Local semiconductor fabs are a must-have capability, and needed to leverage India's large and rapidly growing internal market for electronics, including defense exports.
"There is one statistic that remains with me from the time I was running an incubator for electronic hardware companies, funded by the Government of India: that India’s import bill for electronics would exceed India’s import bill for oil sometime in the next year or two"
Except that nothing really stops India from designing chips and getting them fabricated in Taiwan or S Korea today. The per-chip cost would turn out to be much lower than the massive bill we are currently paying. This is what China currently does for its leading-edge chips. Even with their massive R&D budget, they still rely on external fabs for 70% of their chips -- although this will shrink over time.
We should pursue designing chips today and transition to Indian fabs whenever they are ready. Designing is very much within our capability, but we are not doing justice to it.
The other thing we should aim for are those software tools that are used in semiconductor industry -- electronic design automation (EDA) is a lucrative market. Otherwise we will end up with an exorbitant import bill while providing these for our universities and training centers and for our start-ups.
indian designers don't have visibility into the full product development cycle. they are focused on only a few limited parts of it, because they are in service mode, doing those things that are outsourced to them by the MNCs. the MNCs have no incentive to change this, but somebody like the tatas or vedanta who have announced plans to make chips may do this. yes, they could do full fab-less product development (as the IIT madras shakti processor demonstrated, although they apparently got the chip fabricated at SCL in chandigarh, too). the other part is to create market demand by preferring the use of indian-developed chips through fiat or guidance by the government. not sure we can do much about EDA software: there is no chance whatsoever that any indian software company will clone it, unless there is some open-source EDA available.
There are a zillion parts to chip making and each area has a need for its own set of tools. Relying wholly on imported tools will make it that much more expensive to launch our domestic industry. We can derive inspiration from the many Indian-owned EDA startups in the US that eventually got acquired by the big players.
If we don't have the appetite to make commercial tools, we should at least help the fledgling open-source EDA community. Indians will gain immensely from working closely with the global open-source community.
I was talking to a Masters student in VLSI in India. His university doesn't provide most of the EDA tools. We can hope that students study abroad and return and build India's semiconductor industry. Or we can raise standards at our universities like the Chinese and Koreans did.
Chinese EDA:
----------------
Completely cut off from the state-of-the-art U.S. semiconductor industry and complementary markets such as that for EDA tools, Chinese innovators have had no choice but to quickly pull together resources and develop homegrown tools with support from the pockets of venture capitalists and, to a large extent, the country’s domestic semiconductor industry. The Chinese government is also heavily involved with its own “IC Big Fund”, which has been involved in several financing rounds for EDA start-ups in China. This commitment to domestic EDA is definitely paying off.
if india focuses on EDA alone, that means we're condemning ourselves to being a 'semiconductor design services' entity, which we already are. there are a lot of other jobs we should train people for, eg. packaging engineer, chip testing engineer, etc so we can cover the complete value chain for semiconductor manufacturing. note that china already has all these infrastructural skills/physical infra in place, and they're trying to do design as well. we need to go in the exact opposite direction.
1. We should go all in -- real chips plus the s/w tools that go with it. EDA is very leading edge and won't be a cake-walk.
2. Even if we get fabs today, very likely we will be starting with dated technology for the foreseeable future. EDA, on the other hand will have no such barriers. It can potentially be much more leading-edge and lucrative.
3. If we cannot crack software tools, supposedly our strength, we will fail to innovate in semiconductors even.
4. Other than the fact the it will move Indian software companies up into more specialized products, EDA will provide affordable domestic tools for use at our Universities that can be backyard testing grounds as well.
5. Talking of design services, no fab deals with end-customers directly (other than with giants like Apple and AMD). They first build satellite design services companies (not software services) for all customer-facing issues. That's the only way they can focus on core business.
"There is one statistic that remains with me from the time I was running an incubator for electronic hardware companies, funded by the Government of India: that India’s import bill for electronics would exceed India’s import bill for oil sometime in the next year or two"
Except that nothing really stops India from designing chips and getting them fabricated in Taiwan or S Korea today. The per-chip cost would turn out to be much lower than the massive bill we are currently paying. This is what China currently does for its leading-edge chips. Even with their massive R&D budget, they still rely on external fabs for 70% of their chips -- although this will shrink over time.
We should pursue designing chips today and transition to Indian fabs whenever they are ready. Designing is very much within our capability, but we are not doing justice to it.
The other thing we should aim for are those software tools that are used in semiconductor industry -- electronic design automation (EDA) is a lucrative market. Otherwise we will end up with an exorbitant import bill while providing these for our universities and training centers and for our start-ups.
indian designers don't have visibility into the full product development cycle. they are focused on only a few limited parts of it, because they are in service mode, doing those things that are outsourced to them by the MNCs. the MNCs have no incentive to change this, but somebody like the tatas or vedanta who have announced plans to make chips may do this. yes, they could do full fab-less product development (as the IIT madras shakti processor demonstrated, although they apparently got the chip fabricated at SCL in chandigarh, too). the other part is to create market demand by preferring the use of indian-developed chips through fiat or guidance by the government. not sure we can do much about EDA software: there is no chance whatsoever that any indian software company will clone it, unless there is some open-source EDA available.
There are a zillion parts to chip making and each area has a need for its own set of tools. Relying wholly on imported tools will make it that much more expensive to launch our domestic industry. We can derive inspiration from the many Indian-owned EDA startups in the US that eventually got acquired by the big players.
If we don't have the appetite to make commercial tools, we should at least help the fledgling open-source EDA community. Indians will gain immensely from working closely with the global open-source community.
I was talking to a Masters student in VLSI in India. His university doesn't provide most of the EDA tools. We can hope that students study abroad and return and build India's semiconductor industry. Or we can raise standards at our universities like the Chinese and Koreans did.
Chinese EDA:
----------------
Completely cut off from the state-of-the-art U.S. semiconductor industry and complementary markets such as that for EDA tools, Chinese innovators have had no choice but to quickly pull together resources and develop homegrown tools with support from the pockets of venture capitalists and, to a large extent, the country’s domestic semiconductor industry. The Chinese government is also heavily involved with its own “IC Big Fund”, which has been involved in several financing rounds for EDA start-ups in China. This commitment to domestic EDA is definitely paying off.
if india focuses on EDA alone, that means we're condemning ourselves to being a 'semiconductor design services' entity, which we already are. there are a lot of other jobs we should train people for, eg. packaging engineer, chip testing engineer, etc so we can cover the complete value chain for semiconductor manufacturing. note that china already has all these infrastructural skills/physical infra in place, and they're trying to do design as well. we need to go in the exact opposite direction.
1. We should go all in -- real chips plus the s/w tools that go with it. EDA is very leading edge and won't be a cake-walk.
2. Even if we get fabs today, very likely we will be starting with dated technology for the foreseeable future. EDA, on the other hand will have no such barriers. It can potentially be much more leading-edge and lucrative.
3. If we cannot crack software tools, supposedly our strength, we will fail to innovate in semiconductors even.
4. Other than the fact the it will move Indian software companies up into more specialized products, EDA will provide affordable domestic tools for use at our Universities that can be backyard testing grounds as well.
5. Talking of design services, no fab deals with end-customers directly (other than with giants like Apple and AMD). They first build satellite design services companies (not software services) for all customer-facing issues. That's the only way they can focus on core business.