Indian semiconductor ecosystem: News, Updates & Discussions.

New chip units to be key cog in India's 6G localization strategy

By Kalyan Parbat, ET Bureau
Last Updated: Dec 17, 2024, 03:23:00 PM IST
1734884314679.png

Synopsis
India is leveraging new semiconductor plants in Assam and Gujarat to produce 6G network hardware domestically. This initiative aims to reduce reliance on imports, bolster national security, and drive the 'Make in India' program. The government is encouraging collaborations across industry and academia to develop a complete 6G product chain, from materials to finished goods, anticipating 6G's arrival by 2030.

The government will capitalize on the upcoming semiconductor manufacturing units in Assam and Gujarat to drive local production of sixth-generation network hardware, including advanced chipsets that will be used for running 6G mobile broadband networks.

The move is part of India’s strategy to boost the ‘Make in India’ initiative, with an aim to cut dependence on imported telecom network hardware and enhance national security, senior government officials said.

“A critical component of India's 6G strategy is local development and production of advanced chipsets that will enable devices to process and transmit data efficiently…the importance of cutting-edge chipsets cannot be overstated for 6G technology, which promises ultra-high-speed connectivity and low latency,” one of the officials told ET.

The advent of 6G technology is slated to increase the demand for semiconductors, especially since advanced chipsets typically offer superior performance and energy efficiency.

Since the 6G technology product chain will involve a complex network of suppliers, manufacturers and service providers, the government is taking steps to encourage collaborations among the Centre, academia and industry to drive investments across the entire product chain — from raw materials to finished products, the official said.

6G wireless broadband technology is slated to arrive globally, including in India, by 2030 and is expected to pave the way for faster artificial intelligence-powered broadband networks with data speeds over 100 times faster than 5G. 6G is also expected to leverage terahertz frequencies, offering significantly wider bandwidths than 5G.

The decision to locally manufacture 6G technology chipsets comes at a time when the government, via initiatives like the Semiconductor Mission, has outlined ambitions to establish a robust local semiconductor manufacturing ecosystem.

The Union Cabinet has already given its clearance to set up multiple semiconductor units, including by Tata Electronics in Dholera (Gujarat), CG Power in Sanand (Gujarat) and Kayes Semiconductor Pvt Ltd (also in Sanand, Gujarat). Another mega semiconductor unit, backed by Tata Semiconductor Assembly & Test Pvt Ltd, is coming up at Morigaon, Assam.

This apart, the Department of Telecommunications (DoT) is driving multiple 6G-focused R&D projects with a total allocation of Rs 650 crore. These include 111 approved-research proposals, costing roughly Rs 239 crore, for promoting a 6G network ecosystem, and another 19 projects, backed by the Telecom Technology Development Fund, focusing on advancing 5G and paving the way for 6G. Separately, the Department of Science and Technology is funding advanced communication systems focused on 6G.

Senior DoT officials, meanwhile, downplayed recent concerns around India’s challenges in securing large funding for 6G R&D and upgrading existing 4G/5G networks to support 6G wireless broadband technology.

“India's 6G ambitions are not solely defined by budget allocations but by a comprehensive strategy that includes fostering innovation, building infrastructure and developing a skilled workforce,” said a DoT spokesperson.

Another government official said India's approach to 6G development is more strategic, in that, the Centre is adopting a more comprehensive approach to foster a robust end-to-end telecom product ecosystem with a strong emphasis on Make in India.

Global professional services firm Grant Thornton, in a recent report, said India faced significant challenges in pursuing 6G leadership, particularly in infrastructure-readiness and securing adequate funding for R&D.

Industry experts believe the modest Rs 400 crore FY25 budgetary allocation to DoT for telecom R&D may not be adequate to realize the government’s aspirations to have Indian entities own at least 10% of 6G patents globally by 2030.

