India has shifted to high gear when it comes to developing its semiconductor industry. Government initiatives and programs designed to attract investments into its semiconductor ecosystem are already bearing fruit.

For instance, Tata Electronics Pvt Ltd, a wholly owned subsidiary of Tata Sons Pvt Ltd, will build a state-of-the-art, greenfield semiconductor assembly and test facility in Jagiroad, Assam. The facility will be built with an investment outlay of INR 27,000 crore and is expected to generate over 27,000 direct and indirect jobs in the region.

In addition to this, the company is also partnering with Taiwan’s Powerchip Semiconductor Manufacturing Corp. (PSMC) to build India’s first 12-inch wafer fab in Dholera, Gujarat. With fab construction scheduled to begin this year, the new semiconductor company is expected to create more than 20,000 local job opportunities in the future.

Meanwhile, Renesas Electronics Corp., Stars Microelectronics (Thailand) Public Co. Ltd, and CG Power and Industrial Solutions Ltd signed a joint venture agreement to build and operate an outsourced semiconductor assembly and test (OSAT) facility in the country. The project was approved under India’s Semiconductor scheme by the Union Cabinet early this year.

“The government is committed to getting it done this time,” says Jaswinder Ahuja, Corporate VP – International Headquarters and India Managing Director at Cadence Design Systems. “I’ve been in this industry and in India for the last 35 years, and I’ve seen various government starting in the mid-1990s wanting to have semiconductor manufacturing done in India. They have come up with various schemes, but none of them have resulted in any outcomes.”

Ahuja says the one thing that is very different this time is the commitment to getting it done. “Obviously, the schemes and the incentives that the government has put together this time around are based on feedback and learnings from the previous attempts that were made by past governments. But most importantly, they are 100% committed to getting it done, which means whatever is on the table is not the end game as far as they are concerned, right? Their attitude is, ‘We need to get it done.

We need to have very credible players who are willing to commit with us for the long term,’” he says.

He notes that a lot of consortia and companies have submitted proposals in the past. “Even an ordinary observer like me would look at it and question their credibility and intentions, asking, ‘Are they doing this just to get some incentives from the government?” and “Is it a fly-by-night operation or are they serious?” says Ahuja. “But if you look at the players that are on the table now—companies like Micron Technology, the Tatas—I think they are really trying to get the right people engaged and they are trying to take a long-term view, saying, ‘Look, this is going to take us 10–15 years to really get to where we want to get to. It is not going to happen in three to four years, but we need to make a beginning somewhere.’”

These companies know they are practically starting with a clean sheet of paper, as there is no existing ecosystem and infrastructure in place for the semiconductor industry in India right now.

“For semiconductor manufacturing, you need specialty chemicals, specialty gases, other infrastructure,” explains Ahuja. “So, they are really committed to start the journey, to build up the infrastructure and the ecosystem, and get all the players in.”

India is playing catch-up with countries and regions that have been doing this for 20 to 30 years, and the government notes that it will take more than 10 years to get to where India needs to get to. But Ahuja says he is extremely encouraged by the approach that the Indian government is taking this time, and very positive about what the industry can expect to happen over the next five to 10 years.

Being a design powerhouse is not enough

India has a top-notch technical workforce. In fact, Accenture just acquired two IC design companies in India: Excelmax Technologies, which provides custom silicon solutions used in consumer devices, data centers, artificial intelligence (AI) and computational platforms that enable edge AI deployments; and Cientra, a company that offers custom silicon solutions for global clients. While the latter is headquartered in New Jersey, United States, it also has offices in Frankfurt, Germany, as well as in Bangalore, Hyderabad, and New Delhi—underscoring India’s importance in the IC design space. These acquisitions are expected to significantly strengthen Accenture’s Advanced Technology Centers in India.

Meanwhile, another company, Astera Labs Inc., will expand its operations in India by opening a new site in Bangalore to tap into the region’s rich engineering talent and establish a robust pipeline of cutting-edge interconnect technologies that enable AI architectures throughout the data center. In a press statement, Sanjay Gajendra, President and Chief Operating Officer of Astera Labs, said, “As a thriving international market with a world-class technical workforce, India is a clear choice for strategic expansion to help scale our engineering resources and support the next phase of growth for the company. We look forward to harnessing Bangalore’s tremendous business and engineering prowess to deliver on our vision to be the leader in purpose-built connectivity solutions that unlock the true potential of AI and cloud infrastructure.”

But for India’s semiconductor industry to flourish, this is not enough. A complete ecosystem should be present for any industry to succeed. And that continues to be a challenge in India’s semiconductor industry.

Jaswinder Ahuja

“Today, we do not have that yet; we are in the process of building up. As Micron and Tata get their operation off the ground, I think it is inevitable that the ecosystem will start to come into place. So, the first thing India needs is the ecosystem,” says Ahuja.

Next is the infrastructure, he says. “The government is fully committed—it recognizes that you need to have special infrastructure for certain quality of water, electricity, waste disposal, among others. They are saying that we cannot build this infrastructure all over the country—which is why they have identified certain regions or clusters where they are trying to build this ecosystem, so that you can put everything in place both from an infrastructure and an ecosystem perspective, for things to flourish.”

