Sunday, December 14, 2008

At least this economic recession has had one benefit - it has flushed out a number of financial cheats and scoundrels. The greedy have been found out and in many cases dealt with. Companies that used financial tricks have run out of options. Companies are going to have to make money the old fashioned way - by creating something of value. I've never understood a system that rewards people for moving money around without creating any real value. But that's just the engineer in me.
As the agony of AMD's spiral down continues, it couldn't have helped when the San Jose Mercury columnist Chris O'Brien put AMD on his list of companies that are in serious trouble.

Why did AMD fail to sustain its success after its triumph over Intel with Opteron and Athlon 64?

First off, AMD underestimated Intel's ability to rebound from the Pentium 4/Tejas debacle. Intel is a very resilient competitor with a lot of resources and a process and manufacturing group second to none. Intel, to some extend, was also lucky. It had a very sharp team in Israel that was working on lower power processors that could be enhanced for more performance. And so Conroe (Core 2 Duo) was born and the start of the long slide down for AMD.

AMD's management conceived of a grand idea of merging the raw performance of GPUs with the general purpose CPU (Fusion). It's an elegant and exciting concept. It was design to put AMD back ahead of Intel. The problem was that AMD grossly overpaid for the number 2 GPU vendor, ATI, and AMD's execution of the Fusion development concept has been slow and painful with a number of canceled versions.

Meanwhile, Intel has embarked on it's own GPU and its own version of Fusion. Right now, it looks like Intel might have a Fusion part in the market ahead of AMD. This was clearly a project that AMD could manage with too many new teams trying to work together. It just hasn't jelled.

So, AMD struggles to keep afloat by selling off anything it can. The latest part of the company to be jettisoned is the fabs. Before that it was Legarity, Spansion, Alchemy, and its PLD division.

Frankly, I think AMD's given up competing with Intel. AMD will try to hold onto enough market share to keep in business, but right now, there's very little upside until 2010 or 2011 when 32nm Bulldozer-based cores hit the market. No, instead, AMD management is going to try to eek out some minimal self respect by taking on NVIDIA as much as possible. But we're as resourceful and has adaptable as Intel. So I see no
opportunity there. And, afterall, ATI was rarely very profitable, so even a healthy ATI can't hold up a sick CPU business.

There a re a lot of companies that don't want AMD to go out of business, but at the same time, I don't see these companies digging into their own pockets to save AMD.