New chip units to be key cog in India's 6G localisation strategy
 

1735058861820.png

IIT-Madras launches research centre for AMOLED displays, funded by MeitY, DRDO, Tata Sons

By Puran Choudhary, ET Tech
Last Updated: Dec 23, 2024, 07:14:00 PM IST
1735058753402.png
Synopsis
IIT Madras announced the launch of an AMOLED Research Centre to create next-generation displays for smartphones, tablets, watches, and wearables. Funded by MeitY, DRDO, and Tata Sons, the centre aims to develop and manufacture these advanced displays, promoting modular micro-factories. The initiative supports India's vision for self-reliance and aims to make a global impact.

The Indian Institute of Technology Madras (IIT Madras) on Monday announced that a research centre will develop the next generation of AMOLED displays for smartphones, tablets, watches and wearables. These displays adjust brightness, consume less battery and produce better colours and contrast.

The research centre will be called a national Centre of Excellence (CoE)--‘AMOLED Research Centre’ (ARC) which has been funded by MeitY, Defence Research and Development Organisation (DRDO) and Tata Sons. “The aim of ARC is to create a technology that will enable the growth of modular display factories (micro-factories). It has capabilities of both current and next-generation growth processes. The centre currently focuses on small-size displays such as those used in tablets, smartphones, and watches,” IIT-Madras said.

As per data from Statista – a data collection platform, around 46 million smartphone units were shipped across India in the third quarter of 2024, highlighting the growing demand for manufacturing these displays.

Secretary of MeitY, S Krishnan, during the launch, said, “The Centre is working to develop a technique to manufacture AMOLED displays which can be used in digital devices including smartphones, smartwatches and tablets. It is envisioned that the new method, which is based on the ‘economies-of-speed’, will enable a modular micro-factory concept of manufacture of these devices.” He added that the centre is also developing the prototypes of OLED Lighting and OPV power source for mobile phones driving India’s vision for becoming self-reliant and manufacturing it for the global market.

According to IndustryARC– a market research report firm, the market size for AMOLED displays is forecasted to reach $46.5 billion by 2030. The research paper added that the APAC region held the largest market share with 38% in 2023 and is expected to grow at a CAGR of 18.9%.

The IIT-Madras centre will comprise researchers from various fields working on developing the displays. It consists of a state-of-the-art cleanroom which houses fabrication and characterization equipment.

ARC’s principal investigator, Dr. Debdutta Ray, who is also an associate professor in the Department of Electrical Engineering, IIT Madras, said, “This is the only centre in India that has the capability to develop smartphone-sized AMOLED displays. The Centre will work towards turnkey solutions, where we plan to develop, demonstrate, deploy micro-factories for smartphone displays.”

He added that the goal is to keep the unit fabrication times the same or better where the ‘economies-of-speed’ allow smaller mother-glass size during production. “This will lead to a modular production system.”

IIT-Madras launches research centre for AMOLED displays, funded by MeitY, DRDO, Tata Sons
 

India is now self-reliant in GaN MMIC technology. DRDO has developed 1 kW CW GaN UHF high power amplified & GaN MMIC X-Band power amplifiers with power output of 20W & 30W. MMIC designers can de& produce indigenous GaN MMICs for various applications & 30 frequencies upto X-band.

Source- The Forty-Second Report to the Standing Committee on Defence 2023-24.
 
Are we building or can we build modern lithography machines which contains 100% Indian parts ? @Gautam
Nope, we aren't building any lithography machines yet. Most our lab scale lithography is done through manual machines.

Can we build a modern lithography machine? We have most of the building blocks for it. I can give you a general breakdown of the tech that we have/need:

1. Laser source: Modern lithographic machines use CO2 gas lasers as the source. DRDO's LASTEC has developed a 100kW gas dynamic laser about a decade ago. We don't need that much power. We can have a 100% domestic laser source for lithography within 2-3 years.

2. Highly purified Tin: Our domestic reserve for Tin is limited. But we do have Tin refinement capability. Especially EV companies that make their own batteries or chemical industries that supply to battery makers. TATA Chemicals for example. In a lithography machine CO2 laser is used to strike a drop of Tin. This causes a secondary emission from the tin droplet. This emission is usually a flash of extreme UV rays that will be directed to etch the substrate.

