (SNDK). At these levels most analysts are pointing to a bottom in these names and a chance to buy them into an improving supply/demand balance which brings better pricing and improved earnings for these companies.

The problem that I have with this line of reasoning is that everyone knows that the industry balance is fragile and very cyclical. There was a time from 1992 to 1995 where memory (and semiconductors in general) were said to be growing based on secular demand trends. This propelled earnings and the stocks to huge and sustained gains. There may have been a little price fixing around this period but we’ll let that one go.

Since then it just seems to be very easy for companies to increase supply when prices look like they will be slightly attractive. A holder in Micron stock would actually be down today versus the price paid anytime in the last 10 years. So where is the secular growth story here? Lots more bits are shipped every year but at an every declining price per bit.

The commodity run has investors saying that maybe memory will work this way too since we aren’t adding to capacity at these prices. This parallel is problematic in part because one has to discover sources in commodities and the costs and lead times required to deal with the physical and hazardous extraction and processing are fairly high. While a brand new fab is not cheap, there seems to be plenty of capacity for increases in production by starting up or adding new lines.

SanDisk at least has evolved into a memory company with a consumer retail facet and trades at a much higher multiple of sales than a pure commodity producer like Micron. However the same logic about better supply and demand applies here. Even if everyone is right and SanDisk has a big December quarter due to higher revenues and slightly better margins, nobody will think it can be sustained so the multiple stays low and the stock appreciation is limited.

If the commodity producers don’t have strong pricing power then the argument favors the branded device makers like Apple (AAPL), Research in Motion (RIMM) and even Dell (DELL) over the component suppliers. If memory is cheap it also helps move units and provides software makers like Microsoft (MSFT) very favorable revenue and margin dynamics that are more sustainable.

We welcome arguments to the contrary but we haven’t owned Micron since the mid-90’s and have been tempted by SanDisk but always scared off by what seems to be the same long-run truth about the memory business.

Comments

[rd4sndk:]

You're exactly why Sandisk price doesn't reflect its earnings and earnings growth. If you knew anything about sandisk you would know that its not a semi-conductor component supplier--despite analyst attempts to paint it as one. What Eli has done was to evolve sandisk into a vertically integrated mfg of flash memory cards and Consumer Electronics business model that is protected by fundamental Nand IP.

This is unique in that most companies abandoned vertically integrated semiconductor business when the Japanese and then the Koreans took over the semiconductor fab business and made it into a commondity business.

Eli correctly assumed as you do in your article that the business of consuming chips will be better than the business of making them. By partnering with Toshiba, Sandisk has an advantage over all consumers of nand as long as they stay at fabrication leading edge. It also has an advantage over the semiconductor suppliers by its fundamental IP.

There was only once in the past 7 years that Sandisk business model didn't work and that was last year when they absorbed both matrix and msystem in one year. Prior to that, you analyst kept looking for Sandisk earnings to turn south against lexar and the other nand consumers as their business model was based on Nand as a commodity. We all know what happened to Lexar.

With the msystem acquisition behind them, we longs are looking for Sandisk to get back to operating margins of 20% and earnings growth of 25%. It's interesting that if you plot the earnings of Sandisk and the price of its stock on the same graph, you will see that Sandisk is the most misunderstood company I know.

2007 Nov 28 06:32 AM

[jsam:]

The commodity memory market has always been a feast/famine cycle, as your article correctly points out. The nature of Moore's law points out that this is a race to the bottom: whoever gets there first maintains a temporary edge over the competition. Added to that, the nonlinear nature of adding production ($4Billion and 2 years buys the next increment in capacity) keeps things highly volatile.
True, Sandisk used to be above the fray. However their partnered investment in Fab 3 in '05-'07, and again in Fab 4 in '06-'08, has fundamentally changed their cost structure. They now will suffer the enormous capex and depreciation costs along with the rest of the commodity suppliers. Their "fundamental IP" is also fragile in the over-patented field of Flash technology (note the '517 patent ruled invalid), though it will continue to provide cash flow as licensing contracts are generally long term. Combined with the higher margin consumer market, they do deserve a higher multiple, but it is an open question as to which direction the multiple is moving.
In the end it is all simply a high risk bet, and a long term waiting game: The supply/demand balance is unpredictable in both directions.

2007 Nov 28 07:53 PM

[jcrash:]

Well, I'm long SNDK and would never think about buying MU. I never understood the appeal of Micron as an investment in that it had nothing to differentiate it. The company basically bleeds cash and has never had an attractive price.

SNDK, however, is a different beast. STX and WDC are "commodities" yet people are awakening to their appeal and future as an investment. This same awakening will happen for SNDK, and their future is even brighter.

Everything is going to have Flash in it. Who wants moving parts when they can have flash?

Bouncing off a 3 yr low with growing profits and revenues, this is when you should be buying SNDK.

I held back until the bounce in the Soxx and the bounce off the 3 yr lows for SNDK. Then I jumped in. Semis are breaking out. Get some NVDA and SNDK and watch your account grow.

2007 Nov 30 08:56 PM

[jsam:]

The commodity memory market has always been a feast/famine cycle, as your article correctly points out. The nature of Moore's law points out that this is a race to the bottom: whoever gets there first maintains a temporary edge over the competition. Added to that, the nonlinear nature of adding production ($4Billion and 2 years buys the next increment in capacity) keeps things highly volatile.
True, Sandisk used to be above the fray. However their partnered investment in Fab 3 in '05-'07, and again in Fab 4 in '06-'08, has fundamentally changed their cost structure. They now will suffer the enormous capex and depreciation costs along with the rest of the commodity suppliers. Their "fundamental IP" is also fragile in the over-patented field of Flash technology (note the '517 patent ruled invalid), though it will continue to provide cash flow as licensing contracts are generally long term. Combined with the higher margin consumer market, they do deserve a higher multiple, but it is an open question as to which direction the multiple is moving.
In the end it is all simply a high risk bet, and a long term waiting game: The supply/demand balance is unpredictable in both directions.

2007 Nov 28 09:31 PM

[jsam:]

The commodity memory market has always been a feast/famine cycle, as your article correctly points out. The nature of Moore's law points out that this is a race to the bottom: whoever gets there first maintains a temporary edge over the competition. Added to that, the nonlinear nature of adding production ($4Billion and 2 years buys the next increment in capacity) keeps things highly volatile.
True, Sandisk used to be above the fray. However their partnered investment in Fab 3 in '05-'07, and again in Fab 4 in '06-'08, has fundamentally changed their cost structure. They now will suffer the enormous capex and depreciation costs along with the rest of the commodity suppliers. Their "fundamental IP" is also fragile in the over-patented field of Flash technology (note the '517 patent ruled invalid), though it will continue to provide cash flow as licensing contracts are generally long term. Combined with the higher margin consumer market, they do deserve a higher multiple, but it is an open question as to which direction the multiple is moving.
In the end it is all simply a high risk bet, and a long term waiting game: The supply/demand balance is unpredictable in both directions.

2007 Nov 28 09:32 PM