Barry, re:<You are correct about Intel's guidance re long-term margins>
And you are correct that many of the new businesses, such as set top boxes and digital TV can not bring margins of 58% like the CPU market. But I don't think their margins will be anything like 50%, more like 30%. These areas have no particular monopolistic advantage to Intel and, in fact, nearly 100% of the Internet boxes you can buy today use RISC CPU's selling at $20-$40. There is no room for a 50% profit margin on such a device.
I believe the x86 CPU market is also becoming a commodity-type business and Intel's margins in CPU's will fall to the 50% area even as its competitor's (AMD, NSM, IDT) margins rise from negative to the 30% area. A single fab is all that is necessary to take a big chunk of the market.
5000 8" wafers/week * 52 weeks * (25400 mm2 area per wafer) / (81 mm2 area per chip) * 0.9 (theoretical yield) * 0.6 (actual yield) =
44 Million CPU's
The above arithmetic shows that at 60% yield, a company like AMD could produce 44 million CPU's in a single fab. (The die size used is for the K6-3D processor to be introduced in few months.) Even a 1.5 billion dollar fab is cheap, if it can make 44 million chips at $100 each = 4.4 Billion. Clearly, CPU prices will be falling fast, and Intel's 58% gross margin will soon be history.
PS, never buy a stock with a P/S (price to sales ratio) above 1/10 the growth rate in sales. (i.e., P/S of 3 is OK only if sales growth of 30% can be maintained.)
Petz |
