TSMC has already won the battle for 40nm node proving once again that it is a formidable and agile foundry services leader. However one defines success GlobalFoundries will be successful -- the entire landscape has fundamentally changed. Instead of Big-4 now we have Big-3 in the 28nm node and -- the players are very different - TSMC, GF and - Samsung! And Samsung is indeed the dark horse in 28nm - it is ahead of both GF and TSMC in HK/MG technology. Although Samsung agenda is likely entirely different - Samsung senses that their decades-long objectibe of becoming #1 IC vendor, that is, bigger than Intel, is finally within reach.
GF #1 problem might be at top management levels, starting with ex-CEO of CHRT about whom there seems to be a universal consensus that he is a failed foundry executive and leader with a long list of critical strategic failures. My bet that a candidate that ATIC investor is already trying to attract is ex-Infineon and ex-Grace CEO - Dr. Ulrich Schumacher, an industry leader who is polar opposite of Chia Song Hwee. Dr. Schumacher is a prominent and charismatic industry leader, technologicaly competent and highly respected by customers and all his functional teams, decisive and visionary -- I am confident that he would be a critical addition to the current weakness in GF's top management ranks. We will see...
Even though 28 nm is leading edge, there are simply too many design rule restrictions already. However, GlobalFoundries does not have mainstream base like TSMC at 65 nm and earlier. So I think they are in trouble in being forced to fight at the leading edge for survival.
According to GLOBALFOUNDRIES web site, it states that “GLOBALFOUNDRIES works hard to develop intimate relationships with each and every customer to jointly create solutions that align with their technology and business objectives. The end result is a superior level of performance and innovation, brought to market with amazing speed - benefiting chip-makers, product manufacturers and end users.” Based on GLOBALFOUNDRIES balance sheet, they should take a chance and establish and aggressive roadmap. Competition will always be there, it is the new ideas and the good service that the customers are seeking.
This depends on 3 factors: 1. who is leading below 0.03X technology node 2. who has the capacity available 3. most important: how patient is ATIC. GF is only where they are now because of strong financial backing by ATIC. Right now they are still ramping out from their AMD area which failed. We will see by 2011/2012 when the new fabs and new tech nodes come on-line.
Many have tried to unseat TSMC. This time GlobalFoundry is trying to take all the unprofitable foundries, virtually merge them and defeat the only profitable foundry. This strategy will only work if TSMC lets it work, either through technologic stagnation or price lethargy. Given both variables of failure are completely within TSMC's control, seems not a very SunTzu-like strategy on GF's part to depend on your very seasoned competitor to fail and the upstart to out-perform.
The loss of design portability is due more to the lack of effort on the part of the designers and EDA vendors. It is not due to anything fundamental.
We are working on a new generation of design portability technology that can deliver DFM optimized and electrically optimized layout across different foundries.
I believe innovation in EDA will eventually restore layout portability.
IC Scope Research Inc.
Blog Doing Math in FPGAs Tom Burke 23 comments For a recent project, I explored doing "real" (that is, non-integer) math on a Spartan 3 FPGA. FPGAs, by their nature, do integer math. That is, there's no floating-point ...