I am thinking the same thing, Why would India want to shell out big bucks to build a fab in India? Their infrastructures are horrible and they are not a country known for making tangible goods. Electricity/water/high purity chemicals etc. are severely lacking. Building/operating a modern fab there would be very tough and I highly doubt that they would even be making any money, if they really ended up finishing the fab.
Interestingly, there are two schools of thought here. There was this group, one of them being led by a leading EDA company which felt that there was no need for a India fab - if India could concentrate on its design work, it could go a long way. That was the idealist/no-nonsense kind of group.
But there was this other group which was pitching for a fab and since everyone said it was in "national interest", they had no choice but to toe the line and say putting up the fab was a a great idea. Since ISA was part of this second group and most of the semicon design companies are members of ISA, there was a picture projected that ALL semicon design are for the fab. Nobody wants to stand out and oppose it, even the CEO of the aforementioned EDA company.
No doubt the market is huge and sometimes when you think about it, taking the China angle ( where almost all the Indian telecom equipment comes from China), also appealed to the patriotic streak in Indians. Why depend on Chine/or any single country for something so crucial to Indian interests?
Rumours are also spiralling that India is going to biuld its own microprocessor and that has been going for several years. Maybe that too would happen.
Lets wait and watch. Just yesterday there was a report that the Telecom Commission has approved the enhancement of the foreign direct investment (FDI) limit in the telecom sector from 74 per cent to 100 per cent. We didnt think it would happen but it has.
Who knows?Tthe fab would come up too.. this is incredible India!
When something as big as this is debated, inevitably, it involves power struggle among stakeholderrs. To protect interests of some, discussions often end up with not a real debate, but they offer a means to smoke out who is on whose side. That's a shame but that happens all the time.
That said, I do see the point of the two diverging arguments here.
On one hand, as one of the readers pointed out in this forum, who is to sayIndia shouldn't have a fab? Seriusly, why shouldn't they?
Peter Clarke, my colleague, also pointed out that this is what people said to Japan, Taiwan, Korea and China. And look where we are now.
On the other hand, there is this huge reality. India's lack of interest is a big one. But again, a big project like this might bring order to the chaotic infrastructure.
Am sure it would be great to have a fab or for that matter two -with different geometries. But the whole Indian political arena is very tricky - with the elections coming up next year, no one can really predict whom/which party the ruiling coalition favors and what is really happening behind the scenes. Generally, no one wants to take a call on this - several lobbies might be at work in New Delhi.. it cannot be a decision based on pure merit or feasiblity... too many unseen strings would be attached. Let see which is the strongest one. Like it is said better late than never and if it happens, it could be good for India, despite all the naysayers .
Let me start off by saying that the new EE Times format is utterly illogical and an unmitigated disaster - both functionally and aesthetically. It goes against the native American wisdom of " if it ain't broke don't fix it " ! But then EE Times is no longer owned or managed by US residents.
Now to the subject at hand.
The reason semiconductor manufacturing has not taken off in India ( in spite of starting way back in 1983 at the same time as in Taiwan ) is because there is very little understanding there of its importance - be it among the Govt. scientists & bureaucrats in Delhi or even among the software / chip design types of Bangalore who have dominated the rather inappropriately named Indian Semiconductor Association & through their Ignorance of Semiconductor Physics have skewed its goals & misled it.
For at least a decade these naive people have displayed the totally unrealistic expectation of jumping on the fast moving train of the latest Logic Fab node pulled by Moore's Law and nothing has come out of it. They ought to learn from China, even Taiwan or So. Korea & now Vietnam ( Intel looked at India for its latest A&T plant and then went to Saigon ! ).
Instead of over-estimating their capabilities, India ought to start with in - country assembly of Consumer Electronics ( mostly imported from China now ), gradually move up the Food Chain ( PCBs, chemicals, chip packaging, test,.. ) to low - end Semiconductors e,g. PV cells, LEDs, RF IDs, MEMS, Power Electronics, Mixed Signal, LCD panels and so on, and give up on this simple-minded / insane desire to start with Processors.
In the meantime the Govt needs to fund for at least 5 years at least 3 competing Centers for Semiconductor R&D ( would cost $ 2 to 3 billion but small compared to the import bill for Electronics ). The charter for these Centers should be to develop in depth world - class expertise ( at least 500 Scientists & Engr.s, a third of who should have PhDs ) & pilot scale infrastructure in all the materials, tools, testing, design etc. leading to a smooth transition to late node Fabs and thus avoid loss / stagnation when a Commercial Fab is built eventually
But then India has never disappointed in displaying the worst aspects of popular democracy & crony capitalism combined with hilarious chest - thumping ( post Y2K ) !
How illogical is it for the biggest economies to have a fab?
Not withstanding that its also the largest silicon design power house?
Rather than spend transportation costs from the highest cost fab in the US, it's better to ramp locally for the large(est) economy.