3. Vacuum chambers/pumps: All lithographic work is done is high vacuum conditions to prevent the inclusion of any impurities. We have sufficient domestic capacity of making vacuum chambers & pumps. ISRO's has a number of large thermo-vacuum chambers that they use to test their satellites. A lithography machine will require much smaller vacuum chamber.

4. Precise servomotors: Our CNC machine manufacturers can provide at best 2-5 microns of precision with their servomotor drives. We need 1/10th to 1/100th of that level of precision to produce say a 65nm node chip. There is very little domestic capability of this front. This is a major hurdle & there is no easy way around it.

5. Photoresists/Photolithographic masks: At a lab scale we have capacity to make photolithographic masks up to 14nm. This is not a hurdle.

Broadly, photoresists can be of 2 types positive & negative. Both are needed to make modern chips. Indian chemical companies manufacture positive photoresists in bulk. These materials are readily available. Negative photoresists are harder to make. We do have some production capacity for negative photoresists.

Mastery over negative photoresists is a must if we want to do lithography of <100nm. Positive photoresists can give us >100nm but negative photoresists can go up to 7nm.
 
Nope, we aren't building any lithography machines yet. Most our lab scale lithography is done through manual machines.

Can we build a modern lithography machine? We have most of the building blocks for it. I can give you a general breakdown of the tech that we have/need:

1. Laser source: Modern lithographic machines use CO2 gas lasers as the source. DRDO's LASTEC has developed a 100kW gas dynamic laser about a decade ago. We don't need that much power. We can have a 100% domestic laser source for lithography within 2-3 years.

2. Highly purified Tin: Our domestic reserve for Tin is limited. But we do have Tin refinement capability. Especially EV companies that make their own batteries or chemical industries that supply to battery makers. TATA Chemicals for example. In a lithography machine CO2 laser is used to strike a drop of Tin. This causes a secondary emission from the tin droplet. This emission is usually a flash of extreme UV rays that will be directed to etch the substrate.

3. Vacuum chambers/pumps: All lithographic work is done is high vacuum conditions to prevent the inclusion of any impurities. We have sufficient domestic capacity of making vacuum chambers & pumps. ISRO's has a number of large thermo-vacuum chambers that they use to test their satellites. A lithography machine will require much smaller vacuum chamber.

4. Precise servomotors: Our CNC machine manufacturers can provide at best 2-5 microns of precision with their servomotor drives. We need 1/10th to 1/100th of that level of precision to produce say a 65nm node chip. There is very little domestic capability of this front. This is a major hurdle & there is no easy way around it.

5. Photoresists/Photolithographic masks: At a lab scale we have capacity to make photolithographic masks up to 14nm. This is not a hurdle.

Broadly, photoresists can be of 2 types positive & negative. Both are needed to make modern chips. Indian chemical companies manufacture positive photoresists in bulk. These materials are readily available. Negative photoresists are harder to make. We do have some production capacity for negative photoresists.

Mastery over negative photoresists is a must if we want to do lithography of <100nm. Positive photoresists can give us >100nm but negative photoresists can go up to 7nm.
You first need to master Precision Manufacturing and Ultra Low Tolerance Machining before the above steps. Its not coincidence that Japan and Germany are the only ones with EUV Machines and also proficient in Precision Manufacturing Machines. There is the "Art of Precision Manufacturing" then there is "Precision Manufacturing Equipment" itself. China couldn't do something Japan and Germany can because some of the Machines are off limit to export. Ofcourse they are coming up with their own Equipment.

You are also assuming "Great Satan" sits aside and allow OEMs to export Precision Manufacturing Equipment. In my opinion, the steps are
1. Pump Indigenous CNC Machinery companies like Jyothi CNC with extensive support to move into Ultra Low Tolerance ranges
2. Then use that equipment to manufacture sub-components.
3. Put the machine together.
 
