Welcome to Anandtech Henry!
You bring up some good points, but many of your supporting reasons are not correct IMO. I'll take this apart piece by piece because it's more fun that way! After all, you did say, "Where am I wrong????."
AMD's cash position has fallen from about $1.3B to $.9B in just 9 months.
AMD cash position has fallen rapidly in just 9 months, but you have to realize that this is probably the worst
downturn for the seminconductor industry ever! That's right...ever. It's not surprising that AMD's hurting given it's exposure to the flash industry and the amount of capital it has had to expend for Dresden and FASL.
It will fall again given that their forecast Q4 loss will be about the same as their Q3 loss. (Intel's cash has fallen by the same percentage but that's attributable to stock buybacks and Intel still has $9.0B in the bank.)
Actually, AMD's outlook on Q4 has improved. They're now
saying 10% better than Q3. That won't put them in the black, but it certainly bodes well for them should the economy continue to stabilize.
AMD may get back to profitability in Q2'02 assuming the CPU recovery is for real but by that time they'll probably only have $.6-.7B in cash reserves. That may sound like a lot but it costs about $.5- 1.0B to re-tool just one fab. AMD is not as diversified as Intel. CPU's and Flash are everything to AMD and Flash hasn't shown any signs of picking up as yet.
Re-tooling a fab is indeed expensive, but AMD does have a few advantages here:
1. The majority of the equipment to be used for the .13um shift has already been purchased.
2. Dresden though built with .18um assembly lines was designed to re-tool for .13um from the beginning.
3. AMD may very well supply their newest chips through foundries such as UMC.
In their last Investor meeting, Jerry Sanders said that not taking advantage of the advances that the foundries have made in the last few years would be a mistake. As you may already know, TSMC has been producing .13um products for Transmeta and VIA for some time, and UMC is already sampling .13um. As for the comeback of flash, it will be needed if AMD plans to reach their goals. I wouldn't hazard a guess for the return to profitability of this industry if you begged me...
AMD is underpricing it's Athlon XP's relative to it's value in the marketplace. They must edge prices up and get to 2Ghz in the 1H. Executing on their roadmap must be flawless and they have no margin for error. ('01 was a good year for AMD's roadmap, '02 will be a better year for the Intel roadmap due to Intel's lead with .13 micron). AMD must do it in an environment of significant cost cutting.
I wouldn't say AMD is pricing their chips too low...at least not in the last month or so. Athlon 1.4Ghz CPU's were as low as $102! Now, even with better chips such as the XP series, they generally sell for $110 or more. The 1500+ XP line can't be had for anything less than $109 and most vendors will charge closer to $130 The XP 1800+ sells for $210 OEM, its P4 counterpart (1.8Ghz) sells for $260. That's considerably better than the what it used to be. We should also mention that AMD has a far greater line up of MP processors than it used to. In the past you were limited to 1Ghz and 1.2Ghz models. Now there are 1.2Ghz-1.8Ghz available. These CPU's command a premium over their non-MP brothers. In the case of the 1800+ MP it's around 80% more, plus these are often bought in pairs!!!
The cost structure of the chips themselves is also more in AMD's favor. The current P4 (.18um) is 217mm square. The .18um Athlon XP is 128mm. The Northwood (P4@.13um) should be around 126mm. That's barely smaller than the XP on the older process, and Intel would have to ship them in high volume compared to their .18um chips to offset the difference. The alternative would be to charge very high prices. Which is something that AMD would love Intel to do. (for two reasons)
They also have to start making some inroads into the commercial space ASAP. (Hammer in '03 may be too late assuming '02 is a recovery year for the industry). They also better pray that the flash business picks up soon. And last but not least, they need to win a big name OEM like Dell, Gateway, IBM, Compaq, or HP. (Joe's no name computer shop won't cut it) They also better hope that there are a few typical Intel fumbles along the way. (that's always a sure bet!!)
This is something I agree with you in, at least partly. Previously the server and portable segments of the CPU industry were areas AMD rarely entered, so any increased sales in these areas is a bonus and should lead to more profits given the price premiums in these areas. I don't see how Hammer could be late considering the business cycles lag the overall economy by quite a bit. It may take an entire "healthy" year of 02' before companies have the cash and comfort to begin buying in volume in 03'. I also don't think a large OEM is critical to AMD's success. It certainly wouldn't hurt, but I think AMD is targeting countries like India, Europe and farther east in their goals of expanding marketshare.
AMD has always been the "comeback kid". This time might be their last chance.
They also better beg Jerry to stay on for one more year. Jerry might be able to pull it off. Jerry can charm the socks off employees, analysts, shareholders, etc. Ruiz has big shoes to fill and I doubt he can pull it off alone.
Heh, Jerry Sanders a charmer? Well, there's a first time for everything... I do think he's a good leader despite what some may say, and his ability to gather intelligent people around him has helped AMD. The shrewd decisions AMD has made over the last 2 years is evidence to that. Whether or not it will continue is anyone's guess. I do think that Hector is a credible replacement. His background with Motorola as a profit maker is proof of that.
Here's an interesting article on him for those who haven't had a chance to meet the man in person.
So overall I think AMD has a future. Intel will probably always be the standard that AMD will have to measure up to, but that's ok. The important thing is that we have a choice and a mechanism (competition) for keeping the price of CPU's and their technological advancement where we want them to be.