Of course you are absolutely right that momentum can drive overpriced merchandise into even more overpriced territory, and without a doubt, the more the market gets overextended on valuation and the attendant risks rise correspondingly, the nervous investors will look for "valuation laggards" and view them as safer harbours, relatively speaking.
You can call INTC's book squeeky clean, but its franchise value does not include the absurd options drain and dillution, which must be included in your calculations if you're really trying to determine long-term enterprise value. Given that there are only a small handful of companies with worse drains (BRCM anyone?), I would not give INTC a 20% premium, I would give it a 20% demerit. Most semis that are receiving higher multiples have long-term top-line growth that is double, triple or more what INTC's is. In fact, it's hard to find a serious semi company that doesn't have better top-line growth prospects.
As for AMD, which would I buy, INTC or AMD? Well, I guess the question means I can't answer neither. If forced to choose on the basis of potential, it wouldn't be a tough choice. INTC is trading about 20x richer than AMD relative to market share, etc. and its market penetration is mature and basically done. AMD's is almost exactly the opposite. AMD has been historically mismanaged, but we are in a new era. Never before has AMD actually had a significant technology lead over INTC as it does now with a fully backward compatible 64 bit chip. INTC's failings in this department are worse than embarrassing. A distant 3rd place behind IBM's PPC 970 (currently already at market and by far and away the new market leader in performance per dollar invested), and the Hammer.
I've no sense of whether AMD will stop dilluting its shareholders (it's worse the INTC), and whether Hector Ruiz will turn out to be a superior manager.
But you should be interested in the findings of Virginia Tech, which reviewed offerings from HP, Dell, INTC, etc., etc. and found none of them could hold a candle to an 1,100 unit supercomputing cluster from Apple featuring IBM's PPC 970 chips. In a few weeks it will be certified as one of the Top 10 supercomputers in the world -- and doing so at a cost of scarcely $5 million is mind-blowing. http://computing.vt.edu/research_computing/terascale/
Intel is a master marketing company but it now faces challenges like never before at the top-end. The Itanium has had a checkered existence, and the CISC dead-end has been more or less reached. The hybrid RISC architecture, leaves Intel in the awkward position of telling its customers that it was lying all these years it has been telling them clock cycles are the way to measure performance -- as it tries to convince a skeptical public that a 1.2 Ghz Itanium is actually more powerful than a 3.0 Ghz P4.
I figure they've got maybe 12-18 months to get their act together on the Itanium. If they don't the tide will start running away from them faster and faster. IBM is a serious competitor, and if Intel can't squash AMD before it gains mometum in the 64-bit arena, then we might just see some real competition. If that happens, not only will INTC not see 58%, they won't be seeing 54% either.
Market Data 
Markets 
AI
