Indian Electronics and Semiconductor manufacturing industry

[tharun]

Can some give the reasons why a semiconductor foundry is not viable in terms of economics and weather with links.

 

[Dark Sorrow]

Can some give the reasons why a semiconductor foundry is not viable in terms of economics and weather with links.

1. Setting up a semi-conductor foundry is very expensive. Why should any company take risk while setting up such foundry in India, a country with no such expertise.
2. Existing supply is far greater than demand.
3. Setting up semi-conductor foundries in PRC, Taiwan and Malaysia is more economical and easy as supporting infrastructure, technology and man-power is readily available.
4. PRC and Taiwan are still manufacturing hot-spots and most Semiconductor Manufacturing Company being either Chinese or Taiwanese, why would the set up a foundry in some third country if they can economically set up foundry in their home country.
5. Universities in these country (specifically Taiwan) are highly geared towards material science resulting in fast advancement in MOS technologies and its alternatives.
6. Japanese are more comfortable working with Chinese due to similarities in work culture than compared to India.
7. Indian electronic market is small compared to PRC. It makes sense for the company have manufacturing units close to their market than manufacture it some where far and then ship the material to the market.

 

[Dark Sorrow]

They are daughter companies given very small tasks at times some are funded in small scale to satisfy domestic demand of small electronic products. I have not mislead, This is my field.

Most of the companies I mentioned are multi-million dollar companies if not multi-billion companies (though most of them are).
The fact remain that we (India) don't have such companies.

We know what they are building. They take the design on license production from big players and they duplicate it with small changes and minute changes.
Eg: Rockchip is a fabless firm designing the IC on Chip under licence from ARM on ARM architecture.

In today's world you can't do everything yourself. It is not practical and economically feasible. You sometimes buy and use IP instead of re-inventing the wheel.
Even companies like Intel buy IP from their rival AMD.
TI uses ARM core in its MPU and MCU so does Infineon Technologies, Qualcomm and every other major western semi-conductor company. This does not make anyone incapable . ARM has become de-facto standard in industry like x86/x64
Even Intel and AMD use ARM architecture.
You reuse IP to reduce cost of development, have ready made of the shelf software, cross-tool chains, etc.
Indian companies are able to accomplish this on mass scale.

Market is there but companies couldn't find a place to set up a fab. It is Not possible that you have one of the largest software markets and production facility and you don't have a market for wafers.

I don't know where you are getting your information from but it is totally in-correct.
Our electronics market is very nascent comapaired to PRC. PRC's electronic market completely dwarf Indian electronic market.
I hope you understand difference between software and hardware market. Our software industry work on completely completely deployed hardware (PCs/Workstation/Servers/SBCs). We don't develop or mass produce any of this hardware or its components.

Most probably there are, but I am telling you the reasons why foreign based firms rejected to put their fabrication units. They are NOT getting a suitable place.

Your statement is like fox and sour grapes story. I have read that article. It is a poor attempt to whitewash failure of our government.
Their is an old saying where their is will their is a way.

Please explain this to those firms who rejected to put the plant based on water pollution

In cooperate negation you can't always give the true reason for denying request or offer. For general public you have to give reason that won't embarrass any involved party specifically governments.
The reason is mentioned by the author and we don't know its authenticity and even if authentic we will never know the reasons this companies may have given behind close doors. The reason mentioned may be just for public consumption.

 

[Vijyes]

it has subtropical climate. The air quality is very high in the region North Taiwan where the foundries are. The temp, does not go above 25 degree centigrade has less fluctuation of temperature. This is why it is suited.

When you have high fluctuation in climatic weather in a season then it becomes tough to manage the humidity as it varies with the temp. and it requires more attention and is prone lapses inside a fab.

Pollution is just an excuse. Real reason is that USA doesn't permit ToT of semiconductor manufacturing outside of US and its associated countries like Taiwan, Japan and Korea. IBM had promised to upgrade ISRO SCL lab in chandigarh to 180nm but backtracked. Finally, Israeli Tower Jazz helped India upgrade.

By the way, foreign exchange is extremely precious and it is not the private sector that invests in semiconductor but government. Semiconductor is tightly controlled by governments even today, especially true for chip manufacturers. IC and board manufacturers are more free and unrestricted.

