Keith,
The fab is producing 65 nm MirrorBit NOR flash now, and next year plans to begin producing 43 nm Spansion MirrorBit Ornand2 flash products.
I don't know how to fully asses the impact of this. It seems that the city built the fab, but Elpida backed away, and now, they have no customers. So Spansion stepped in.
I wonder how much risk Spansion is taking on by this joint venture. I take it Spansion will be selling this product in the market place, but I am curious how the costs will be handled.
Another thing that is unclear is, the current scenario, where Spansion is struggling to sell out its current capacity, where does this new capacity come to the equation?
Initial reports were that the fab will be producing NAND, 1 to 4 Gigabit sizes. I guess that is ORNAND at 65nm. Is there sufficient market where these parts can profitably be sold?
BUGGI posted a lot of calculations about competitiveness of Mirrorbit vs. NAND. If it is roughly at parity, it would still be a tough sell, since the NAND business is in toilet. The only explanation I can see here is that the Sity of Wuhan built, what turned out to be a white elephant, with no customers, and now it is willing to subsidise the startup operations, and is giving Spansion a sweetheart deal on cost. This article seems to be confirming it: tradingmarkets.com
That would play OK to the scenario, as I see it unfolding, that ORNAND2 at 45nm (or 43nm) is truly unique product that can beat NAND players. The 65nm ORNAND production can be viewed as seeding the market from Spansion point of view, and a trial run from SMIC point of getting the fab up to speed, getting ready for 43nm production.
I guess the re-tooling the article mentioned must involve getting ready for immersion lithography, since that's what's needed for 45nm ORNAND2.
Anyway, this is kind of exciting, but I would not be surprised if Q3 ends up with some additional costs helping SMIC ramp flash production.
Joe |