@Sufia: You narrated many points about India with negative connotation. It was very difficult to read and pungent. Most learned people otherwise has lots of praise for India. India also has so many achievements and you should also be very proud of that. You may try to contribute whatever you can.
As for fab, may be India has special need for it - Defence and Space. And India struggels to achieve that. Commercial fab may not be of so much interest to Indian Government. SCL was effort in that direction but story circulates that it was sabotage to SCL as it was actively engaged in Indian Military and Space program.
Delays and delays are quite common in Indian governance system. There is no surprise in that and nothing wrong in telling the truth. India was basically an agriculture country and few years back IT boom came and suddenly everyone started exporting softwrae services. Outsourcing became a big hit as wages here are much lower as compare to Western countries and there was time difference. Even though major portion of the software gets exported from cities like Bangalore, one can easily see the bad condition of infrastructure. You just cant drive on the road. ITs better to have a bicycle but then only God can save your life from potholes and bad drivers.
While software is a comparatively easy business to set up in cities like Bangalore or countries like India but setting up fab is different ball game. Government and politics play very crucial role and you are not just dealing with high profile engineers. In countries like China where communist is the governmnt many decisions can be taken fast and there is no un necessary politics to stop development. But in India situation is totally different, one government change can make different opinions.
First off, let's get this straight - my story is not about negativity on the fab. It is about the way things are going and how delays are going to matter more and more each day. I am an Indian and I love my country and I just would love things to move faster and be executed in a quick way.
I dont get how _hm can think that my post was reeking of negativism.
India is a great country but somewhere down the line, the politicians are just messing it - all the way through. One end of the spectrum would be things like a fab while at the other end it would things like inflation and the sharp rise in all the food items, from tomatoes to rice. They have literally doubled in the past one year or so.
So, it is just not the setting up of the fab that the govt is not getting it right but so many other things. Like anon5664597 says if India's ruling party was like that of China, we would have been way ahead of China, given the expertise of Indian engineers. But that is a totally different way of thinking, so let's not go there now.
What our leaders lack is foresight. When I visited Chongqing about 5 years ago, I was amazed at the size of their airport. Its population would have been around 28 million, I think. But their infrastructure was impeccable - though not a single person could speak English. On the other hand, Mumbai, the fourth most populous city in the world and the highest in India, cannot even boast of an airport even quarter its size. I am not sure of the exact dimensions but just my personal ball-park kind of a figure going by how crowded it all seems.. We just dont seem to think much ahead.. that's the Indian proble. Even simple things like flyovers in cities, we build them over a period of 3-5 years ( delays take place due to corrupt politicians/corporators/contractors) and by the time they are completed, the traffic has tripled and the flyovers are clogged. They were meant to ease the traffic conditions.
This is India and that's how it has been.. Things do improve but the price they pay in terms of time is horrendous.
Back to the fab, it would be good for India to have it, for sure but the question is when.
Am sure some government bureaucreats do read eetimes ( at least they get alerts about any India story) or other publications like these. But to what effect? Even when Indian publications do carry articles about delays, they are just take in the unconcerned-political-stride.
But despite all these hurdles, the Indian spirit somehow manages to stay alive.
@Sufia, I am one of those who took your article in a constructive way. I have commented on this topic many times before in EE Times and else where (for example Pradeep Chakraborty's blog).
My argument in the past and I think that message is still valid in the present is this: India does not need to shell out billions of dollars to build a state of the art fab that offers next generation technology -14nm & 9nm which are going to be less than 5 in the whole world costing more than $5B to build. Older technology fabs like larger than 28nm nodes are still the work horses of today and for some forseeable future. Along with this, India could also encourage MEMS fabs where the process node technology is a minor consideration but technical innovation for newer sensing devices is the enabler.
The beaurocratic brickwall and lack of technology understanding and vision, albeit high-level that Sufia points out are the the major impediments. This can only be solved by active participation of the educated / informed population in the political process.
It is better to be late than never. Why India should not have a fab ? Every one is aware that most of the Semiconductor chip designs are from India. Some of us may not be aware that designs for semiconductor process equipment's modules are being designed in India. Experienced & Intellectural manpower availability will not be an issue. Yes, we all accept the fact that the discussion for setting up of fab in India is going on for more than couple of decades. But, now with the Government's Semiconductor Policy being in place, the infrastructures that are required like Land , Electricity, Water, etc will be provided at subsidised rates. The Roads / connectivity in and around the Bangalore International Airport are being developed and will be completed very soon. There will also be relaxation in custom duties for bringing in New Process Equipments and all other materials required for the manufacturing process. Whether it is 45 / 90 nm it depends on the final device / volume that is to be produced. The delay again may be due to the investors trying to bring in obsolete / old manufacturing line of equipments which are not as per the guide lines set by Government. Sufia, keep posting !
