>>Also why is this simplistic approach so much better than the way analysts arrive at
their numbers. Something is wrong if you really believe Spot and contract sales
across many different memory products by MU can be reduced to Achilles quotes.
You know... my feeling all the while is that you are not interested in answering
anyones concerns...but just getting into a pi$$ing contest. I have seriously asked this
question many times for my own edification and never got a satisfactory answer. <<
dave, valid question. i've answered this question before in many different ways. let me summarize:
1. it has been an excellent predicter the last two qs. does that mean it will be a good predicter this q? not necessarily. does that mean it will be bad this q? not necessarily. an inference can be drawn that whatever achilles tracks is right in line with what mu sells, though. you do need a slack factor of about +-5%, imho, and imho only.
2. when tom kurlack issued his downgrade and mentioned micron's dram pricing, he quoted the achilles number to a "t." so at that point in time, tk's number jived EXACTLY with achilles. the probability of having completely different numbers all q and then having the same numbers at report time has got to be very low. this isn't a guarantee that the same is true going forward, though. but, the inference is clear.
3. jerry denkura, our local dram trader, has always been with +-5% of achilles anytime he has quoted dram pricing for mu. he knows because it is his job.
4. the press has recently reported dram for mu being very close to the number that achilles posts.
add it all up and i know where i'm going to bet my money...
btw, there is lots of different dram with different prices. although i don't know for sure, i'd bet their is one standard chip that is probably the vast majority of sales and that is what achilles tracks and mu sells. the other stuff appears to be specialty chips, hence, the higher price tag. perhaps jerry could weigh in here.
another caveat brought up by peter. mu is going more and more to sram - this isn't followed by achilles and will have some affect on mu's asp and profits.
however, i know that the sram market is also collapsing under the weight of increased supply and 15 million more ics a month from mu is only going to make the situation much worse. sram is priced very closely to dram.
since the sram ramp is new, mu is less efficient at it. mu also has warned of testing problems and inserted a caveat in their 10q.
this may not convince you. however, you WILL NEVER get absolute proof of anything. if you had to prove that you had a job before going to work then you'd never go to work in the morning.
achilles isn't 100% foolproof. but when you sum it with all the other data, you will realize that it is a very good indicator of current dram pricing.
btw, I'D LOVE to have you on my jury if i was accused of a crime. :-) |
