October 24, 2000
Advanced Micro Devices
Posts selected for this feature rarely stand alone. They are usually a part of an ongoing thread, and are out of context when presented here. The material should be read in that light. How are these posts selected? Click here to find out and nominate a post yourself!
How Pessimistic are the Analysts?
The biggest frustration for me as an AMD investor is the severe pessimism of the analysts. BeRealCNBC's comments that the analysts get it "right" as much as us on the message boards really bugs me.
While it is true the analysts were very close to the mark last quarter, it is only because AMD gave very specific guidance. AMD did miss that guidance overall (specifically ASPs were abysmal), but still beat the low analyst numbers. That's the good thing about low analyst targets--I definitely prefer the analysts to have at least a little skepticism! I also believe the analysts will be close to the mark in Q4 given AMD's reduced guidance (AMD should come out a few pennies ahead, but not a blowout).
But what about 2001? Are the analysts right in believing that AMD will only earn $2.55 in 2001? Here it is interesting to take a look further back in history to see what the analysts were predicting around this time last year. In mid-Q4 last year Salomon Smith Barney predicted the following for FY2000 (numbers are adjusted for split):
12-month price target: $16
Amazingly, we effectively reached his price target (16 3/16) just a few days ago! But suffice to say his earnings estimates were not even in the ballpark. To date this year AMD has already earned $1.82, more than double his estimates. Remember, this estimate was given BEFORE the slowdown was predicted by SSB in the summer. Both Intel and AMD had guided that demand would exceed supply throughout the year (which has turned out to be false--Intel had an inventory buildup this quarter). Imagine what SSB would have predicted if they thought there would be a slowdown or price war? Perhaps $0.50/share for AMD?
At the end of December, SSB did upgrade their earnings estimate to $1 and their price target to 17 1/2, but as you can see, their earnings were still vastly under the real numbers, well under half what AMD will actually earn (around $2.50 per share).
Why do analyst estimates matter? Well, presumably investors base their decisions on predicted growth rates for companies. Those growth rates come from analysts. Right now AMD appears to have earnings growth next year of 2%. Even if you factor out the higher tax rate in FY2001, AMD's earnings growth before taxes is still only 20%.
So let's look at that $2.55 FY2001 estimate again. Are the analysts actually doing any real work to make that estimate? Do they make spreadsheets and run different scenarios for ASPs, production capacity, flash growth rates, etc? Or are they just simply guessing? What would have to happen for AMD to get earnings that low in FY2001? Are the analysts predicting doomsday scenarios for AMD in 2001? In my spreadsheet, these are the numbers I have to plug in to get $2.55:
Analyst doomsday recipe?
CPU unit growth: 6% per quarter
CPU ASPs: flat at $90.5
Memory revenue growth: 7% per quarter
Additional share dilution: 60M shares
FY2001 EPS: $2.56 (2.4% growth)
Note that all the numbers above are beyond conservative, they are downright catastrophic. Let's look at each one:
CPU unit growth of 6%
AMD's actual CPU units are increasing by well over 10% per quarter as Dresden ramps up. Last quarter units increased by 11%. A unit growth rate of 6% assumes 8.5M CPUs per quarter average next year. Given that Dresden alone at .18 process technology can produce around 12M Athlon CPUs per quarter fully ramped (assuming yields of 70%), and Dresden will be at 50% ramp in Q1, Dresden alone can average around 9M units per quarter throughout the year. That's entirely ignoring the transition to 130nm process! Now we can wonder whether all that production capacity of CPUs will sell out, but consider that AMD has zero competition at the mainstream level (P4 will be too expensive for a long time).
ASPs flat at $90.5
This is ludicrous to me. Yeah, I know I was one of those that was hoping for $100 ASPs this quarter, so perhaps I am a little over-optimistic. AMD is entering new markets which should improve ASPs, in particular high-end mobile and low-end server markets. In addition, we will see chipset competition next year (new chipsets from AMD, ALi, Via, Micron, etc) which should bring down the price of chipsets and motherboards (right now Athlon mobos are around $30-40 more expensive than equivalent PIII mobos). If we can achieve a $10-20 reduction in mobo cost (which I believe will happen), AMD could potentially ease up on lowering prices of its CPU's.
In Q3, AMD sold a considerable number of Durons (22% of total sales) and a lot of K6's (46% of total sales), with only 32% of sales high-end Athlons. I would guess that K6 ASPs are around $50, Durons around $60, and Athlons around $170 (these numbers add up to $90.5 ASPs, consistent with Q3 results). Let's say in Q1-2001 AMD sells around 65% Durons and 35% Athlons. Assume Athlon ASPs drop to $150 and Duron ASPs stay flat at $60. The end result would be $91.5 ASPs. While not great, this is still more than $90.5! And I can easily imagine Duron ASPs rising as AMD introduces a 1GHz Duron in Q1 next year. Also, I believe that the percentage of Athlons will increase in H2-2001 because Dresden will only be producing Athlons and Athlons will move to .13 micron, further increasing capacity.
Each $1 rise in ASPs corresponds to a roughly $0.056 increase in earnings for FY2001 in my model (with the other low assumptions above).
Memory revenue growth of 7% per quarter
AMD's actual flash growth has exceeded 10% every quarter for the past year. In Q3 of next year, a new flash fab comes online. By the end of next year, Jerry expects flash bit capacity to double from 700Tb to 1400Tb. Flash prices per bit are dropping at around 4% per quarter, but capacity and average bits per package are going up rapidly to compensate. To achieve growth that slow, AMD would have to have capacity growth half of what they plan or prices would have to start dropping 3 times faster than they are. Considering that AMD's flash is mostly sold through contract prices that do not change, that is unlikely.
Share dilution of 60M shares
This I can almost believe. This is a similar share percentage dilution to what occurred to fund Dresden. AMD does have a shelf offering for $2B. The above would correspond to a stock offering of $1.5B at $25/share or $2B at $33/share. I certainly hope AMD tries to avoid diluting the stock any more than necessary! I am hoping that AMD only dilutes the shares by around 40M next year.
My conservative analyst model for FY2001
This is what I think a reasonable conservative analyst that has run the numbers should predict. AMD will likely easily exceed all of the targets below.
CPU unit growth: 7% per quarter
CPU ASPs: average $94 for the year
Memory revenue growth: 9% per quarter
Additional share dilution: 60M shares
FY2001 EPS: $3.18 (27% growth)
Sadly, the highest analyst estimate on Yahoo for next year is $2.89, well below my model above.
My reasonable model for FY2001
CPU unit growth: 8% per quarter
CPU ASPs: average $97 for the year
Memory revenue growth: 10% per quarter
Additional share dilution: 40M shares
FY2001 EPS: $3.87 (54% earnings growth)
My fantasy model for FY2001
CPU unit growth: 12% per quarter
CPU ASPs: average $110 for the year
Memory revenue growth: 12% per quarter
Additional share dilution: 40M shares
FY2001 EPS: $5.96 (138% earnings growth)
Note that in order for the "fantasy" model to come true, several miracles would have to happen. I'll leave those up to your imagination. :-). Those are my thoughts on 2001. We'll see what AMD has to say about 2001 at the November analyst meeting...
Read More Posts by This Author
Go To This Post
More Recommended Posts