A fab is in India, is a license for politicians and crony businessman to loot money. Exihibit - Fab city in Hyderabad. $10.2 billion in land deals got done over there. Politicians and insiders made a killing. The investors holding properties in communities such as Oxygen city (they really had some very exotic names) are all holding land that will may be in 10years reach the value they paid for....Until the politicians change their value system (build a fab because your defense and country needs it and not as another avenue to loot money) , the fab is not going to happen. There is a BEL fab in Bangalore - they have land, they have power, water maybe. But they do not have vision. I talked to them in 1995 about starting a foundry. They even had Phillips, who would have given the process technology partnership, but the folks who ran it just were not willing to change. I told them I can get 100 expats to build a foundry, they just needed to make a committment. But als they did not move.... I watched SemIndia come and go... I don't think the fab will happen in my lifetime. If it does happen and if the politicians do not change their mindset, it is surely destined for a big budget flop.....
The main argument put forward two years ago in favor of chip manufacture in India was that India needed to be able to do more to satisfy demand for consumer electronics goods internally.
If you forecast how many smartphones, tablets, TVs and other goods India will be buying by the end of this decade it was clear that it was unaffordable unless India was also earning foreign exchange by exporting consumer electronics to China and elsewhere.
And while India could assemble boxes and export those it would be much better value creation to be making the silicon and designing the boxes as well as constructing them.
Now clearly one or two fabs will not got India there, but the argument is that you have to start somewhere. You have to aim to create wealth and export it for money.
And for those that say it is already a lost game. Well that is what people said to South Korea in the 1970s to Taiwan in the 1980s and to China in the 1990s.
There is no need for a fab in India. Many governments believe that if they have a fab it suddenly transforms thier country into a "high-tech" powerhouse. This is really not the reality. What is always missed in the discussion of a fab is that after you come up with the billion or so dollars (for a lagging-edge fab), you are not finished. Every couple of years at least, you need to come up with some more money, the money train really never stops.
So, envision a non-leading edge foundry fab in India. Want to see what it will look like? Visit Silterra in Malaysia and 1st Silicon in Malaysia before it was acquired by X-Fab. These companies typically sell or sold about $200 million dollars of ICs per year. The government of Malaysia was "all in" at the start and couldn't wait to sell its share in these fabs after a few years of stagnation.
Yes, India may actually eventually put up a non-leading edge foundry fab. It really won't change anything and certainly won't provide some big "technology boost" for the country it is looking for. Once again, examine what a local fab or two has done for Russia or Brazil on the world stage of semiconductor production. Not much you say, my point exactly!
A totally agree that countries underestimate the table-stakes of getting involved in semiconductors by orders of magnitude.
You may remember that during the unionification of Germany, the state of Saxony working with Dubai thought a few hundred million dollars would be enough to leverage a position in SiGe manufacturing through a vehicle called Communicant. I don't think the fab shell got finished.
Take two was Abu Dhabi setting aside $10 billion to get into chip manufacturing, with the possibility of a fab in the emirate, through the acquisition of manfucturing operations belonging to AMD and the city-state of Singapore.
Most of that money has now been used to create Globalfoundries but it is starting to seem unlikely that a wafer fab will be built in Abu Dhabi. Soon GloFo is going to have to stand on its own two feet, make profits and compete with TSMC, Samsung and Intel.
But I still feel that fabs in Malaysia whether owned by the state or multinationals, are good for Malaysia. And even a trailing-edge fab in India churning out MEMS microphones or such like, could be making exports for India and bringing much needed revenue into the country.
A lot of thinking has gone into this and the government wouldn't want it to boomerang again. A number of drafts and recommendations have been made and probably right decisions would be made. No doubt, we missed the bus but there is no harm trying.. better to have what you need than need what you don't have. There are some key people in the government whom I meet quite often and they don't really come across as duds or capitalist cronies. There are highly educated and many of them are Ph.Ds.... I am sure they would have done their research and put forth their views too and none of them is a bribe-taking bureaucrat, at least not the hard-core type. I am talking about IAS officers not politicians.
Incidentally, to reply to chipmonks's post the reason Intel moved to Vietnam was not because it thought that putting up a fab in India was unrealistic, (incidentally, it played a key role ( low-key though) when the talks were going on for the 2006 SemIndia fab) but the reason was that the government had a long wish-list and Intel didn't want to comply with that... that's why it moved to Vietnam. Not because it thought that a fab wouldnt work. If that was the reasoning, how come IBM, (an equally conservative multinational) is planning to be a part of the TowerJazz consortium?
India is not always " displaying the worst aspects of popular democracy & crony capitalism combined with hilarious chest - thumping ( post Y2K )." This is a huge country and for instance, lets just take the political parties -- believe it or not, there are 363 political parties (seven national parties, 34 state parties, and 242 registered and unrecognized parties) and all are supposedly working for the good of the people, But despite all this, there is still widespread poverty and inequality of wealth and of course the great digital divide.
