27 comments

[ 4.1 ms ] story [ 99.9 ms ] thread
The fact that Apple is selling more iPads than Dell is selling PCs in a quarter was the most striking fact from this article.

Apple really is creating the future of the industry rather than just trying to keep up with the trends it sees, and that's what makes the company so valuable.

IMO those numbers are not directly comparable. You buy a PC when your old PC is getting too old; only a portion of the PC market will be buying PCs in a given month.

On the other hand, the iPad is a new and rapidly growing market segment.

In other words, one is a new growing market, and one is an older stable market. The numbers will be more comparable once the tablet market is stabilized.

Also the Dell angle avoids every other PC manufacture. Dell is only ~13% of the market tops, the volume of PCs is 8 times what the iPad moves. Also, PCs are a maturing market and it's no longer necessary to upgrade every couple years, a 5 year old PC probably still works fine for most users.

DRAM issues are probably much more to do with 2GBs being enough for most users in combination with a lengthened upgrade cycle. Moores law would suggest that todays users would buy 8 to 16GB of DRAM which just hasn't materialized.

Moore's Law relates to transistor density on an IC. As far as I know there is no suggestion within Moore's Law that attempts in any way to predict consumer buying patterns or demand.
I'm not so sure you can disconnect Moore's Law from market demand. There is certainly no physical basis for Moore's Law. It is mainly an expression of the rate of capital investment and business profit cycles. It happens that spending the Moore's Law rate on R&D and new fabs approximately matches the profits needed to build the next cycle. If customers stopped needing more memory, for whatever reason, then the business investment cycle represented by Moore's Law would quickly come to a halt. Conversely, if there were a Manhattan Project style crash R&D effort that needed better memory chips, and lack of profit was no barrier, then the density could improve faster than Moore's Law for some time.
>Moores law would suggest that todays users would buy 8 to 16GB of DRAM which just hasn't materialized.

my half year old 4Gb laptop i'd buy with 8Gb at least. Yet they all still come with 4Gb (and its happening for something like 2 years already)

So I guess this explains why ram prices have completely crashed, and SSDs are dropping pretty fast as well.

Good deal, IMO. What's funny about this situation is that Apple charges $200 to upgrade a MacBook Pro to 8gb. I just bought a new 15", and instead of giving that to Apple, I went and grabbed 8gb of ram from a local shop for $40.

Haven't RAM prices been 'crashing' since the 70s?
I'm always surprised when I buy new RAM but I just bought some a few days ago and I really was shocked. Almost exactly one year ago, I bought 2GB of Mushkin for $50. This week I bought 8GB(2x4) of Kingston for $40 ($30w/rebate). Both are DDR3 1333 so maybe it's just a large overstock near the end of it's cycle but that price drop is huge.
Well, since the mid 90s, every time I built a computer, a 'good' amount of ram landed somewhere in the $200-300 price range. So the cost of a modern amount of ram has stayed relatively constant in my experience.

But maxing out the ram in a machine for $40 certainly seems a hell of a lot cheaper than it's ever been.

So I understand that prices are falling now, while there is excess surplus. But the was I see it, down the line, after manufactures stop supplying or get wiped out, PC's will only become prohibitively expensive, and perhaps some of the comments from this earlier HN post (http://news.ycombinator.com/item?id=3295603) may hold some truth.
It goes in cycles; there's the Kitchin cycle, which deals with inventory problems, and the fixed investment cycle, which is driven by the need to replace ageing equipment. The former is more relevant to the consumer electronic industry given the rapid acceleration of technology over relatively short timescales as an additional driver of demand.
So shouldn't this mean chipmakers should be adapting their business model?

Also, incredible that Apple is selling more iPads than Dell is selling full PCs.

Can we really credit this entirely to the increased efficiency of iPads? There are lots of usages of RAM--now people carry around cell phones with more RAM than before. And data centers can make use of tons of RAM--the main constraint is its price.

Maybe this is actually a result of bad general macroeconomic conditions. Nobody wants to spend too much money, whether it's consumers or people setting up data centers. That's always been the case, but we're poorer than before now, and RAM is one of those things that you can cut down on, or stop increasing as quickly on.

This news is unfortunate.

>And data centers can make use of tons of RAM--the main constraint is its price.

price through several aspects. Among the main constraints is the number of memory slots (as we have mainly 4Gb and 8Gb RAM modules to work with). And 8Gb ECC (ECC/registered memory is the only way to utilize more than 4 slots per CPU) is only recently became reasonably priced (ie. close to 2x4GB). Once the slots are filled with 4GB modules - throw and replace with 8GB. Painful. Once filled with 8Gb modules - replace the motherboard+CPUs. With what? There is of course monsters around - http://www.provantage.com/supermicro-mbd-x8qb6-f~ASUPM3AU.ht... which require monster priced CPUs :)

So, thanks to manufacturers for the cheap 4Gb and 8Gb, yet if they built the large overcapacities for these modules instead of gearing up to 16Gb - that would explain their financial problems.

I find it difficult to believe that the iPad is the root cause of the surplus.

The timeline is all wrong:

>DRAM chipmakers including Elpida Memory and Hynix Semiconductor have lost $14 billion over the past three years

2011 minus 3 years puts it at 2008. Something important did happen in 2008, but it wasn't the iPad launch.

The fact that the story comes from appleinsider.com doesn't lend much credibility to the iPad angle either.

(comment deleted)
FWIW, it's capitalized as an iPad in all circumstances.
I believe HN automatically capitalizes the first letter.

(I may be wrong, I vaguely remember this conversation going on in an iOS post a few months ago)

> Less DRAM, more NAND Flash

These guys just silly. They don’t get difference between these types of memory.

Actually, at the clock speed of iPad’s processor you can’t run programs directly from NAND, you have to copy it into RAM and only then run it from RAM.

NAND doesn't replace RAM functionally, but if you're a chip maker, you can make money from both. If sales of the one drop and those of the other go up, a shift in strategy (less DRAM, more NAND) is decidedly unsilly. This is especially true of the two developments are related, which they are: RAM sales are dropping because there's not a lot of RAM in the iPad, while NAND sales are going up because (among other reasons) there's an SSD in the iPad.