Pam, That's not exactly a foregone conclusion just yet: At this moment OneNAND rev is just a small portion of Samsung's total NAND rev and mobile phones is just one of the major applications for OneNAND. Also I think there are more flash technologies that stand a good chance below 45nm in the NOR space.
NOR market has been experiencing a slow, painful death over the last few years (and it has been discussed over at the Sandisk thread for a long time- read posts from a year and a half or two ago) and whether you like it or not, the pace is just going to accelerate. NOR's main advantage was code execution and NAND, in the form of OneNAND, mDoC, etc. is fast replacing NOR based solutions because they are able to deliver XIP just as reliably AND at a much lower price! NAND is way too cheap compared to NOR and will continue to do so because NAND cell has a much smaller feature size and is on a much steeper cost-reduction slope going into the future. So a rapid decline in usage of NOR in cell phones is just going to accelerate. The other thing is that there is a growing demand for data storage in cell phones and they all want their phones in a slim form-factor. If products like OneNAND and mDoC or other MCP based solutions can provide data storage as well as code-execution capabilities either on a single chip (OneNAND) or in a MCP form (mDoC) it saves space in the device and gets the job done at a much lower price. Replacement of NOR with NAND based solutions is a no brainer, IMHO.
I agree, revs from OneNAND are much lower than NOR revenues today, but this is changing and changing fast. In fact, even if NAND based solutions do not prevail, overall market for NOR in terms of PBs sold is going to grow at a much lower pace than NAND market and companies like SPSN, Numonyx and others will chase the same opportunities in a same size or a shrinking size market (in $$ terms) and what that means is that they will have no pricing-power! Vendors will squeeze out the suppliers and there will be no margins. In the meantime expenses continue to go up for the industry as they are forced to move production to 300mm and finer geometries to lower their cost per bit.
I do not know what you mean when you say that NOR has a better chance below 45nm? Toshiba is ready to go into production at 43nm and introduce 3 bits/cell NAND in 1H08 and they are already shipping 16Gb NAND MLC monolithic chips at very mature yields on 56nm today! SPSN just announced that they are sampling 1Gb chips at 65nm today. There is absolutely no match here when it comes cost/MB of flash and SPSN will not even be able to catch-up with Toshiba in terms of lowering costs for another 5 years at minimum, if at all! NAND will scale another couple of generations and after that they will move to 3D. NAND is here to stay for and dies will shrink one way or another for at least 5-7 years at minimum.
Indirectly important: As Samsung moves into the traditional NOR market space mostly at the cost of Numonyx (btw this trend isn't likely to break any time soon), Spansion has recently started to move into the traditional NAND market space. Better late than never. In a few years the differences will blur even more.
Samsung will have plenty of spare 200mm capacity which is fully depreciated and they would be glad to get whatever RoI they can from these investments because they are not able to do much with this capacity in DRAM and NAND. They will not only take away market share from Numonyx but also from Spansion and others. As for Spansion entering NAND, do you really think they are any match to Samsung or Toshiba with their Mega-Fabs? Even Hynix and IMFT are losing money in NAND against Samsung and Toshiba despite their years of expertise at fabricating semiconductors and ample financial resources. Do you realize that Toshiba/Sandisk together spend a little less than twice the current market-cap of Spansion in capex every Q on NAND? Spansion will not be able to compete in NAND period.
This is of course only a nuance but an important one imo. Although I can't completely agree with your prior post that the cross licensing deal makes the propietary technologies de facto standards, I do agree with this earlier more carefully worded statement. There are actually a few well accepted standards. Imo it would need a commanding share (>>30% rev wise) in the traditional NOR space to make it a de facto standard. Mirrorbit is getting there too.
Cross-licensing deal between Toshiba and Samsung is not making it a de facto standard the way you describe it but it is facilitating it. OneNAND is a better solution and is able to deliver what cell phone makers are looking for at a much lower price and that is what will make it a de facto standard if both Samsung and Toshiba start pushing it without having to pay L&R to each other. Spansion is too late to the NAND party and they need to worry about survival rather than aiming to compete with Samsung and Toshiba in NAND. They are not going to profitable in 2008 or even in 2009.
My view is that both Samsung and Spansion are reasonably likely to continue to gain share in the NOR space at the cost of Numonyx plus the smaller players for some time to come. The cross licensing deal is highly important, but not big enough of a thing that would make Spansion's share gains a thing of the past all of a sudden. The basis for my assumptions here is that I think that technologies like charge trapping and SONOS will win over floating gate and that Numonyx is moving to the newer technologies way too slow. Just my 2c...
I disagree. NOR market will decline or remain flat at best and what that means is that industry participants are going to pounce on every design opportunity and kill each other with lower prices to just stay alive in the business. Eventually, some capacity will have to be retired and the weaker players without resources will have to fold-up or get merged with someone stronger. Investors at Numonyx are going to cut expenses aggressively and try and make it profitable so that they can spin-off as an IPO in a couple of years and make a good RoI for their 150mm at risk in the project. Competition in NOR is going to be fierce and making money will get harder and that is the true picture of NOR as I see it and all IMHO. |
