Re: At the same time, the Intel folks moan and groan about how AMD should simply give up and go back to a niche market with 20% MSS, where they can have sustained profitability.
>> I don't believe this at all.
First of all, I am a proponent of suggesting a scenario where AMD may find profitability if they set their sights on a lesser MSS ambitions, but let me tell you why.
I am a believer in the non-linearity of market share vs. effort. Basically, that means you can spend one unit of effort to earn a point of market share, but in order to earn 5 points of market share, it may take >>5x as much effort. In AMD's case, I think they have enough design wins and vendor support to keep 20% of the market with very minimal effort. Vendors want dual suppliers, and let's be honest: 20% is not a niche... it's a pretty sizable chunk!
But look how hard AMD has to fight, just to get to 25% of the market. They reached 20% almost effortlessly when they were ahead, but they also maximized margins, and won some coveted OEM designs (which they will keep now regardless). But so far, the climb the 25% has been rather steep and costly for them. Last year, when Intel had only a fraction of their production shifted to Core micro-architecture, AMD went from hundreds of millions in profit from their CPU division, down to almost a loss. And this quarter, they will most likely post a complete loss.
And does it get any easier to get to 30%? No, it gets harder. It means AMD has to stuff even more parts in the channel, convince OEMs to commit to even more designs for a less competitive product, and spend even more market dollars to get long time Intel users to switch (which they are less likely to do now that Intel is ahead).
You see what I mean? Getting to 30% may kill AMD in the process, but it doesn't necessarily win them anything. It is a fictitious goal with no prize to be awarded at the end. It just means that if AMD gets here, they can then aim for 40%.
This is why I think the whole race for unit market share to be a bit idiotic. The real goal should be for revenue share and/or margin dollars. AMD may actually increase revenue share by decreasing unit share (depending on how steep the curve is for them right now, relative to how much investment it is costing them to go for the next point of unit share). This would obviously involve raising ASPs to a reasonable level, and putting more focused efforts into the markets where AMD still has somewhat of an advantage (such as investing in low power designs, as opposed to idiotically trying to market and justify the existence of Quad FX).
Right now, AMD is casting a broad net that doesn't necessarily show an advantage anywhere. It's unfocused, reactive, and it promises much while delivering little. Slashing prices isn't the answer. Maximizing margin is.
Market Data 
Markets 
AI
