Re: Merced, I can only speculate. If you suppose that each new
generation of chip from Intel carries with it some risk that Intel
will stop using REAL for burn-in services, then the Merced delay
means REAL's Intel revenue is assured for a bit longer than we
previously knew. As Merced is a complete departure from all the
microprocessors Intel has made since at least the 8088/8086, the
risk of losing Intel would seem to be somewhat larger with Merced
than with previous generations.
If we know that REAL will win the Merced burn-in business, the delay
is probably irrelevant. As Scott says, unit volume is what is most
important to REAL. Fees for burning in Merceds will probably be
higher than fees for burning in Pentiums, but there will be additional
costs for R&D, new equipment, and transitional expenses as well. (I'm
assuming that past comments by Edwards re: the conversion to 64 bit
drams are also relevant to new CPU architectures.)
A fair question for one of us to put to the company would be do
they expect to win the Merced burn-in business?
Other relevant issues, involve Intel's competitors. Who burns in
AMD and Cyrix CPUs? As Intel's market share slips, REAL's growth
will suffer if someone else is doing the burn-in for those chips.
The market is not very impressed with Intel's (or Compaq's) near-term
prospects. If those companies see demand slow-down (as opposed to
pricing pressures), REAL will also suffer. Presently, the market
does not seem able to distinguish between the pricing difficulties
faced by many PC-related companies and the healthy unit demand growth
that will benefit REAL. Unfortunately, the market has shown it is
not very quick to correct undervaluations of REAL.
Paul |
