Re: Your AMD number is off too, and should be 1,219 million, not 285 mil.
Thanks for being the only Intel long with the brains and internal consistency to realize that neither the AMD number nor the Intel made any sense using Paul's interpretive distortion of the post I made.
For anyone who actually follows the financials of the two companies, it should have been obvious that I was listing the quarterly changes for each company -> negative for Intel and positive for AMD.
Bottom line is that, in the first quarter of this year, Intel's capital position was down by 9% (of last quarter's $12,500) while AMD's capital position was up by 30% over last quarter's $934.
And that was in a brutal quarter with AMD lacking a low cost (integrated) chipset, lacking mobile processors, and lacking SMP systems.
This quarter AMD has integrated systems, next quarter (maybe) mobile and SMP systems.
I'm not long AMD any more. I think the current price war will be long and brutal and AMD will be under severe pressure. But Sanders has been planning for this for a long time. I remember when all the AMD longs (including me) were expecting huge increases in AMD ASPs, and were disappointed that AMD wasn't locking itself into huge expansion projects. Now, "suddenly," there is excess capacity as far as the eye can see, with a large number of new FABs coming on line in Taiwan, Korea, and Europe, and all the semi-commodity chips like flash and comm are facing brutal competition. So building new FABs for CPUs, and using the old FABs for those "other" purposes, once a great strategy, is now a great way to lose money.
AMD's costs are very low and they're carrying almost no extraneous, cash draining baggage. AMD can make GHZ+ Durons in its old Austin FAB. By the end of the year fully 50% of its capacity will consist of much faster copper chips from Dresden. With one last $0.5 Billion investment, AMD will have half their capacity in SOI at .13 on copper.
AMD can make money selling those GHZ+ Austin chips for $50 and put them in $70 integrated motherboards with 128 meg SDRAM costing $25. AMD can make money selling it's high end Dresden chips for $100 to be put on a $120 DDR motherboard with $60 of DDR DRAM. Intel, OTOH, will have an increasingly difficult time getting substantially more for Celeron or PIII than what AMD will be charging for GHZ+ Duron. P4 will command a higher price, but keep hitting the ceiling of suspicious overall performance and the cost of $160 dual channel motherboards. Until brookdale or a 3rd party board comes out for P4, it's also saddled with RDRAM's $90 cost for a minimum configuration.
Intel still has $10 Billion to burn, but they will probably need every cent over the next year to put into place copper FABs and to cover the $1+ Billion quarterly losses of Intel's "other" divisions.
The outcome of Intel's dropping of it's prices to about one third of what they have been during the glorious past decade of growth and monopoly profits will take some time to sink in. But when it does, the stock will most likely take a big hit.
The monopoly days are over for Intel. Even if the price war flattens AMD, somebody will buy AMD and keep CPU prices at a level that reflects the existence of competition.
Making little chips out of silicon for $10 and then selling them for an average of $250 was a fabulous business for Intel while it lasted. And it lasted for an unbelievably long time.
With competition forcing manufacturers to put more into each little chip so that they cost $20 to $25, then selling them at an average of $90 is still a great business, but not if 15 years experience in the "old ways" has resulted in the Intel style loading of $150 of overhead costs on top of each little chip.
I think this whole business is in for a rough couple of years. |
