What's new

Is semiconductor manufacturing in India an unworkable project?

India has lots of companies (even Indian ones) doing design and testing work for companies like Qualcomm, Intel and many mid level logical IC companies etc.

Do you really think that a Tata Sky set top box has an entirely Indian designed processor ?

Some years ago a team in Intel Bangalore claimed to have designed a processor and even got an Indian government award for it. The reality was that they had upgraded a four-core Xeon processor to a six-core ( or a ten-core ) one. The base ISA and rest of the design remained the same. They basically lied.

I am sure that no Indian company or project has actually designed a processor from scratch. What they do is build products around licensable cores from ARM and other companies.

OK, can you link me to an article where the processor has been designed in India from scratch ?

Intel came close to establishing a fab in 2006 with Maran was minister when Jaya thwarted his attempt to establish the factory in Chennai. They went to Vietnam.

Hmm, that was unfortunate.

But even now the central government could have picked one from two groups that showed interest in establishing fabs : one group was HSMC and the other Jaypee. HSMC even did a ground breaking ceremony at a site in Gujarat.

There are lots of IC which anyone can manufacture by paying IP rights which are less like 0.01 to 0.10 cents to a chip.

That is how chip "design" companies in India operate.

there are lots of low hanging fruits where more money can be made and more value can be added

Well, I hope to do that through my project. If I can convince VCs to invest that is.
 
.
Do you really think that a Tata Sky set top box has an entirely Indian designed processor ?

Some years ago a team in Intel Bangalore claimed to have designed a processor and even got an Indian government award for it. The reality was that they had upgraded a four-core Xeon processor to a six-core ( or a ten-core ) one. The base ISA and rest of the design remained the same. They basically lied.

I am sure that no Indian company or project has actually designed a processor from scratch. What they do is build products around licensable cores from ARM and other companies.

OK, can you link me to an article where the processor has been designed in India from scratch ?



Hmm, that was unfortunate.

But even now the central government could have picked one from two groups that showed interest in establishing fabs : one group was HSMC and the other Jaypee. HSMC even did a ground breaking ceremony at a site in Gujarat.



That is how chip "design" companies in India operate.



Well, I hope to do that through my project. If I can convince VCs to invest that is.


Nope. I am not talking about consumer products here. Worldwide there are only a few companies which design Set Top Boxes. Like a few companies which makes screens or controller PCBs for Smart TVs. That's a different set of argument which we can talk in different scenarios.

There is no need to designing something which has already been made. Its more easy to buy them. Indian companies are mostly consultants. Contrary to many people here India exports a lot of engineering goods even to China. We are good in design and testing. That's why lots of companies have R&D bases here. Its also cheap to train people here. But it's costly for a product being up. That's the problem. Indian chip design companies do chip design for established chip designers like Amd, Intel, QC etc and many mid.level logical IC like (timers, counters, DIO, 4G, 3G etc) chip sets and they are supposed to sell the design after testing. Indian companies do not have the sales or skills or name to do the marketing themselves. They have no experience, zero visibility and their products even if they are good, they will fail.

I will give you an example of 555 Timer IC. The logic of 555 is same across various manufacturers. Someone made it, IPed it and gave it. They took the core design and they just manufacture it under their name. Like STM, or TI etc. Its like added pocket money for those companies. Indian company can also do the same, but will it be worth it? 555 or thousands of such logical circuits are redundant in the sense that it will result in a loss if they try to marker it under their name even if TSMC manufactures it which they do. Since Indian companies export their design we do not hear much about it. My friend was laughing when he saw the news about Shakthi processors cos it seems they have designed even more powerful cores for different microprocessors companies which we don't use in normal windows computers or servers.

Chip architecture is redundant. That's the truth. If we have to come up with entirely new architecture then entire spawn of systems has to designed to work with that architecture. Some Arch are now free. Like Stock Android. The problem is manufacturing. Who is going to be the buyer? An fab has to manufacture couple of hundred millions of chips of different varieties to be profitable. Can Indian companies and manufacturers absorb that much output? The answer is no. Even China doesn't manufacture all components. Memory modules or rams are manufactured in Thailand. Screens in Korea or Japan, Chips in Taiwan etc. But they have the rest in the mainland. We don't have the basic resistor or capacitor manufacturing here. And that's where we should concentrate.
 
.
Is semiconductor manufacturing in India an unworkable project?

