The Athlon 64 is admittedly generally better than the Pentium 4 is most situations and applications, but AMD is a vastly inferior company and that has not changed in the least. For example, Intel offers much wider solutions including chipsets, motherboards, software, marketing dollars, etc... So, it is not just the microprocessor but also the company you have to deal with. On top of this, having fewer different parts makes life easier for the company for logistical reasons, and of course they receive a lot of incentives from Intel to stay Intel only.
Normally, these rumors have no truth at all, and the performance of the microprocessor has not made any difference. This is not the first time AMD has had anything faster than Intel. If you remember the Athlon and the Katmai you had a similar situation, maybe even greater in the sense that the Athlon was faster in even more applications. That is, it is fairly easy to find an application where the Pentium 4 is faster, but finding one where the Katmai was faster was very difficult. It made no difference at all. On the other hand, the Athlon was a pig and the Katmai was much lower power and much smaller. Pentium 4 is a pig, and bigger as well as being slower. So, probably the Pentium 4 represents the weakest competitive position Intel has ever had.
On top of this, the Katmai never hurt Dell. They kept growing, and it was irrelevant. If you look at Dell's recent performance vis-a-vis Hewlett-Packard, it is relatively poor. Their recent performance has shown much slower growth rates than have been historical for Dell, and not too coincidentally AMD has shown market share growth. So, it seems fair to assume that the inferiority of the Pentium 4 is impacting Dell right now more than any other competitive position Intel has had with AMD.
The other issue is the successor to the Pentium 4. It should be better than the Athlon 64, also a relatively inferior processor that artificially looks good because the Pentium 4 is so bad. When compared to the Pentium III and derivatives, particularly for multi-cores, the Athlon is still a pig that offers very poor performance for the power it uses. I think most people would much prefer having a Pentium M to an Athlon 64, particularly when they have the same features (for example x86-64 mode). The perform is roughly the same despite the Pentium M having a very low power ceiling. On top of this, the P8 is rumored (I have seen it mentioned in two places, so who knows if it is true?) to have four integer pipelines, which if true should give it superior performance to the Athlon 64. At any rate, it will be better than the Pentium M and the Pentium M is already a better processor than the Athlon 64 for most people. It costs too much though.
So, if Dell can wait, the competitive situation will change once again in the processor field and Intel should minimally attain parity. Their other advantages will remain and AMD has shown no effort to close them. However, if Dell goes with AMD, that has to say something negative about the P8 or minimally something very positive about the AMD brand name. It may be that there is a decent percentage of people that now perceive AMD as better and have brand loyalty. If so, even if the P8 is better, Dell may find it still makes sense to switch.
Despite all this, I think Dell will remain Intel only for the near future. I could easily be wrong because their slow growth is a new situation, but there are so many advantages to being only Intel based I think they will outweigh the current disadvantage in product mix. If it perpetuates to the P8 as well, I think Intel is done. If the P8 fails the way the P7 has, Intel is going to be a great company to short sell. It still is difficult for me to imagine how it could be so bad though, but who guessed the P7 would be so bad before it was released? I certainly did not.