Nope, we aren't building any lithography machines yet. Most our lab scale lithography is done through manual machines.

Can we build a modern lithography machine? We have most of the building blocks for it. I can give you a general breakdown of the tech that we have/need:

1. Laser source: Modern lithographic machines use CO2 gas lasers as the source. DRDO's LASTEC has developed a 100kW gas dynamic laser about a decade ago. We don't need that much power. We can have a 100% domestic laser source for lithography within 2-3 years.

2. Highly purified Tin: Our domestic reserve for Tin is limited. But we do have Tin refinement capability. Especially EV companies that make their own batteries or chemical industries that supply to battery makers. TATA Chemicals for example. In a lithography machine CO2 laser is used to strike a drop of Tin. This causes a secondary emission from the tin droplet. This emission is usually a flash of extreme UV rays that will be directed to etch the substrate.

3. Vacuum chambers/pumps: All lithographic work is done is high vacuum conditions to prevent the inclusion of any impurities. We have sufficient domestic capacity of making vacuum chambers & pumps. ISRO's has a number of large thermo-vacuum chambers that they use to test their satellites. A lithography machine will require much smaller vacuum chamber.

4. Precise servomotors: Our CNC machine manufacturers can provide at best 2-5 microns of precision with their servomotor drives. We need 1/10th to 1/100th of that level of precision to produce say a 65nm node chip. There is very little domestic capability of this front. This is a major hurdle & there is no easy way around it.

5. Photoresists/Photolithographic masks: At a lab scale we have capacity to make photolithographic masks up to 14nm. This is not a hurdle.

Broadly, photoresists can be of 2 types positive & negative. Both are needed to make modern chips. Indian chemical companies manufacture positive photoresists in bulk. These materials are readily available. Negative photoresists are harder to make. We do have some production capacity for negative photoresists.

Mastery over negative photoresists is a must if we want to do lithography of <100nm. Positive photoresists can give us >100nm but negative photoresists can go up to 7nm.
Principles according to ASML required to be mastered to produce capable Lithographic machine

Not an Expert in this field, but I would add precise Lenses (for DUV) and Mirrors (for EUV) are also a critical technology for a Lithographic machine required to be mastered and these lenses and mirrors have ISR applications for our own KeyHole series of satellites. BTW US LLNL has helped with the development of EUVL in 1990s by roping in Europeans and Japanese and transferring required IP to them, see pg 13-17 in this PDF:
In order to ensure the acceptance of EUVL as a global standard for next generation lithography, LLC members admitted ASML Holding N.V., a company based in the Netherlands, into the CRADA. The member companies realized the costs of developing an EUVL system were so great that commercial viability would require both U.S. and foreign sales. Jim Glaze, VNL director, said the agreement “was a necessary step to achieving international acceptance” of EUV technology. Negotiations with ASML were led by Mary Egger, Deputy General Counsel, and Paul Gottlieb, Assistant General Counsel for Tech Transfer and IP (retired).“The agreement followed DOE guidelines with respect to U.S. manufacturing.” Mr. Gottlieb continued, “DOE negotiations maximized the benefit to the US economy in light of the economic reality about where chips were manufactured at the time. The chips would be made available to US companies to use.” Gottlieb concluded, “That is the goal when you deal with these U.S. manufacture waivers. You take a look at economic reality and then you try to negotiate the maximum net benefit to the U.S. economy not withstanding the political pressure.”

ASML wasn’t the only international company involved in the EUVL project. Early EUVL infrastructure work took place in Japan particularly on resists and mask blanks. European companies ASML, ZEISS, and TRUMPF also made significant contributions to the EUVL technology. The LLC eventually licensed the EUVL technology to ASML for commercialization. Although the EUVLLC CRADA was initially funded for $250 million over three years, ASML spent more than $9 billion dollars over nearly 20 years to fully commercialize the technology, requiring optics redesigns, source approach changes, flexible illuminator development, and significant effort on vacuum chamber and components cleanliness.