Now,my questions are-
1)does India have the chip manufacturing for defence needs like Tejas, missiles/rockets and other avionics and control rooms? Can the SCL of Chandigarh manufacture these needs of DRDO and ISRO?

2) Can India get a chip fabrication facility if government funds 100%? Do we have the technology to do that ourselves or are there any foreign company willing to help?

 

[Screambowl]

Pollution is just an excuse. Real reason is that USA doesn't permit ToT of semiconductor manufacturing outside of US and its associated countries like Taiwan, Japan and Korea.

There are two types of foundries 1) which designs it's own and produces it, 2) which produces the wafers as per client's design.
India has a technology to develop wafers but only restricted to ISRO, SCL. Of course there are fabs working but they are manufacturing and further putting up of manufacturing plants will take time.

HSMC is doing really well and has good capacity to manufacture wafers.


Semiconductor Manufacturing[edit]

• Gujarat is expected to be home to India's first semiconductor wafer fabrication manufacturing facility by late 2017 in Prantij of Sabarkantha district. To be set up by anchor partner Hindustan Semiconductor Manufacturing Corporation (HSMC) and copartners STMicroelectronics N.V. (France/Italy) and Silterra (Malaysia), it will employ a workforce of over 25,000 including 4,000 direct employees. The group will establish two manufacturing units at an expense of over INR 29,000 crore or about US$4.5 billion, each capable of producing 20,000 wafers per month. Technology nodes currently proposed by this consortium are 90, 65 and 45 nm nodes in Phase I and 45, 28 and 22 nm nodes in Phase II.[82]

In March 2016, HSMC received ₹700 crore worth of seed investment for the project from Mumbai-based private equity fund Next Orbit Ventures (NOV).[83]

• Another consortium, led by Jaiprakash Associates in collaboration with IBM and Tower Semiconductor Ltd., proposed to build a wafer fab in Greater Noida near Yamuna Expressway in Uttar Pradesh at an expense of over INR 34,000 crore or about US$5 billion, capable of producing 40,000 300mm-diameter wafers per month in an advanced CMOS with 90, 65 and 45 nm CMOS nodes initially before gradually switching over to 28 nm and 22 nm CMOS nodes in later phases. As of April 2016, the fate of the project remains uncertain as the debt-ridden lead partner, JPA, exited the project, citing commercial infeasibility of the project.[84]

In February 2014, the union cabinet approved setting up of these fab proposals with decision to extend incentives as follows:

• • 25% subsidy on capital expenditure and tax reimbursement under M-SIPS Policy.
  • Exemption of Basic Customs Duty (BCD) for non-covered capital items.
  • 200% deduction on expenditure on R&D under Section 35(2AB) of the Income Tax Act.
  • Investment-linked deductions under Section 35AD of the IT Act.
  • Interest free loan of around INR 5124 crore to each.[85]
• SunEdison and Adani Group have signed an MoU to build the largest vertically integrated solar photovoltaic fab facility in India with an investment of up to US$4 billion in Gujarat's Mundra, creating 4,500 direct jobs and more than 15,000 indirect jobs by integrating all aspects of solar panel production on site, including polysilicon refining and ingot, cell, and module production.[86]

• The U.S.-based company called Cricket Semiconductor has evinced interest in investing US$1 billion for building an analog integrated-circuit and power supply integrated-circuit specific semiconductor fab in Madhya Pradesh.[87][88]

 

[G10]

There has been no activity from HSMC in ground at site. Most probably that project is also dead. Jayprakash project died long time back.

 

[cobra commando]

An ARM killer from IIT, Madras? Meet the brains behind India’s ambitious processor project

 

[Screambowl]

Guys this is a big achievement now.
This can boost India's computing capability in defence and research
specially in particle physics research and other applications.
It's vast!!

https://www.ndtv.com/india-news/iit...akthrough-in-memory-device-technology-1775555

Dehradun: Researchers at IIT Roorkee today claimed to have made a breakthrough in memory device technology which, they said, usher in a new technological revolution.

