Replies to post #39999 on Advanced Micro Devices Inc (AMD)

[CombJelly]

"looks more like they’ve been scrambling to keep from falling back into the loss column."

Probably fair. Anectdotally, it looks as if Hpaq and eMachines both lowballed A64 demand and ran out of machines about a month ago. It's hard to gain marketshare when the OEMs won't build the machines to be sold. Hpaq, in particular, seems to have been caught with their pants down in another area. Apparently they felt the r3000z and its twin the zv5000z was only going to be of interest to a few bottom feeders. But since its introduction it has expanded its already long list of options to encompass higher resolution screens, bigger disks, more memory, etc. The lead time has also stretched out on occasion. Usually it is about 2 weeks, but it can be 6. It wouldn't surprise me if a higher end version with a better video card were to pop up soon.

If 90nm pans out, then AMD can turn to scrambling the competition...

[JackDye]

I am still playing with the numbers based on my understanding of what Q1 should have looked like and it raises some interesting questions.

If the 100% margin on the previously written off products had not been in the Q1 numbers, they computational group would have reported $571M in rev and closer to $58M in operating income ($67 reported less 50% normal cost of sales on the $19M).

Looking at the Q2 results, the $554M revenue ($17M less q/q)drove the same $58M in operating income. This says despite the lower revenue, cost decreases occured and in fact margin expansion occured that offset the lower revenue. This assumes no other one-timers of which we are not aware.

On the flash side, the q/q revenue increase of $45M drove a fall through of $31M in operating income. That is a hefty 70%, well above the normal gross margin. That also says margins in that space are expanding nicely.

So where does that get us? If we do "moderately" better than normal seasonality of 7-9% in both business (say 11%), we will end up adding $61M in computational revenue and $74M in flash revenue for a total of an incremental $135M.

Given Opterons and A64 should be increasing and statement from the cc that AUPs should be going up, I will use a 60% margin on the incremental computation revenue. On the flash, falling through at the 70% of Q2 seems pretty rich, so I am going to back that off to 50%. This gives us incremental margin of $37 in comp and $37 (funny enough) in Flash.

Offset that with the 10-15% increase in SG&A (I choose 12.5%) and that comes to $22M. Then give up the Spansion share of increased profits of $15M ($37M x 40%). That gives us a total offset of $37M again - just the way the numbers work...

Anyway, that says we get to keep the incremental $37M or basically 10 cents. Add that to this quarters 9 cents and we get a total of 19 cents for Q3.

That says there is no improvement in the Other category or o,i&e compared to this quarter. Accelerating cost savings due to 90nm transition, richer mix, or anything else would improve upon my assumptions.

Jack.