The provisions of the CRADA gave the EUV LLC complete ownership of the EUVL IP developed within the program. The rights retained by the LLC were subject to retained government rights that included a royalty-free government license and certain requirements to promote U.S. manufacture of any new technology. The DOE retained the right to use the technology, royalty-free, for use in defense applications.




No wonder the only other near competitors to ASML are both Lense maker themselves Canon, Nikon




I would add EDA softwares also required for Chip Design which are US based, So US still holds many keys to the semiconductor industry not just with market size but also on tech side, which make ASML etc to obey US sanctions as the key tech was funded by US itself.
 
SCL has fabricated C-DAC's 64-bit Vega processor at their 180nm node.


Until now the 32-bit processor was known to be fabricated.
1737046766102.png

They are running a year behind schedule on their HPC processor development roadmap:
1737046829914.png

Design work for C-DAC's AMR based HPC processor "Aum" had started in 2024. C-DAC has tied up with Indian SoC design firm MosChip Technologies and Japan based Socionext Inc. for the design and development of "Aum". They intend to fabricate the chip on TSMC's 5nm node:

C-DAC partners with MosChip and Socionext for design of HPC Processor AUM based on Arm Architecture - Socionext America
 

How come nobody took notice of this huge development & its implications on India in the future ?

India's been relegated to Tier -2 in spite of being a strategic partner behind US vassal states essentially in Europe ( who're Tier 1 ) which has Eastern Europe in it too who're incidentally NATO allies along with Portugal & in some cases we're at par with China.

What're your views ? @Gautam et al .
 
What're your views ? @Gautam et al .
its implications on India in the future ?
This is a leverage they have & they will use this. Why are you surprised?

They will say we will not sell you AI chips unless you completely cut off any supplies to Russia or any other concession that they want. New Delhi will counter this by pressing on with data localization thus making it harder & more expensive for US companies to get Indian user data to train their AI.

With AI we will eventually take the same model as digital payments. At some point VISA & Mastercard ruled the roost in the Indian domestic digital payments scene. Now they have been effectively sidelined.
India's been relegated to Tier -2 in spite of being a strategic partner
We have been over this. Phrases like:

"Strategic partner"
"Most consequential relationship of the 21st century"
"The oldest & the largest democracies"
"Shared values"
"Alliance of democracies"

These are hyperbolic political talking points. Used opportunistically by both American & Indian govts. when convenient. Not to be taken seriously.
 
IIT-M's SHAKTI microprocessors are near ready to be used by the Indian defence/aerospace industry. From design, to fabricating, to packaging, to backend, entire thing is Indian:


View attachment 40407
View attachment 40408
View attachment 40409
View attachment 40410



IIT Madras & ISRO jointly developed and successfully booted an Atmanirbhar aerospace quality SHAKTI-based Semiconductor Chip

CHENNAI
11th February 2025
Press Release
1739300638960.png

Indian Institute of Technology Madras (IIT Madras) and ISRO have led the way in developing and successfully booting an Atmanirbhar aerospace quality SHAKTI-based Semiconductor Chip.

The SHAKTI microprocessor project is led by Prof. V. Kamakoti at Prathap Subrahmanyam Centre for Digital Intelligence and Secure Hardware Architecture (PSCDISHA) in Department of Computer Science and Engineering, IIT Madras.

The SHAKTI class of systems are based on RISC-V, an open-source Instruction Set Architecture (ISA), for designing custom processors. ‘SHAKTI’ is backed by Ministry of Electronics and Information Technology, Government of India, under its ‘Digital India RISC-V’ initiative (DIRV). It aims to promote indigenous development of microprocessor based products that offer best-in-class security and visibility for users adopting RISC-V technology.