A team of researchers from the Department of Physics and Centre for Nanotechnology has developed a high-density, energy-efficient and four-logic state memory device named Magnetoelectric Random Access Memory (MeRAM).


The device could provide a massive boost to overall computing processes and memory-intensive tasks like video and multimedia signal processing, pattern recognition, virtual reality, artificial intelligence and machine learning.

"MeRAM has immense potential to be used in future memory chips for almost all electronic applications, including smart phones, tablets, computers, microprocessors, and for large data storage," Davinder Kaur Walia, a professor at the Department of Physics and Centre for Nanotechnology, said.

Ads by ZINC
The device was constructed in the Functional Nanomaterials Research Laboratory using magnetron sputtering technique, she said.

MeRAM's key advantage over existing technologies is that it combines extraordinary low energy with very high density, high-speed reading and writing times, and non-volatility--the ability to retain data when no power is applied, Ms Walia said.


"The world is rapidly moving towards faster, smaller and quantum technologies which has created an ever-increasing demand for small and more efficient devices and technology. Our focus was to achieve a four-logic state as we knew that then we will be able to create a device which could probably usher in a new technological revolution," she said.

"To achieve this, we used a new material called Ferromagnetic Shape Memory Alloys (FSMA) and the concept of composite barrier were chosen which helps us in achieving the goal of distinguishable memory logic states. The current ultimate memory cell has shown a tremendous improvement of nearly 140 per cent in the memory functions," she added.

 

[Dark Sorrow]

^^ The above article completely forgets to mention about the most important aspect i.e. commercialization.

The article also fails to mention about 2 most important parameters i.e. write endurance and data retention as well as fabrication challenge.

Typical Indian journalism.

 

[HariPrasad-1]

There are two types of foundries 1) which designs it's own and produces it, 2) which produces the wafers as per client's design.
India has a technology to develop wafers but only restricted to ISRO, SCL. Of course there are fabs working but they are manufacturing and further putting up of manufacturing plants will take time.

HSMC is doing really well and has good capacity to manufacture wafers.


Semiconductor Manufacturing[edit]

• Gujarat is expected to be home to India's first semiconductor wafer fabrication manufacturing facility by late 2017 in Prantij of Sabarkantha district. To be set up by anchor partner Hindustan Semiconductor Manufacturing Corporation (HSMC) and copartners STMicroelectronics N.V. (France/Italy) and Silterra (Malaysia), it will employ a workforce of over 25,000 including 4,000 direct employees. The group will establish two manufacturing units at an expense of over INR 29,000 crore or about US$4.5 billion, each capable of producing 20,000 wafers per month. Technology nodes currently proposed by this consortium are 90, 65 and 45 nm nodes in Phase I and 45, 28 and 22 nm nodes in Phase II.[82]

In March 2016, HSMC received ₹700 crore worth of seed investment for the project from Mumbai-based private equity fund Next Orbit Ventures (NOV).[83]

• Another consortium, led by Jaiprakash Associates in collaboration with IBM and Tower Semiconductor Ltd., proposed to build a wafer fab in Greater Noida near Yamuna Expressway in Uttar Pradesh at an expense of over INR 34,000 crore or about US$5 billion, capable of producing 40,000 300mm-diameter wafers per month in an advanced CMOS with 90, 65 and 45 nm CMOS nodes initially before gradually switching over to 28 nm and 22 nm CMOS nodes in later phases. As of April 2016, the fate of the project remains uncertain as the debt-ridden lead partner, JPA, exited the project, citing commercial infeasibility of the project.[84]

In February 2014, the union cabinet approved setting up of these fab proposals with decision to extend incentives as follows:

• • 25% subsidy on capital expenditure and tax reimbursement under M-SIPS Policy.
  • Exemption of Basic Customs Duty (BCD) for non-covered capital items.
  • 200% deduction on expenditure on R&D under Section 35(2AB) of the Income Tax Act.
  • Investment-linked deductions under Section 35AD of the IT Act.
  • Interest free loan of around INR 5124 crore to each.[85]
• SunEdison and Adani Group have signed an MoU to build the largest vertically integrated solar photovoltaic fab facility in India with an investment of up to US$4 billion in Gujarat's Mundra, creating 4,500 direct jobs and more than 15,000 indirect jobs by integrating all aspects of solar panel production on site, including polysilicon refining and ingot, cell, and module production.[86]

• The U.S.-based company called Cricket Semiconductor has evinced interest in investing US$1 billion for building an analog integrated-circuit and power supply integrated-circuit specific semiconductor fab in Madhya Pradesh.[87][88]

It is awesome. We have a lots of talent and we need to Invest in R & D. If we put our money in R & D, I am sure that we can emerge as a very big Hub in R & D that we can compete with whole world and not only india.

