Mannerisms

Should the industry gamble on boom or bust in 2007? In the old days the chip industry made the bet regularly.

One of the dicta of Gordon Moore, co-founder and former CEO of Intel, was: “It’s a peculiar business. The only sane strategy is to bet the company regularly.”

Modern CEOs don’t usually have the guts, or foresight, of Moore, and, for the overall industry this years nice gentle projection of low double-digit growth means most execs won’t have to take too much of a bet.

But, for one segment of the industry, 2007 is going to be a massive gamble. That sector is the NAND flash market.

Everyone knows that prices are falling like a stone; everyone knows that capacity is being built like there’s no tomorrow, and everyone is banking on the fact that new applications will soak up the over-production and underpin the price.

The Intel/Micron joint venture, IM Flash, is currently bringing up three 300mm fabs for NAND flash, says it will start a new fab in 2007 in Singapore and that it will build one a year thereafter.

The STMicroelectronics/Hynix joint venture DRAM/NAND fab in Wuxi, China produced its first wafers in Q4 2006; the new SanDisk/Toshiba fab starts at 2,500 wafers a month in 2007; SMIC the Chinese foundry, has started sampling NAND flash; Samsung plans to build eight 300mm fabs between 2005 and 2011 and has prototyped a 32Gbit NAND on 40nm; Spansion is producing a 4-bit-per-cell NAND which could, potentially, revolutionise the industry’s economics.

Yet: “There’s going to be an almighty collapse in the NAND market”, says Andrew Norwood, senior analyst at Gartner Dataquest.

“New applications can change that,” says Ralf Ebert, manager for flash marketing at Samsung, and Samsung is promoting some e.g. a 36GByte solid state disc to replace HDD in ultra mobile computers, and a hybrid HDD incorporating NAND flash, supported by Microsoft’s new OS Vista, which allows a PC to boot up as quickly as a TV.

Added to that are new applications in portable video products and there is no doubt that demand will grow faster than last year’s 60 per cent for units and 200 per cent for bits.

But will it grow as fast as the increase in capacity? That’s the big bet of 2007.

Execs are talking much like their counterparts did 25 years ago. When ten players were each investing to take 20 per cent market share, there was always one response: ‘elasticity of demand’.

If prices go lower, they argued, lower prices will foster new applications, and the more chips people will buy.

It was a fine old argument in the old days, but nowadays, when chip companies have stockbrokers’ analysts crawling all over them with a single obsession – financial performance - it is problematical whether companies will absorb below-cost prices for long.

The Big NAND Bet is a fine, ballsy, feisty bet which will be one of the things which will make 2007 a very interesting year.