Fabless companies have to accept a slow ramp-up of wafer volumes because of the IP qualification and parametric yield challenges.
Design teams will need to be manufacturing-aware and have in depth understanding of the process variables that have to be included as part of design disciplines. There will be the need for collaboration between fabless company and foundry on understanding the process variations.
Fabless companies cannot assume that their designs will give high yields based on their legacy interfaces that are in place with foundries. There is the need to anticipate potential yield problems and address them at the design implementation phase, and DFM factors have to be carefully analyzed.
Fabless companies will need to build stronger process and product engineering and DFM expertise. It means a extra cost of 50-100M USD per year.
Only a few high-end chip makers today can even afford the exorbitant cost of NEXT-GENERATION RESEARCH AND DESIGN, much less the fabs to build them.
R. Colin Johnson, “IBM Fellow: Moore’s Law Defunct,” EE Times, 4/07/09
[With Intel cranking out new fabs every 2 years based on its huge R&D budget, they are in the catbird seat. The ARM foundries because of the broken fabless model are hamstrung with a failed approach, dropping players and inferior technology. The end game is upon us...]
And now back to the slideshow:
Comments on Fabless Culture
Traditional culture is not tolerant enough to allow failure. Because
innovation is too risky, “Me Too” has widely adopted not only in
products development but also in business model. The result is that
products from different companies are homogenized and price war
Most fabless companies are still under the control of founders
which are in general the technical experts so that they “Prefer to
being a chicken head rather than being the tail of phoenix”
(Chinese proverb). Merge and acquisition are extremely difficult.
Local government does not want the companies they have
supported to close door. They try their best to help these “dwarfs” ,
even if some of them already become “plant men”.
Venture capital does not take venture. The reason behind is that
most venture capital comes from government. They undertake a
responsibility to maintain the value of “State Asset”.
There are not enough real entrepreneurs. Entrepreneurship
becomes a big (problem?)
Because of extremely high cost in process technology and SoC
development, the relationship between fabless company and
foundry become sensitive.
Foundries will select their customers very carefully, and the criteria
for acceptance will involve financial returns as well as technical
capabilities. In addition, foundries will need a vehicle for ramping
volumes of wafers rapidly in new generation of process
There is also the complementary factor that, with the increasing
cost of implementing SoC designs at most advanced technology
node, only a small number of fabless companies can involve.
Whenever the partnership between a fabless company and foundry
established, it would not be easy to change, as the change means
high cost that should be paid by both parties.
At 20nm and 14nm, the foundry business model will probably be
slightly changed. An IDM-like collaboration between fabless
companies and foundries will be expected.
[For anyone having trouble interpreting this, let me help. This means all of the fabless ARM shops are in trouble, big and small. They face fewer choices, less control, bigger delays, worse economics and rising costs. And they won't have FinFET for years. The ARM Bubble descent into the land of sharp objects just picked up speed...]
From a slide at: gsaglobalDOTorg/events/2012/1107/docs/slft2012_wei.pdf
[The number of designs that the ARM foundries can support at each node is dropping. At 65nm it was 65 to 80, at 40nm, it was 40 to 50, at 28nm, it was 20 to 30, at 20nm, it will be 12-16 and at 14nm it will be 5 to 8.
Just another way the ARM foundries are putting themselves out of business...]