 

[Silent_Monk]

I know that there are many chip design companies in Bangalore and maybe even in Hyderabad and chennai, and electronic manufacturing seems to be coming up noida and Gurgaon.

Are chip design companies coming up in north?

Yes.. Qualcomm, ST Micro, NXP are to name a few.. But almost all companies have huge presence in Bangalore only.

 

[Silent_Monk]

It is awesome. We have a lots of talent and we need to Invest in R & D. If we put our money in R & D, I am sure that we can emerge as a very big Hub in R & D that we can compete with whole world and not only india.

Semi conductor industry is totally controlled by Patents from US. Every small detail like even Design of a storage element or new algorithm etc are patented. That is the reason it's very hard to really become stable and to become market leader in semi-conductor industry. What we really have in India (all private semiconductor companies) are just development centers. Every work which is new for example newer architecture or newer technology, generally starts in US. Then on from second revision, It will be in India where we will just do blind folded implementation.

 

[HariPrasad-1]

Semi conductor industry is totally controlled by Patents from US. Every small detail like even Design of a storage element or new algorithm etc are patented. That is the reason it's very hard to really become stable and to become market leader in semi-conductor industry. What we really have in India (all private semiconductor companies) are just development centers. Every work which is new for example newer architecture or newer technology, generally starts in US. Then on from second revision, It will be in India where we will just do blind folded implementation.

Ok. I am just a informed layman in this area.

 

[Indx TechStyle]

IIT-Madras researchers design first indigenously-developed microprocessor

This completely indigenous fabrication is that India has now attained independence in designing, developing and fabricating end-to-end systems within the country, says IIT-M


Researchers at the Indian Institute of Technology, Madras (IIT-M) have designed and booted up a microprocessor, India's first indigenously-developed one, that can be used in mobile computing devices, embedded low power wireless systems and networking systems. This microprocessor will be helpful in reducing reliance on imported microprocessors in Communications and Defence Sectors and is on par with International Standards.

The microprocessor, developed under the project 'Shakti' was fabricated at Semi-Conductor Laboratory (SCL) of Indian Space Research Organisations (Isro) in Chandigarh, making it the first 'RISC V Microprocessor' to be completely designed and made in India. The processors developed under Shakti are based on an open, free programming language ISA (Instruction Set Architecture), called RISC-V ISA, which provides software and hardware freedom for future developments.

The indigenous design, development and fabricating approach reduces the risk of deploying systems that may be infected with backdoors and hardware Trojans. This will have more significance when systems based on Shakti processors are adopted by strategic sectors such as defence, nuclear power installations and government agencies and departments, said the institute.

Lead Researcher Kamakoti Veezhinathan, Reconfigurable Intelligent Systems Engineering (RISE) Laboratory, Department of Computer Science and Engineering, IIT Madras, said, "with the advent of Digital India there are several applications that require customisable processor cores.

The 180nm fabrication facility at SCL Chandigarh is crucial in getting these cores manufacturers within our Country."

Shakti processor family targets clock speeds to suit various end-user application devices such as various consumer electronic devices, mobile computing devices, embedded low power wireless systems and networking systems, among others. The Project is funded by the Ministry of Electronics and Information Technology, Government of India.

IIT-M said that the impact of this completely indigenous fabrication is that India has now attained independence in designing, developing and fabricating end-to-end systems within the country, leading to self-sufficiency. With a large percentage of applications requiring sub 200 MHz processors, the current success paves the way to productisation of many hand-held and control application devices.