Samiksha Goel, DHNS, JUL 20 2020, 07:16 IST UPDATED: JUL 20 2020, 10:15 IST

file6w9no2fmi1s141io0o82-1595220324.jpg

Representative Image

The Minister of Commerce and Industry Piyush Goyal had recently asked Indian industry to set up a semiconductor fabrication plant in India, but experts in the sector seem unsure if it’s a pragmatic proposition as the risks are too high and the investment required is too huge.

This comes as a strategic call in the wake of India-China standoff as the government considers the establishment of such facilities as being integral to the development of the electronics industry.

This is not the first time there has been an attempt to set up a semiconductor fab in the country. India has been trying to get a private semiconductor fab since 2006, but all the previous efforts have failed.

Semiconductor manufacturing comprises the front-end fab manufacturing and the back-end assembly, including packaging and testing. There are only a handful of companies globally that do front-end manufacturing at scale.

India has done well in design and verification for the semiconductor industry, with most of the global semiconductor companies having an R&D footprint in India, but 100% of our chips, memory and display are imported into the country. Just on semiconductor chips alone, we are probably looking at around $10-12 billion of imports this year at a minimum.

PVG Menon, President & CEO, VANN Consulting Pvt. Ltd. and former President, India Electronics & Semiconductor Association (IESA) says, “The past track record of the government does not inspire any confidence that the private sector partner will be treated well. When one or two attempts fail, one can blame the private sector, but when successive attempts fail, one has to start critically examining government policy.”

Menon says that the government support to the private sector players is crucial to set up high technology manufacturing like fabs, as has been the case in some countries like the USA, Taiwan, South Korea, Japan and China.

“In India, the govt has tended to view each private sector proposal with extreme suspicion. If the govt constantly wants to play the role of a suspicious policeman, rather than the role of a facilitator, then I am afraid we are very far away from setting up National Technology Infrastructure Assets with the help of the private sector,” he says.

India has two fabs — SITAR, a unit of the Defence Research and Development Organisation (DRDO) in Bengaluru and a semiconductor laboratory in Chandigarh, which build silicon chips for strategic purposes like defence and space and not for commercial use.

Semiconductor fabs have investment starting around $8 billion, and the numbers climb rapidly upwards. They have very heavy running costs, and technology needs to be upgraded typically every 3-4 years, say industry watchers.

Menon says an extremely aggressive and consistent fiscal support package will have to be made available to develop the semiconductor manufacturing ecosystem. “Past history cannot be wished away and I would say that the industry will wait to see the details before they put money on the table.”

Venkata Simhadri, CEO, Moschip says the government needs to understand what the country needs and the type of semiconductor fabs they are talking about.

“Having a land bank is the least important requirement for setting up a fab. It costs billions of dollars to set up a cutting edge CMOS fab and there are only 2-3 companies in the world having that technology and there is no strong motivation or reason for them to set up their fabs in India. In addition, they can’t just rely only on the local market to fill up the capacity in the fab,” he explains.

Simhadri says the the government should Instead focus on developing the fabless semiconductor eco-system in India. “For example, the leading semiconductor companies like Qualcomm, Broadcom and Meditek don’t have their own fabs. India can focus on creating similar companies to develop products focused on Indian market.”

“Government shall come up with a ‘Fabless semiconductor policy’ and a “dedicated fund” to support this initiative. The growing demand for semiconductor components in IOT devices, solar equipment and electric vehicles does not need cutting edge CMOS fabs. Specialty fabs like Gallium Nitride and Silicon Carbide may have enough demand locally and cost much less to set up. In the short term, India should focus on these specialty fabs and come up with a clear support plan,” he adds saying it’s not just about providing land bank.

The Indian semiconductor component market is expected to be worth $32.35 billion by 2025, growing at a CAGR of 10.1% between 2018 and 2025, according to IESA.

---

@Joe Shearer @xeuss - Similar to what we were discussing day before yesterday.
Taiwan can help with 180nm. And I think not much investment will also be needed.
 
.
Taiwan can help with 180nm. And I think not much investment will also be needed.

And who will buy 180nm chips other than old basic ICs which are now like generics? Money or tech isn't an issue here. Making profit is. China can do all that without thinking about profits. Its HSR except for one line until recently is still running in loss for past 20 years. India as a democracy cannot make that decision to bleed money and asking privates to bleed money for govt as well. Chinese think that they will make.profits after 30-40 years. Here the issue is who will foot the bill for 30-40 years after making the initial investment?
 
.

Pakistan Defence Latest Posts

Back
Top Bottom