Tim
Getting major design wins, e.g., having a major PC maker adopt the line as the basis of its PCs, requires more than just a good chip design
This is a concept that one occasional poster here simply can't grasp.
"Parts is Parts" is a phrase I coined to describe the mindset that believes that all you need is a preliminary datasheet and customers will design in your product. If they don't then it's proof that your competition is somehow breaking the law. Manufacturing capacity, trackrecord, quality, reliability, support infrastructure, chipsets, reference designs, drivers, etc are simply non factors to OEMs, or so the thinking goes.
Historically, it must have been clear by the late 1990s or early 2000s that AMD's business model was unsustainable. The cost of process development and new fabs was escalating at such a pace that it was clear AMD couldn't finance the pace to remain competitive. Good designs that couldn't be manufactured in high volumes wouldn't get those design wins AMD needed to remain competitive, meaning that "Parts is Parts" just wasn't true. AMD needed to turn to foundries for added volume but the x86 cross licensing agreement with Intel limited AMD to 20% of their volume from outside foundries, so AMD was handcuffed with seemingly no way out. Their only chance was to bring a trumped up lawsuit against Intel in hopes of getting a settlement that lifted the 20% foundry volume restriction. It worked to some extent but years of poor quality, limited support, erratic delivery, disastrous design blunders and generally poor relations with OEMs took it's toll. The rest is history.
BTW, I think Global Foundries is still in the foundry hunt.
