Silicon Investor (SI) -- The First Internet Community

STOCKTALK

We've detected that you're using an ad content blocking browser plug-in or feature. Ads provide a critical source of revenue to the continued operation of Silicon Investor. We ask that you disable ad blocking while on Silicon Investor in the best interests of our community. If you are not using an ad blocker but are still receiving this message, make sure your browser's tracking protection is set to the 'standard' level.

Technology Stocks : Semi Equipment Analysis -- Ignore unavailable to you. Want to Upgrade?

To: cluka who wrote (52787)	7/16/2011 12:14:27 AM
From: Sam	2 Recommendations Respond to of 95761

Well, I've had this discussion several times now, and I'm not going to rehash the whole thing again. But briefly, in the words of one of the Sandisk posters, it may be a "commodity," but it isn't "just" a commodity. There are real differences between it and commodities like: oil, the metals, foodstuffs, DRAM, hard drives. Furthermore, Sandisk doesn't just sell NAND chips; they sell NAND solutions. This value added allows to capture margins that wouldn't be there for a product that is "just" a commodity. Some of the factors that differentiate NAND from other commodities and Sandisk from other NAND vendors include: 1. Very expensive to make a NAND fab. That in itself is a barrier to entry. That is one reason why there are only 6 NAND producers, or 4 if you count the Sandisk/Toshiba and Intel/Micron partnerships as one apiece. 2. But an even bigger barrier is the technology that is necessary to make NAND. Even "good enough" NAND. As Eli Harari used to say, it isn't easy to make NAND. And it has gotten progressively harder as smaller geometries have become state of the art. Look at dramexchange.com, and check out the differences between the "High Price" and the "Low Price for the contract prices for DRAM and NAND. The DRAM spread is extremely close; the NAND spread is very wide. There is a reason for this: NAND from different vendors isn't simply interchangeable, contrary to what some people have said. One chip isn't exactly like another chip, and can't simply be substituted. There is high quality NAND and low quality NAND. Sandisk is a proud engineering company, and they make high quality NAND. 3. There are also different kinds of NAND. Besides SLC and MLC, there is also TLC, or "tri-level-chips." TLC is cheaper than MLC, but is to MLC as MLC is to SLC--not as fast or as durable. But Sandisk has invented a technique to make TLC emulate MLC, and, according to Harari a year ago, even SLC in some applications. They started to qualify it in embedded applications last year, and have said that they are making progress in qualifying it in more apps this year. This is not a "commodity" type of difference. They say that their technique, which they call Adaptive Flash Management or AFM, is patented, and should allow them to charge MLC prices for TLC chips, giving them a margin advantage that no other NAND company has. 4. The fact that there are so few vendors is one reason why pricing probably won't get as vicious as DRAM pricing did. A related reason is that two of the vendors--Samsung and Sandisk/Toshiba--each control about 35% of the market, and both are technologically able companies. Hynix doesn't have the money or the technological prowess to challenge them, while Intel has not joined Micron in its latest fab in Singapore and has expressed ambivalence about remaining in the partnership. It will be difficult for Micron to keep up in both DRAM and NAND without Intel's financial help in the future. And even Samsung has apparently fallen slightly behind in node shrinks, with Sanshiba transitioning to 19nm and Intel/Micron to 20nm over the next few months, while Samsung is still transitioning to 26nm along with Hynix. These facts in themselves don't mean that it isn't a commodity, but it is a commodity that isn't easy to grow like corn, find like gold or oil or produce like DRAM. It is unlikely that either Samsung or Sanshiba will start a price war because they have so little to gain by it--it is unlikely that they would take meaningful share gains from each other, and Micron and Hynix don't pose real long term threats to them, especially since the market itself is expanding so rapidly. It is possible that as the output from the fabs that are just now coming into production gets out, there could be an oversupply until SSDs really take off. But the reason I put that article from Dan Amir in my previous post is that it is really only a matter of time before the SSD market takes off and sucks up the output from the new fabs, creating shortages. Even traditional commodity producers profit well when demand for the commodity outstrips the supply.