During this year’s IDF event in San Francisco, California, many of the company’s experts debated on the validity of the well-known Moore’s “Law” concept. While the current pace of development can’t continue indefinitely, they believe we still have a decade before alternatives must be developed.
We reported here
on Intel’s 14nm and 10nm manufacturing plans and we also talked about its competitors here
Intel boldly went ahead and decided to skip or transform the 16nm node and go directly to 14nm manufacturing as low-power and low manufacturing costs are more important these days than the highest achievable frequency and the best performance.
Mark Bohr, Intel
’s senior fellow and director for process architecture and integration reportedly
commented on the issue as follows:
"I do not accept that it has to happen, but it probably will.
I do not see the end of Moore's law for about 10 years."
The reality is that developing new and advanced manufacturing technologies is very expensive now. So expensive that TSMC
needs to invest over $10 billion just to develop the necessary tools and methods for 10nm manufacturing.
Expanding 10nm manufacturing to other FABs and increasing the production volume will certainly cost the Taiwanese company another $10 billion or even more.
Two things will most certainly happen in the future: the shrinking will become costly enough that it won’t be economically viable anymore and the physical limits for transporting electrons through gates that are only a few atoms thick will be reached.
This is exactly the reason for which Intel’s fellows are planning for a fundamental change in technology that will take course over the next decade.