The front-end design of the Shakti processors is developed with Bluespec, an open-source High-Level Synthesis (HLS) language. Shakti and its eco-system with its modular design approach enables design and development of application-specific end-user computing and communicating systems.

In July 2018, an initial batch of 300 chips, named RISECREEK was produced under Project Shakti, that were fabricated at the Multinational Chip Manufacturer Intel's facility at Oregon, US, that successfully booted the Linux operating system. Now, the fabrication has been done in India.

https://www.business-standard.com/a...-developed-microprocessor-118102600747_1.html

 

[Indx TechStyle]

IIT-M said that the impact of this completely indigenous fabrication is that India has now attained independence in designing, developing and fabricating end-to-end systems within the country, leading to self-sufficiency. With a large percentage of applications requiring sub 200 MHz processors, the current success paves the way to productisation of many hand-held and control application devices.

The front-end design of the Shakti processors is developed with Bluespec, an open-source High-Level Synthesis (HLS) language. Shakti and its eco-system with its modular design approach enables design and development of application-specific end-user computing and communicating systems.

In July 2018, an initial batch of 300 chips, named RISECREEK was produced under Project Shakti, that were fabricated at the Multinational Chip Manufacturer Intel's facility at Oregon, US, that successfully booted the Linux operating system. Now, the fabrication has been done in India.

Significant but commercial implementation needs time.

 

[Indx TechStyle]

'India's indigenous chip 'Shakti' will not be outdated' says lead researcher of the project

 

[Indx TechStyle]

Articles from July I missed.
India could develop a gallium nitride fab, if the government is willing to invest

Just over ten years later, India is now in a position to develop a gallium nitride fab and a large industry around it, if the government is willing to invest Rs 2,500 crore over five years.

If India develops a gallium nitride fabrication unit, it will join a small club of countries with the technology to grow this material and make devices. A fab can also seed a large electronics industry around it, provided a new set of design and manufacturing startups are also seeded along with it.

A gallium nitride fab is different. It is much less expensive and the technology is still in its early days. Only a few companies now make gallium nitride chips but they would be widespread within a decade. India is expected to be a heavy user of gallium nitride chips for power electronics, especially if the electric vehicles market expands within a decade.

These two institutions now have some of the best facilities in the world for nanoscience and technology research. IISc has a 1,400-sq feet clean room, where dust particles are at a level that is a few millionth that of the environment just outside the building. With it, scientists can make — on a small scale — many devices that are now becoming important for the electronics industry. It can also test these devices as they are made. These facilities are not found beyond some developed countries. For example, even Australia does not have some of the equipment found at IISc.

If the government does not fund the full amount of Rs 2,500 crore, IISc has a back-up plan. The gallium nitride team will set up a company and seek investments of Rs 300 crore, probably from the government itself. It will then be in a position to sell some material that it can make in the facility, and start attracting investment.

Govt to study IISc’s Rs 2,500-crore semiconductor fab proposal

India has two fabs — SITAR in Bengaluru and Semiconductor Laboratory in Chandigarh which builds silicon chips for strategic purposes like defence and space and not for commercial use. The country is a net importer of semiconductor chips for its $ 100 billion electronics industry with limited local capabilities. The Niti Aayogbacked proposal is to build a fabrication facility using Galium Nitride, a material that is commonly used in light emiting diodes or LEDs and an emerging technology, while skipping the already mature silicon-based fab technology. US and China dominate silicon chip technology. The IISc project includes not just building the chips, but also systems for applications in power electronics and radio frequency electronics used for cellphone towers in 5G applications and radars.

The IISc team has submitted two proposals: the first Rs 2,500 crore greenfield facility and managing it for five years. It includes design and roll out of a qualified product in 36 months. The second one is Rs 300 crore line in CeNSe facility and roll out of the first product in 24 months. “If we have to become leaders of this technology, the government has to fund it. That is the way they have done it in China, Taiwan and Korea. It has to be a government-funded and privately operated company,” he said. The government is yet to finalise the funding mechanism, said the Meity official.

$35 million isn't too much & viable to invest too. @ezsasa
@aditya10r was asking for update. So, here is the most active plan. Prantji & Noida Silicon fans are close dead projects as no one's willing to invest fearing losses.

