Mark and Intel Investors
Re: Intel's Decision to Delay Construction on Fab 16 (Fort Worth)
Background
On Saturday I had a telephone conversation with Jules Garfunkel. During that conversation we talked about the recent events at Intel and how the Street had penalized Intel's stock price on Friday, (down more than 2) because it had first interpreted Intel's announcement, (that it would delay Fab 16 plans) as a sign of weakening demand. But suppose, it was ypothesis. While I volunteered to write it up for SI, the development of the hypothesis was a joint effort, and a testament to the productivity that can come from "brainstorming" an idea. However, the reader should be reminded that the hypothesis is still highly speculative in nature.
Hypothesis
On Friday Intel announced that Fab 16 (Fort Worth, Texas) had been
delayed for one year, coupled with a change in Fab 18 (Kiriat Gat,
Israel) production plans had caused much speculation and confusion as to Intel's reasons, or need for this change. Much of this original confusion centered around what appeared to be the demand situation for Flash Memory and Microprocessors. The then "current" understanding was that the demand for both products remained high, but by diverting Fab 18 from a pure Flash process make up, - to a dual
process capability - microprocessors and flash -will actually accelerate the availability of new CPU capacity for Intel. At the same time capital expenditures for 1998 would be lowered.
However, we believed that what had been overlooked during that brief
time period was the effect of a possible settlement between Digital Equipment Corporation (DEC) and Intel regarding DEC's patent infringement law suit against Intel, and Intel's countersuit. Although there is no guarantee that any settlement is imminent, let alone what the terms of that settlement could be, rumors had abounded concerning DEC's Hudson, Mass wafer Fab #6 and the DEC Alpha chip as being part of that settlement.
Therefore we reasoned, could Intel potentially, (I emphasize the word potentially) be acquiring wafer fab capacity that was not planned on when they forecasted future demand and manufacturing requirements that led to the decision to build Fab 16 in the first place. And if so, should Intel's shift in Fab utilization plans be viewed as a strong positive for the company and not a negative.
Why is this significant?
There are several reasons for this.
First, Intel may possibly pay cash for DEC's fab thereby increasing
its capital equipment spending much faster than previously forecasted, but in the long run it would save Intel an estimated $ 1 Billion.
Second, the acquisition of this fab and its employees, who may shortly become Intel employees, would instantaneously increase headcount, make demands on retraining employees to Intel standards, methodology and procedures, and increase further capital equipment spending to bring the fab line up to Intel standards. However, by taking in a "foreign" unanticipated fabrication plant, Intel would get instant trained personnel, and a new Fab on-line faster than any other way and practically without missing a beat.
If Fab 16 in Fort Worth proceeds on schedule, Intel will have to
continue (we are assuming they have already begun) hiring professional and technician staff well in advance so as to train these people in existing Intel fabs in a production environment working shoulder to shoulder with existing, experienced Intel staff. If the sudden acquisition of DEC's Fab #6 becomes a reality, the burden of adding an entirely new facility with people and staff requiring training/retraining would stretch Intel's ability to "expand" beyond its limits. Thereffore taking a bigger bite out of the Fort Worth Fab. But why the need for such rapid deployment of new Fab capacity we asked?
Third, Intel during its recent Conference Call told analysts they
planned to address the sub zero segment of the market in the very near future. Therefore, we believe, that a new chip set which integrates many PC system functions - graphics display, modem, audio, disk controllers, etc. would be required "instantly" by Intel.
This chip set, more than likely would be one piece of silicon to keep costs down. However, because this new requirement was not planned for, and was brought on by the recent, but unanticipated success of the sub zero PC, Intel finds itself with the need to expand Fab capacity quickly to meet that demand. Therefore, acquiring DEC's Hudson Mass plant would allow Intel to develop a "second" chip for addressing the sub $1000 PC market and act in concert with a STANDARD Pentium MMX, in the minimum time period.
Fourth, when DEC's Hudson Mass. Fab, was first built, it was touted to be the most similar chip Fab to Intel's standards, than any Fab outside of Intel's domain. Therefore, we estimated that at a purchase price to Intel of $700 Million, plus another $300 Million to convert the Fab would cost a total of $1 Billion instead of $2 Billion for a new plant. An overall saving of $1 Billion.
If the sub zero market is indeed growing as rapidly as some of the PC analysts would like us to believe, and Intel is going to address this market segment and they have promised to do in their CC, then the acquisition of DEC's Fab #6 in Hudson, Mass. would then prove to be very opportunistic. If this arrangement is part and parcel of Intel closing the door on their legal entanglement with DEC, as well as solidifying DEC's future support of Intel architecture devices (read -
NOT AMD devices), then this may be a compelling solution to THREE problems for Intel - the proverbial killing of three birds with one stone!, or snatching victory out of the jaws of problems. (poetic license)
Thus, the implications of a possible DEC-Intel settlement, along with addressing the sub $1000 PC market with an integrated solution, should be considered as a possible contributing factor in Intel's recent decision to place Fab 16 (Fort Worth) on temporary hold and redirect the plans for Fab 18 (Israel) production processes to include CPU/Logic as well as the original Flash process capability.
Paul and Jules |