The market is huge, like I said earlier in the report and if India can work on a project (Aadhaar) whch is about giving a unique identification number to over one billion people and handle all the logistics and supply chain management of a project of this scale ( by the way, it is headed by Nandan Nilekani, co-founder $ former co-chairman of Infosys), I am sure it can work its way around a fab too. Sure, tons of problems are going to be there but things have a way of working themselves our or dying out. Lets hope the latter doesn't happen again. And, even if it does, so what?.. life goes on minus a fab. Havent we heard of the word, resilience?
Kris, its good if India sticks to software and buy the chips from somewhere else. But looking at the consumption this country has and low wages, I guess many industry leaders would make a lucrative business plan in front of the government that once we have silicon fabs everyone will make money. Because in India politics is only about making money. Many politicians have no strong educational background and have come up only because of family politics. They still are living in the era of bullock cart and caste politics.
I am not saying India doesnt deserve to have the silicon fabs or cannot have one. With the second highest population in the world and already proved in software services they can do it. But fab needs discipline, planning, government committment etc. In India, they hardly plan for future. One can see a high end BMW and cow on the same road. One flyover taken many years to build.
Cost of setting up fab, maintaining it and bringing it to break even needs lot of committment from industry, people and government. And like someone mentioned people in decision making position do not have right attitude to get these things going.
thank you Anon...the problem I have with this fab investment is that modern fabs employ few people and are very capital intensive...I would think the software model with large employment and low capital investment fits India much better...Kris
You have stated that you have frequent contact with many of the key figures involved in setting up a Wafer Fab in India. Wondering if they ever check EE Times ? I believe that it would help them considerably if they were to review the comments made here.
You could do them ( especially the Ministry types ) and India a big favor by directing them to these comments.
BTW what is IBM's stake in the Tower / Jazz proposal ? Would they invest ? or is it just Technical Consultancy for a fee ?
Well in country like India where there are so many different languages, so many mainstream religions, every state has their own language and culture. On the top of this country has the second highest population in the world and there is democracy. Government cannot do anything even if they are willing to do. Citizens are ready to take their right but no one wants to do their duty, the country is lucky that IT software boom happened and companies are still running. With such a political instability and corruption, its a risky game to invite manufacturing. Its easy to export software but for Fab everything from politics to local manpower matters, it wont be wasy for country like India. And absolutely no surprises that the project is delayed, here things take many year to become a reality.
The data being circulated everywhere in India is that India will import USD200 Billion worth of Electronics Products by 2020 . Indian Government agencies know that such a big import bill is not possible and hence something need to be done . Building a fab is one of the solutions being proposed in that big puzzle . But , the question is what will be produced in those fabs when there are no big Systems and Product design houses in India . India is importing almost its entire requirement of more than USD 50 billion of telecom equipments . Hence , it is very obvious that the cutting edge Fab should not be the first priority . Firstly , India should groom big System design houses and Fabless semicinductor companies , almost 100 to 200 of them and at least 20-30 of them having revenue of more than USD500 million . This process will take 4-5 years and in between it will be very obvious whether a fab is needed . Presently , just be smart and use Fab facilities around the globe .
......but with local chip manufacturing capabilities, even if it is is specialzed More-than-Moore processes away from the leading edge, it is easier to foster those fabless chip companies, which do need to be up to speed in an understandng of Design for Manufacture....
On the one hand it is sensible to start with a fab plan while also nurturing fabless chip companies. On the other the fab and process part of the chip manufacturing process consumes the vast majority of the capital expense and R&D funds.
A friend and colleague of mine was directly involved in setting up the 1983 fab in India. His frustration was with the government, starting with the shipment of photoresist that sat on the airport tarmac for several months before they would release it. Of course, the photoresist turned into glue!
Was reading an article and found something interesting in the same context.
Indians can't sacrifice democracy to an authoritative model even when delivery of prosperity might be clearly visible in an authoritative regime. The authoritative model of China, arguably, might border on autocracy and perfect opaqueness but has taken them back to their centuries-old glory with a strong global imprint. India might or might not wish thay route the whole hog, but there is a trajectory beyond the current model of democracy.
Would India connsider manufacture of electronic discretes such as transitors, diodes, etc. or probably even passive elements? In fact, I'm working in transferring technologies, Tower being one of many at the receiving end.
A fab in India would sure lure the Malaysia/ Singapore living Indians towards a switch of job. So I was told.
What are the engineering and design challenges in creating successful IoT devices? These devices are usually small, resource-constrained electronics designed to sense, collect, send, and/or interpret data. Some of the devices need to be smart enough to act upon data in real time, 24/7. Specifically the guests will discuss sensors, security, and lessons from IoT deployments.