"Price War" thoughts - ASP development
First, let me say that this posting will be quite a bit
long and will also contain some maths. Don't be frustrated,
its always very simple and I hope that I could write
it so, that everyone could easy follow the lines.
Why did I made a spreadsheet? You know, that I made many of
these things in the past, but because AMD didn't gave enough
data outside, it was even harder to follow the ongoing de-
velopments. But because we heard in the last weeks so many
comments that AMD will be doomed after the NGA launch and
Intel will crush our margin, I said to me, that I have to
do some math just to get a better feeling of possible num-
bers. I also hope, that I could give some light into this
topic. I wrote two posts on our german board and these
thoughts I will summarize here too.
First I will made a "Basis Scenario" and will then go to a
"Worst Case" ... the better ones I will let for you. Don't
nail me to this or that number. I had to do guesses and its
not that important whether one number is 5$ higher or lower
or if the units are 500k more or less. We don't know the true
ones and everyone could analyze the development by its own.
So I hope I could start now and I will appreciate if others
are doing the same things. The final numbers in the con-
clusion will show one main point and it will not be VERY
important if the numbers are by a fine margin correct, they
show the message and thats fine for me.
I gave AMD 5 segments. Each segment will get an segment-ASP
and a segment-volume-number. First I give my starting point
numbers from Q4-2005 and then I will go on and make my
guesses. Again, don't take this or that soooo seriously, the
amount of parameters is huge and we could ALWAYS change ALL
things. So for Q4 a made these numbers for:
A64-Sempron-X2-Mobile-Server
5 - 5,5 - 0,5 - 2 - 0,3 Vol. -> 13,3
the segment ASP are
100 - 50 - 300 - 125 - 500 -> ASP = 99,6$ -> Rev. = 1325
Now I will go on with this volume numbers:
Q1: 5 - 5 - 1 - 2 - 0,4 = 13,4
Q2: 4,5 - 4,5 - 1,5 - 2,5 - 0,5 = 13,5
Q3: 5 - 5,5 - 2 - 3 - 0,6 = 16,1
Q4: 5,5 - 5,5 - 2,5 - 3 - 0,7 = 17,2
-> volume for 2006 = 60,2
You could see, that I didn't anticipate a HUGE DC ramp just
to be on the safe side. Lets take a look at the DC % numbers
overall and compared to only desktop:
Q1: 7,5%/9,1%
Q2: 11,1%/14,3%
Q3: 12,4%/16%
Q4: 14.5%/18,5%
Just to start, I let the Q4-2005 segment ASP in the first
stable to get a feeling, what these shiftes will do for the
whole ASP.
Q4-2005 = 99,6
Q1: 111,9
Q2: 125
Q3: 127,3
Q4: 133,7
Keep in mind, that these numbers are unrealistic, so I made
further changes to the starting points, just to WEAKEN AMDs
position.
A64 Q2 ASP 100->95, Q3 = 90, Q4 = 90
Sempron Q2 = 49, Q3 = 48, Q4 = 47
Mobile Q2 = 120, Q3 = 130, Q4 = 140 (X2 Effect)
Server Q2 = 480, Q3 = 460, Q4 = 440
So, we have more realistic numbers going forward and I'm now
calculating the DC (X2) segment ASPs, just to let the overall
ASP stable at 100$, which is the starting point in Q4-2005.
The given volume numbers didn't change, only the segment ASP.
So, with these numbers, the X2 ASP could drop to ... just to
let AMD reach 100$ in each quarter!!!
Q4 - 2005 = 300 (starting point)
Q1 = 140
Q2 = 110
Q3 = 120
Q4 = 100
You could see the amazing drops in the DC line, which AMD
could "easily" face - no problem. I couldn't say this often
enough - keep in mind, what this means. With 100$ DC ASPs
in Q4, the overall ASP would stay at 100$ with these circ.
So, to be even more unfair to AMD, a made a bad "worst case"
scenario. Changes are:
Starting Point Q4-2005 (ASP)
A64 100->105
Sempron 50->50
X2 300->260(very low)
Mobile 125->120
Server 500->500
Now the segment ASP will change to:
Q4->Q1->Q2->Q3->Q4
A64: 105-105-97-90-82 (look how bad this is)
Sempron: 50-49-47-46-45
Mobile: 120-115-120-120-120
Server: 500-480-460-450-420
Again, I'm calculating the X2 segment ASPs, which leads to
stable overall ASPs at 100$, with the given volume numbers
from above. They could change and vary, of course, but I
have to take ones ... the new "needed" X2 ASP would be:
old -> new
Q4 - 2005 = 300 -> 260
Q1 = 140 -> 150
Q2 = 110 -> 120
Q3 = 120 -> 140
Q4 = 100 -> 150
So, you could see, that from the very low 260$ X2 ASP point
in Q4-2005, the number could drop even 40% without sacrif.
the overall ASP. Now you could say, with this low segment
ASP, the volume number is too low, just because the demand
will shift to more DC. Thats correct, but I made bad choices
to show whats possible in worst case, but just these numbers
show, how big the ASP potential for AMD is. So I made two
more equations, where I upped the DC volume number in Q4-2006
and decreased the A64, just to let the overall volume number
stable. First I go from 2,5Mio. to 3Mio., so the segment
ASP could drop from 150 to 140. If I go to 3,5Mio. in Q4, the
segment ASP could drop to 130, which is 50% less then the
very low 260$ guess for Q4-2005.
Do your own math, try to think about these implications and
try to analyze how Intels prices will and could destroy our
margin level. I will go that far - its not possible as shown,
when Intel doesn't want to destroy the own numbers in a huge
way.
BUGGI |
