Replies to post #516 on Intel Corp (INTC)

[wbmw]

John, Re: So you could be looking at 5 consequtive quarters, with an average (WAG) $500 million loss.

I don't see that likelihood of that at all. AMD's break-even point is close to $900M. Do you think they will only earn $400M the next 5 quarters? I think that is beyond realistic. However, my guess for Q3 is $650M in revenue, and a $250M loss. It should be the worst of all upcoming quarters for them, so I don't see more than a $100-200M loss average over the next 5 quarters. Don't get me wrong, though. That could still be enough to bury the company. The only way for AMD to avoid ruin is for a quick turn-around in market demand, and that's a very questionable bet, IMO.

wbmw

[wbmw]

New AMD roadmap

Patrick Tanner makes some observations

http://home.attbi.com/~ptanner/roadmaps/roadmap2.htm

Comments:
(1) Bartons (mobile and desktop) slip one quarter.
(2) Clawhammer (desktop) slips one quarter
(3) Mobile Clawhammer noted as 0.13u (previously not noted)
(4) Opteron (Sledgehammer) noted as 1-8 instead of 2-8 way MP

For the first, note this interesting post from Kapkan.

http://www.siliconinvestor.com/stocktalk/msg.gsp?msgid=17986092

The second has been discussed, but the third indicates that AMD is being more aggressive with their mobile roadmap, so I expect high power Hammer processors being shoehorned into desktop replacement systems, much like what Intel is aiming to do with their "transportable" roadmap.

With regards to the last comment, here's my theory. I think that AMD will launch Opteron at the same time as the Clawhammer based Athlon, except they'll do it without a platform ready. Therefore, they will market Opteron as a uniprocessor CPU for entry level servers, and paper launch a 2-way design at the same time, which will show up several months later. The 2-way design will go into mainstream DP servers, and compete with Xeon. Meanwhile, the 4-way and 8-way designs will eventually be trashed, due to the enormous difficulty with validation and building infrastructure for large scale systems. AMD simply cannot do everything with their current cash burn, and low volume, highly competitive ideas will be the first to go.

wbmw