The Business Case
Semiconductor technology must operate not only in accordance with the laws
of physics, but also in accordance with the laws of the marketplace, and the
market is less bullish than it used to be. Many industry leaders say that we
should forget about the relentless drive to microminiaturization. No one really
knows what to do with the level of integration we have today.
In the good old days, large systems helped pay for DRAM, logic, and
microprocessor development. Now, more and more, developers are relying on
PCs to pay the bills. While it would be wonderful to make 16-Gb DRAMs
and billion-transistor microprocessors, where is the business case? There are
already two generations of DRAMs sitting on the shelf. At this rate, we may
not see the 1-Gb DRAM for 10 years.
Another quandary for the industry is system-scale integration. The technology
is already here to merge DRAM, logic, and essentially every function of a
computer onto a single slice of silicon. When system integration takes place
at the silicon level, the industry will certainly experience a paradigm shift as
dramatic as the introduction of integrated circuits in the 1960s and
microprocessors in the 1970s. System-level integration is a natural evolution.
And the consequences would be profound. The challenge would be to adroitly
manage such a profound change.
|
