Fire At Hynix FAB May Bump DRAM Prices 77
Lucas123 writes "A fire that engulfed portions of a major Hynix memory FAB in Wuxi, China earlier today did not do as much damage as reports originally claimed and the company said it expects to be back in production soon. According to a Hynix statement, the fire occurred during equipment installation at around 16:50 Korean time and it was extinguished in under two hours. The company said while published photos showed the FAB facility surrounded by smoke and engulfed in flames, 'the damage is not as severe as it seems as the smoke created was because the fire was concentrated in the air purification facilities that are linked to the rooftop of the fab.' The company also said that there is no material damage to the fab equipment in the clean room, and Hynix expects to resume operations in a short time period, so overall production and supply volume should not be 'materially affected.' Even so, the spot price of DRAM is expected to leap as a result of the news."
Crap ... (Score:5, Interesting)
Anybody who bought memory in 1994/1995 knows about this kind of thing.
I remember spending $600+ dollars to get 16MB of RAM back in the day, and that was considered a good price back then.
Of course, cynically I believe companies will latch onto anything which allows them to claim increased scarcity and jack up prices.
And that there is a spot market for DRAM tells me that, once again, speculative investors are fucking it up for everybody -- kinda like oil, where the price goes up because people believe that other people believe the price will go up, and not for any actual market factors.
This is why we can't have nice things.
Sumitomo all over again (Score:5, Interesting)
Anybody else remember when the Sumitomo epoxy resin plant went up in the early 90s? RAM prices TRIPLED OVER NIGHT and remained high for the next two years, even though chip manufacturers had 6~10 months of the product stockpiled, Sumitomo had 6~8 months worth stockpiled (and in fact, weren't producing for that reason at the time), and several other chemical companies could have been up and running to produce the resin (which sold for $6/lb and THAT price never changed).
In the end, the plant came back up ahead of schedule, and nobody else jumped in because at $6/lb, it simply wasn't worth it to make the stuff, which was use not only in RAM chips, but a lot of other chip packages as well (oddly, none of those other chips went up in price).
In short, the price jump was artificial, had nothing to do with supply and demand, but simply companies taking advantage of news to increase profit margins. ....the Thai floods from a couple of years ago are another example, though those floods did at least have a small impact on supplies (though again, the prices for platter drives remain unreasonably high)
Do they need to re-qualify the equipment? (Score:4, Interesting)
The main question is whether they had to cut power to the manufacturing equipment to fight the fire. If not, the would easily be back up and running in two weeks if the equipment was on but idle.
But if power got cut, even if the equipment wasn't damaged, all of it will need to be brought up, reinitialized, and re-qualified, a process that will most likely take days for each piece of equipment. The tolerances for semiconductor equipment are so tight (not only for pattern geometries, but also for things like high vac, chemical flow rate, temperature (try less than half a degree F variation across all 151 individual heating elements) that it takes a lot of time, effort, and expertise to hone things in just right. Expect 2-3 months for requal if that's the case.