I'm not surprised their investors were unhappy with the way Q2 turned out. It's just one idiotic move after another, and this was done after some smart moves -- buying SeaMicro and paying their way out of the GloFo contract, both of which I think were great for AMD.
They maintained their margins, 46% (quite high), but because they oversold their Llano supplies and shorted on the other end, they decided to keep the prices (bulk) the same? People aren't going to pay the same prices for a Llano chip when they know Trinity is due or is already here.
Instead, they delayed Trinity, or limited its supply (it's not suffering from yield issues), in order to sell the overstocked Llanos. But if you don't drop the price, why would anybody want to buy them in the first place? How long is it going to take to sell them at those same prices? The only prices that will change is on account of the OEMs and that only hurts their own bottom line. If you look at the machines Llano is being sold in and the current competition and its respective price, there really is no reason to buy them unless the price for the Llano chips is lower. Apparently, AMD disagrees.
If you've got a better product that's not suffering any constraints then you push it out as fast as possible, particularly when you're so far behind your competitor. I can understand not gunning for market share, but it's also going to hurt your bottom line when you lose market share because you're unwilling to adapt to the current climate. Llano isn't going up against SB anymore, it's going up against IB and a few Trinity models. Who in their right mind would buy the overpriced Llano chip when there's clearly better alternatives?
AMD needs to pick out an OEM they get along with and push forward a new product, Ultrabook style (though not necessarily Ultrabooks; I think Ultrabooks still need another year or two or -$300 to actually make sense). Work with the OEM to provide a small laptop with decent specs and quality that can compete with Ultrabooks but doesn't suffer from the same issues (heat, throttling, space constraints, ULV processors, etc). Develop a standard, see how it goes and then if it does well others will follow suit. This "do what you want with our chips" deal only works if the OEMs make smart decisions. Intel has already learned that's not the case.