Don't panic!

It looks like AMD has learned some things from the ATI acquisition in the form of using major foundries. Hector actually stated that in the conference call!

The difference here compared to the recent move by Texas Instruments is that leading-edge production will remain in-house at Fab 36 and later at Fab 38. Trailing-edge MPUs, which are a big part of the market and especially the ‘channel,' look like perfect products for shipping out to foundries.

This could give some great flexibility to AMD, who could hint that not only will they use Chartered Semiconductor more for 90nm devices, but that they could also use IBM's Building 323!

The fact that Fab36 is fully converted to 65nm and is not being affected by the $500 million capex cut means that we can expect continued ramp of dual- and quad-core for the rest of the year. This is important as this area is the server market honey pot.

The sacrifice being made is Fab 30. Basically, the rate of pace of the 200mm to 300mm conversion at the Dresden fab will slow, though it might be more likely that the pace will stop completely!

The rate of 200mm output will still fall, though 300mm increases at Fab 30 will stop. Overall 300mm production at the 65nm node will grow due to the company's use of Chartered, increased capacity at Fab36 and of course the possible use of IBM's services by the end of the year!

I have no idea what the impact of this asset-lite strategy will be on the decision to go ahead with a 300mm fab in Luther Forrest in NY State, but it doesn't look good without gaining more market share in 2007/8.

Market share gain strategy is still at the core of this restructuring change. AMD claimed today that the pipeline of design-ins are still growing and expects market share gains in 2Q.

That may well mean that 1Q market share fell across the board.