Silicon Investor (SI) -- The First Internet Community

STOCKTALK

We've detected that you're using an ad content blocking browser plug-in or feature. Ads provide a critical source of revenue to the continued operation of Silicon Investor. We ask that you disable ad blocking while on Silicon Investor in the best interests of our community. If you are not using an ad blocker but are still receiving this message, make sure your browser's tracking protection is set to the 'standard' level.

Technology Stocks : Intel Corporation (INTC) -- Ignore unavailable to you. Want to Upgrade?

To: burn2learn who wrote (169921)	8/24/2002 2:24:22 AM
From: wanna_bmw	Read Replies (1) \| Respond to of 186894

burn2learn, Re: "Do you think they are not smart enough,don't have the resources, what is the barrier for them to catch up?" This is just my common sense guess, since I am not the expert like some others are, but I see the main "barrier" to competing in the CPU market is the simple tradeoff between performance and efficiency. If you look at the driving force behind foundries, it's to increase the amount of quality customers and ship record wafers at record margins. The margins come from the ASPs they get for promising top notch process technology and timely delivery, as well as cost cutting measures like larger 300mm wafers and further manufacturing efficiency improvements. Customers, on the other hand, can come from anywhere. A large customer like AMD is always good for a foundry like UMC, but they aren't going to sacrifice their margins to make AMD more competitive. If AMD can't afford to compete in performance with UMC's process, then other customers will be there to make use of the same weekly wafer starts. Instead of AMD, consider any CPU manufacturer, and the story should be the same. It's always much higher performance to run things in your own fab, where you control all the production and improvements in manufacturing. The foundry is just going to do what maximizes yield and wafer starts, and that probably won't coincide with optimizations that could make a CPU, say, run 300MHz faster to be more competitive. It's a long winded explanation, and hard to make more concise. I think the sentiment from what I have heard is simply that foundries make bad competition vs Intel, since Intel has all the manufacturing power that they need to fine tune their process for performance. Increasing yield and wafer starts is the secondary motive - maintaining their competitive place in the market is their primary. It's simply a different set of values between a semiconductor giant who owns their own fabs, and a foundry who simply cares about selling off wafers to the highest bidder. wbmw

To: burn2learn who wrote (169921)	8/24/2002 10:23:34 PM
From: Yousef	Read Replies (3) \| Respond to of 186894

B2L, Re: "Yousef, Since you have worked with foundries can you describe ... How have they improved. I think about the additional resources they are getting and the learning curve." B2L, the foundries have considerably improved over the past decade. They are now much closer to providing leading edge technology. In the early days, most foundries received processes from a partnership or "sending Fab". Now, the foundries are quite capable of developing ASIC processes. The FACT is ... today, the foundries are at least one-two years behind INTC in equivalent process at equivalent yield/defect densities. All three leading foundries are NOT in production for .13um 200mm LowK let alone at .13um 300mm technology. Two of the foundries are capable at .13um 200mm FSG technology, but yields are still not competitive. The other "weakness" that the foundries have is in their transfer of technology between the development facility and then between each manufacturing Fab. SPICE models change with each of these transfers. Commodity ASIC's might not be bothered by these changes, but high performance CPU's surely are. Questions to ponder: 1) Once AMD commits to significantly outsource their CPU's, how many process development engineers will "stick around" ?? 2) Once AMD loses these engineers, will it be easy to get them back if the outsourcing strategy needs to come "back in house" ?? 3) Can AMD wait for the foundries to "catch up" to INTC ... How do these foundries get back on the same technology schedule as INTC. Right now, they are 1-2 years behind. 4) As BMW points out, why would a foundry not expect a profit equivalent to the profit derived from wafers sold to other customers ?? 5) Did you know that UMC has >50 different customers ... AMD won't even be #1 in volume. They will be in the top 5. Should AMD be able to call all the "shots' on process changes and priority lot needs ?? If you start to think about the answers to these questions and watch the foundry troubles at .13um (Cu/LowK), I think AMD is taking considerable risk. Make It So, Yousef