CPU Market Shift

[wolverinero79]

Over the past decade, we've seen a fun battle between AMD and Intel over the home and business consumer. Lots of interesting new technology and lots of new products that have computers whizzing along.

However, what happens now? Intel believes 22nm is possibly with current technologies and is rumored to have working 45nm parts, but they are driving costs lower and seem to be focusing on non-US markets. AMD is doing interesting things like investing in ATI, building new fabs (Intel's building 2 new ones as well, but out of technology necessity, not capacity - I expect them to close others), and trying to take more MSS.

After 2007, will we see processors turn into more of a commodity market where manufacturing strength is valued over squeezing that last bit of performance out of a chip? I mean, Core 2's huge leap is one thing, but how many times can one of the companies increase performance that incredibly?

Plus, consumers could definitely get into the space of "who cares" in terms of PC computing speed. If you ask 50 people on the street if they care more about how many songs can they fit on their iPod vs how many FPS they can get on their computer, guess what they'll pick ("FPS? What's that?").

The stock analysts already treat the companies as involved in a zero-sum game. What do you guys think? Do you believe that the CPU manufacturers will focus more on consumer products, or squeezing performance out their chips? Certainly going to something like 80 cores means that although processor prices may retain their ASPs (they actually haven't already though...they're ticking downwards), the per die profit will definitely be shrinking - which is fine as long as you keep shrinking the dimensions, but when you can't go smaller, what then?

Exciting time, but very unknown and very suspenseful.

 

[BGP_Spook]

Please Note: I am making these percentages up on the fly using my impression of the computing space as justification.

Plus, consumers could definitely get into the space of "who cares" in terms of PC computing speed.

They are already there.

Take a broad look at the computing spectrum.

50% of users are the grandma, the kiddies, and Joe Schmos who only want to check their email, browse the net, and deposit their mp3s.

They buy Dells or whatever is cheap. They don't care about the latest technology developments, the move to 64-bit computing and problem of reprogramming software to better use multiple cores. All they know is what they want to know.

Probably 40% of users are into computers because they have to be. They work on whatever is cheap use specialized software or Office. They could care less about how many Gigaflops their video card can do. They don't know and don't care to know the difference between on-board and discrete video cards. All they know is what they have to know, which typically isn't much.

Of that last 10%, half are like most of those who read THG. They worry about this stuff, not because they have to but because they enjoy it. Not unlike how some people enjoy hearing about the latest fashions and romances of Hollywood starlets.

The other half of that 10% are the people who are concerned because technology hasn't quite met their highly specialized need, yet.

It is only this 10% that still cares about their hardware.

As for where I think the CPU market is headed, I think it is headed in this direction.

Those who still care about their hardware are going to fragment off and be addressed separately from the average user. The PC space is breaking down into categories of those who's needs have been met by past computing technologies and those who care--for whatever reason--about future computing technologies.