 

[ezsasa]

Articles from July I missed.
India could develop a gallium nitride fab, if the government is willing to invest

Govt to study IISc’s Rs 2,500-crore semiconductor fab proposal

$35 million isn't too much & viable to invest too. @ezsasa
@aditya10r was asking for update. So, here is the most active plan. Prantji & Noida Silicon fans are close dead projects as no one's willing to invest fearing losses.

My view was that they should not go thru Govt funding route, they should rather try for the market itself. if not a single big investor, they can try to impress upon their potential indian customers to form a consortium to invest. This consortium of potential customers will keep the future R&D strong.

 

[Dark Sorrow]

My view was that they should not go thru Govt funding route, they should rather try for the market itself. if not a single big investor, they can try to impress upon their potential indian customers to form a consortium to invest. This consortium of potential customers will keep the future R&D strong.

The problem with the mentioned approach is that what IISC is offering is an academic, untested, unproven and immature product with no support eco-system and developers. I doubt they have not even identified target applications (communication and defense are very vast field with very different processing requirement).
Main question left to answer is will developers be ready to use this product when competing against western giants who have fully developed, proven, mature and easy to use environment.
Now industry is moving towards AFEs and SiP (as Soc have become common) what is their plan an integration of their chip with physical world.
All they are talking about is process technology (fabrication) where as they should be talking about performance/bench-marking of their ISA, AFEs, cryptography implemented in ISA, communication/interfacing peripherals, peripheral and instruction set for controlling application, compiler support, RTOS support, IDE support etc. and the list goes on. Only after successfully answering this their product may become viable for actual implementation in system.
I want to conclude with saying who is going to provide software support for their processor. They have not made this point clear!!!

 

[Indx TechStyle]

Indian firm makes electronic chip to help curb call drop, facilitate 5G

The company already has clients in the US and China and bagged orders to supply 5 million chipsets for various kinds of devices.

NEW DELHI: Bengaluru-based Saankhya Labs Wednesday unveiled the first indigenously developed electronic chipset that can be used for functions like direct TV broadcast on mobile devices, curbing call drops and 5G connections.

"Saankhya Labs, a Bengaluru-based company is launching its indigenously designed and developed, the world's first and most advanced multi-standard next-generation TV system on the chip," Telecom Minister Manoj Sinha said after unveiling the chipset.

All electronic chipsets, considered as the heart of modern devices, have been developed by foreign companies so far. None of them is made in India because there is no modern semiconductor manufacturing plant in the country.

Electronic chipsets of Saankhya Labs are also being manufactured at

Samsung facility in South Korea.
"I am also told that this broadband-broadcast convergence technology has great potential to minimise or eliminate call quality issues faced by telecom operators today," Sinha said.

Saankhya Labs co-founder and CEO Parag Naik said the chipset will help in separating video content from a mobile network and hence reduce the load on the spectrum for improving call quality.

Saankhya's Pruthvi-3 chipset will facilitate direct transmission of video on mobile phones and can also help convert an Android-based smartphone into a satellite phone.

The company expects to launch its chipset-based mobile phone accessories in the form of a dongle and mobile phones within a couple of years, Naik said.

"Semiconductor technology takes time in maturing and adoption. Original design manufacturers will incorporate our chipsets into various products. We will try for getting certain standards based on our technology incorporated in 5G services. Our chipsets are 30 per cent more efficient in output compared to our nearest competitor and 50 per cent cheaper too," he said.

The company already has clients in the US and China and bagged orders to supply 5 million chipsets for various kinds of devices.

"We are looking at 50 per cent year-on-year growth in business. In this financial year, we are expecting sales of around USD 14-15 million (Rs 98-105 crore). We have advance orders for supplying 5 million chipsets to companies in the US that they will use in tablets (PCs), television gateways etc. This order will be completed by the first quarter of 2020," Naik said.

Saankhya Labs has started a pilot for rural broadband services in Scotland and Africa and received a request for the project in the Philippines, Brazil and the US.

"We are in talks with an Indian firm to start rural broadband trials in the country," Naik said.

He said that in the next two years TV gateways, rural broadband technology, satellite communications and defence will be the main contributor to the business.