The ‘IRIS’ (Indigenous RISCV Controller for Space Applications) Chip was developed from ‘SHAKTI’ processor baseline. It can be used in diverse domains from IoT and compute systems for strategic needs. This development was part of the effort to indigenize semiconductors used by ISRO for its applications, Command and Control Systems and other critical functions aligning with its march towards ‘Atmanirbhar Bharat’ in Space Technologies.

The ISRO Inertial Systems Unit (IISU) in Thiruvananthapuram proposed the idea of a 64bit RISC-V-based Controller and collaborated with IIT Madras in defining the specifications and designing of the semiconductor chip.

The chip configuration was arrived at addressing the common functional and computing requirements of existing sensors and systems used in ISRO missions. Fault-tolerant internal memories were interfaced to SHAKTI core, enhancing the reliability of the design.

Custom functional and peripheral interface modules being used in multiple space systems like CORDIC, WATCHDOG Timers and advanced serial buses were integrated. Provisions for expandability to future missions also implemented through multiple boot modes and hybrid memory/device extension interfaces. The finalised design underwent software and hardware testing, targeting a high-reliability, high-performance product.

What makes this semiconductor effort important for the Nation is that it was done completely in India and is a great example of ‘Atmanirbhar Bharat’ - conceived by IISU Thiruvananthapuram, designed and implemented by IIT Madras, manufactured by SCL, Chandigarh, packaged by Tata Advanced Systems, Ltd at Perjenahalli, Karnataka, Motherboard PCB (Printed Circuit Board) manufactured by PCB Power, Gujarat, motherboard assembled and mounted by Syrma SGS, Chennai, software developed by IIT Madras and successfully booted at IIT Madras.

Highlighting the importance of this new microprocessor, Prof. V. Kamakoti, Director, IIT Madras, said, “After RIMO in 2018 and MOUSHIK in 2020, this is the third SHAKTI chip we have fabricated at SCL Chandigarh and successfully booted at IIT Madras. That the chip design, chip fabrication, chip packaging, motherboard design and fabrication, assembly, software and boot - all happened inside India, is yet another validation that the complete semiconductor ecosystem and expertise exists within our country.”

Appreciating this joint effort Dr. V Narayanan, Chairman, ISRO, said, “We at ISRO are very happy that IRIS Controller conceived by IISU based on SHAKTI processor of IIT Madras could be successfully developed end-to-end with Indian resources. This marks truly a milestone in “Make in India” efforts in semiconductor design and fabrication. I congratulate all the teams involved, especially the IISU Team led by Sri Padmakumar ES and the IIT Madras team led by Prof. V. Kamakoti. I am sure that this high-performance controller, realized as per our requirements, will contribute significantly to future embedded controllers for space mission-related applications. It is planned to flight test a product based on this controller shortly and performance will be confirmed.”

Commenting on the third consecutive successful manufacturing of SHAKTI series of chip, Shri Kamaljeet Singh, Director General of SCL Chandigarh, stated, “SCL is proud to be associated with IIT Madras and ISRO in the successful development of IRIS-LV Processor. IRIS-LV Processor is fully indigenous and fabricated in SCL's 180 nm technology node encompassing mask frame design, GDS preparation and testing. Post silicon design validation & extensive electrical testing on wafer-level was conducted at SCL in close collaboration with the IIT Madras team. SCL is committed and continually working in association with academia and startups to facilitate and achieve Atmanirbharta in the realization of niche products.”

IIT Madras & ISRO jointly developed and successfully booted an Atmanirbhar aerospace quality SHAKTI-based Semiconductor Chip | Indian Institute of Technology Madras
 
SHAKTI family microprocessor tape outs till date:

RIMO (C-class processor) fabricated on SCL's 180nm CMOS node:
1739301087941.png

RISECREEK (C-class processor) fabricated on Intel's 22 nm FinFET node:
1739301102896.png

MOUSHIK (E-class processor) fabricated on SCL's 180nm CMOS node:
1739301182726.png

IRIS-LV (C-class processor) fabricated on SCL's 180nm CMOS node:
1739301020254.png