InvestorsHub Logo
Followers 29
Posts 25865
Boards Moderated 0
Alias Born 09/11/2002

Re: bobs10 post# 58704

Sunday, 07/03/2005 1:20:08 PM

Sunday, July 03, 2005 1:20:08 PM

Post# of 97747
Re: It doesn't have to be fully ramped to make a huge difference if it is fully 65nm. First, let us assume that the 17% or so of the market is the best that Fab30 can do. Fab36 uses 300mm wafers, each of which can have 2.25 times +/- of a 90nm wafer at the same geometries. In addition, at 65nm, that roughly doubles the number of die per wafer. Assuming that Fab36 is only running 1k wafer starts per week by the end of 2006, that would mean that AMD can roughly double its market share with the two fabs alone.

I think these expectations are far off from what is truly realistic. First, you need to recognize that AMD's current capacity is primarily based on how many 84mm^2 Venice cores they can produce (in addition, the socket-A Sempron die size on 130nm also happens to be 84mm^2). These are AMD's most versatile cores with the highest number of combined skus. San Diego, which is used for the 4000+ and FX skus are low volume right now, and dual core products are also in very low volume. So you have to keep in mind that they are capturing 17% of the market with an average die size of about ~85-90mm^2. This isn't likely to fall much at 65nm either, since dual cores will have become popular at this process node. In fact, at 199mm^2 for their Toledo core, I think it's quite likely that the average 65nm die size will increase, especially considering new features such as Presidio and Pacifica. This will partially offset the benefit of having 300mm wafers.

In terms of capacity, I found this article had some interesting things to say about Fab36:

http://www.yeald.com/Yeald/a/21511/advanced_micro_devices__amd___fab_36_construction_announcement.ht...

Capacity. The fab will have initial equipped capacity of 13K wafer starts/mo. but the size of the facility allows for maximum potential capacity of 20k wafer starts per month. As soon as Fab 36 goes online, which is expected to be in 2006, AMD will start transferring the production of its microprocessors to that facility. Given that it will take about 18-24 months to ramp to 100% capacity utilization but also taking into account the size of the facility, we estimate that somewhere around the second half of 2007 all of AMD's microprocessor production will be taking place in Fab 36.

This suggests that Fab36 will have 3.25k WSPW as part of the first round of ramping, which will probably be ready by the end of '06. By the end of '07, it will have 5k WSPW capacity, which is the same as what Fab30 can do today on 200mm.

http://www3pub.amd.com/locations/fab30.html

However, the first link expects AMD to cease production on Fab30 once Fab36 is fully ramped. At similar die size and at face value, that would seem to imply the full 2.25x capacity advantage of 300mm wafers by the end of '07 (Chartered capacity notwithstanding). But as I mentioned above, 65nm average die sizes are likely to increase as AMD integrates more features to be competitive with Intel's next generation cores. Traditionally, Intel's high volume cores tends to be in the neighborhood of 140mm^2, and I think AMD will use some of their extra capacity to grow their die size competitively.

But keep in mind that these fab investments are costing AMD dearly. From the same article above:

Financing. The estimated initial cost of Fab 36 will be $2.4B. AMD has secured $500M in grants and allowances from the Saxony government. In addition, the company has raised $320M in outside equity funding for the fab construction. On top of that, AMD will take a $700M loan from a syndicate of banks. Thus, on aggregate about $1.5B of the facility cost will be externally financed. As it stands today, AMD's operations must generate an incremental $900M over the next 24 months to cover this new capex requirement. Given AMD's cash flow generating capability, this will not be easy.

I think this says it all, and possibly a reason for frustration that gave way to the recent lawsuit. This article came out about a year and a half ago, and AMD has surely not made the $900M needed to cover the capex expenditures. It's no wonder they are willing to fight by any means necessary.
Volume:
Day Range:
Bid:
Ask:
Last Trade Time:
Total Trades:
  • 1D
  • 1M
  • 3M
  • 6M
  • 1Y
  • 5Y
Recent AMD News