They have a Trade / Lobby group over there calling itself grandly the Indian Semiconductor Association and has been pushing for CMOS logic Fabs now for 7 or 8 years. Doubtful if a single one of the members have ever seen the insides of a Fab or have any inkling of Device Physics. At best some are just surplus Programmers, be it Embedded or higher level or EDA vendors focused on Logic / Processors. But most of them are just salesmen working for various overseas chip companies.
Their argument for pushing for a multi-billion dollar CMOS Foundry ( actually 2 ) has been that it will give Indian chip design companies an opportunity to have their designs fabbed in India itself and thus the nation would keep a larger part of the revenue. But are there really any Fabless chip companies in India that can create business for these multi-billion $ CMOS Foundries ?
All that India now has are a large no. of international semiconductor companies who get just small bits of their chip designs done there. Unlike Taiwan ( MediaTek ) or even China ( Spreadturm ) India has NO independent companies who can design a whole SoC. So where are they going to get proven designs with no IP issues to fill the domestic Fabs ? Instead of first promoting a few domestic chip design companies in India that would probably cost no more than 100s of million $ ( incl. making sure that they are on the ball by getting the designs turned into Si at TSMC etc. and then tested ) the Govt. over there seem to be bent on blowing several billion $ to subsidize old node ( 65 nm and older ) Fabs !
Have they not researched first SMIC and other Chinese Foundries that were started almost 15 years ago w/ no in-house R&D but with plans to jump to 90 nm DRAM based on imported technologies ? Even though the Chinese Govt. has enough muscle to favor local mfr. and China now has several Fabless Co.s to generate business, these Foundries have struggled and had to be bailed out repeatedly with billions in new loans.
The dilemma of new Fabs / Foundries that start with no technical expertise of their own is thus : even if they get Govt. support ( in this case $2 to 3 billion ) and start by licensing technology at an old node ( the talk is 65 nm or older ) it still takes them 4 to 5 years to absorb that old process and they never manage to make enough profit to qualify for a loan ( another $3 to 4 billion ) to upgrade to the next / more profitable node ( while the leading edge / most profitable node has perhaps already moved 3 to 4 generations ahead ). Without a huge bail-out ( e,g. by the Govt. ) to leapfrog to the leading edge, it becomes a slow but inevitable slog to shut-down.
One would have thought that at least in India they would have known all about WHITE ELEPHANTS !!
It has been a while since this has been in the works, and the details remain few and far between. Still, the creation of fabs for semiconductors in this part of the world represent a seminal moment in their emergence as true competitors to the United States and the few other countries that produce semiconductors to this point.
What fabs? After reading the article several times I still have no idea of what fabs this is all about. Is it a 300mm? 200mm? Logic? Memory? Packaging? What is the budget? $50billion? $20billion? $100million? That makes a huge difference in attracting people's attention, or not.
I suggest this decision should wait for general election. This is safe becasue, if government changes after election, this decision will become controversial and vendors may not be happy with Indian government. Like that happen to Enron.
NASA's Orion Flight Software Production Systems Manager Darrel G. Raines joins Planet Analog Editor Steve Taranovich and Embedded.com Editor Max Maxfield to talk about embedded flight software used in Orion Spacecraft, part of NASA's Mars mission. Live radio show and live chat. Get your questions ready.
Brought to you by