Re: AMD's roadmap seems much more interesting. I'm not sure if they can really sell 50 million processors next year (9 Q1, 11 Q2, 14 Q3, 16 Q4), but the profit at $100+ ASP is mind boggling.
Pravin,
I've been thinking about this for a few days and came to some conclusions for the next few years.
I'm basing this off of really only one assumption, the world economy won't collapse. The other half-assumption is that Intel executes well.
I think Intel hasn't been a position to be forced to get products out quickly in a long time. Now, AMD is pushing them to release products before they are ready (both companies did this for 1 Ghz). Under normal circumstances, I think Intel would've been flawless for the past year. Now they are suffering the "IBM" effect. This is what IBM has been going through for a long time as they have been under pressure by smaller more nimble competitors. To compete, you have to act as quickly as your more nimble competitor and speed up roadmaps, which sometimes causes you to release faulty products. I think more of Intel's current problems are related to unexpected competition from AMD (causing them to speed up roadmaps) than from poor execution off of original expectations.
I think this is only going to get worse.
I believe Intel that demand is very high, making supply tight. AMD selling 50 million processors next year represents .83% of the worlds population buying an AMD based computer (50 mil/6 bil population). Who knows how much business could come from China's population which is 4 times the US. Who knows how much business could come from Europe because of Dresdan (you're damb right I'd be marketing AMD as a European company over in Europe, ie patriotism). The US market should see average demand I think. The US is small compared to some of these other countries that are starting to catch up in technology. If the world economy remains strong both AMD and Intel should be able to sell everything they can make.
Then on to Intel's "supply/demand" problems...
Austin: Durons on .18um Al will be a big seller, although I don't expect much to the bottom line.
Flash: Not an issue. Sold out longer than most of us will probably hold AMD stock.
Dresden: Dresden is the key, bottom line.
Dresden is running .18um copper. We're expecting .15um or .13um over the next few years. Dresden has proved its weight in gold to me already. I wasn't expecting T-birds to blow away Athlons in benchmarks. It is my understanding that copper helps you shrink below what aluminum allows. A small performance gain at the same shrink is just an additional benefit. The real gains will be seen when AMD goes to .13 on copper. So how far ahead is AMD in technology? A tough question, but I'd say 7 days. This is June 5th when Dresden opened. We will know exactly how far AMD is ahead when Intel opens a copper fab in volume production. So this number will grow every day. My WAG is that AMD is one full year ahead of Intel.
From Intel's front we will see the same. Demand should be strong, but Intel isn't ready to handle it. Intel will also be having a lot of down time. Conversion to .18um is still ongoing, conversion to .13um, conversion to willy, etc. Then will start Al to copper conversion. It is my understanding doing copper and Al in the same plant is a no no. So to convert to copper will mean big down time. AMD should benefit by having Dresden running full steam this whole time.
Anyways, that's how I see the next few years. There are things I'd like to see done better. Intel's recent execution has made AMD's seem brilliant, so it's hard to complain.
chic (who knows Dresden copper is the key) |