Finally, another challenge that India needs to overcome is manpower. India has no shortage of skills when it comes to chip design. In fact, Ahuja says India is recognized as perhaps having the largest number of semiconductor designers in the world.

“But when it comes to manufacturing, there is no history, there is no expertise in the country. So, I think the companies that are trying to build manufacturing activities in India, like Micron and the Tatas and others, are going to have to source manufacturing expertise from other geographies where the talent is present. Subsequently, there must be an aggressive push in trying to create programs for training the manufacturing talent domestically, which is already in flight—but these programs will take some time to bear fruit. I think this is going to be two to three years before we can start to build up a local talent base (for manufacturing),” Ahuja explains. “They will have to go through appropriate training, go through some level of apprenticeship under the guidance of experts, and then build that capability. And then, we will have to replicate that. So, I think it is a two- to three-year journey to build up each of these elements—the ecosystem of suppliers, the infrastructure, and the talent. I would

say that these are the three big focus areas that we need to work on as a country to get the manufacturing off the ground.”

Focus should be on 28nm

Should India be successful in building up its semiconductor manufacturing base, the next item on the agenda is the technology node to focus on, because observers say that targeting the advanced nodes for India will not be economically feasible.

“I have been involved with various discussions with the government in the India Semiconductor Mission, and I personally have been a huge advocate of us starting with the 28nm node, which, I believe is what the Tatas have shared,” Ahuja explains. “We do not have all the details available, but I think their intent is to start with a 28nm node. The trailing nodes continue in some very low-end applications, but the bulk of the manufacturing happens at three primary nodes. And I think the dynamic has changed now and that the 28nm, I believe, is going to be the longest-living and the highest-volume node in the history of the semiconductor industry. The reason is that it is the one node that can take care of a very fast set of application needs and allow for analog, mixed-signal, and digital integration on the same chip.”

He adds that the economics of semiconductor starts to change when you go down to 16nm and below, where manufacturing costs are so high, and yields are becoming a challenge.

“Historically, what we used to do is when a new node comes, everything moves to the new node. That automatic transition to the new node broke down after 28nm—it did not make economic sense anymore to move everything automatically to the next,” says Ahuja. “So, I believe, considering the history and that we need to start somewhere, and given the fact that the first investment must yield good results, the logical choice is to start at 28nm because you know that it is a secure investment. We are not likely to have huge challenges in getting high yields. It is a proven, established technology that will be relevant for the next 10–15 years. And it is a great place to start and a great place to build up our own ecosystem in our own capability.”

Opportunities abound

Looking at the various market segments right now, there is a very vast opportunity for India’s semiconductor ecosystem. Ahuja highlights the three biggest markets: energy, communications, and automotive.

As the world progressively move towards more renewable sources of energy in line with the global net zero emissions drive, power generation is expected to become more distributed in the future.

“Compared to the large hydro plants or thermal plants that we have had, we will move to different, smaller configurations and footprints. This means that the national energy grid needs to be upgraded and must become a smart energy gate—and that is going to require a very large number of semiconductors,” says Ahuja. “Those might be specialty semiconductors, including compound semiconductors, gallium arsenic (GaAs) or gallium nitride (GaN), or other technologies. I think that is on the road map for the government because the next thing that we are expecting is them approving some projects in the compound semiconductor space, which will be very relevant for the energy space. And it will be volumes of those chips that can be designed and manufactured in India: energy meters, all the chips that go into the smart grid, the distributed energy generation infrastructure—there is a huge opportunity there and I think there is a strong preference by the Indian Government to have Indian suppliers or Indian manufactured chips, because it is a matter of national energy security.”

When it comes to the communications sector, India has developed the NavIC standard, a navigation satellite system much like the GPS standard, but with a more regional focus. The Indian government is requiring all devices that would otherwise have GPS to also have NavIC capability in them.

“Fleet management is a huge application here,” says Ahuja. “The number of fleets, both private and public, in India is going up right as the transportation infrastructure builds up. So, whether it is transportation of goods or people, and they are requiring for security reasons, for efficiency reasons, all fleets need to have NavIC capability. That is another example of a huge opportunity for chips to be designed and manufactured in India.”

Another sector, which is the biggest growth driver in the semiconductor industry right now, is automotive, especially with the growing electric vehicle (EV) industry, the increasing functionality and applications in cars that are driving the increasing use of electronic components, as well as autonomous driving.

“And when I say automotive, I do not mean just cars, but also the autonomous robots that are being used in warehouses in ports, for example. And then, of course, what is unique about the India market is the two-wheelers and three-wheelers,” says Ahuja. “Now, if you look at the global players that are focused on the automotive sector, their primary focus is at the high end—flying taxis and autonomous electric cars like the Teslas and BMWs. But there is a massive market opportunity to design and manufacture chips for the automotive sector at the low end of the spectrum—both for cars as well as for two-wheelers, three-wheelers, and other applications.”

Where Cadence comes in

The Indian Government’s comprehensive policy framework for the country’s electronics and semiconductor industry is focused not just on manufacturing alone, but also on design-led manufacturing, like Apple designs its phones and chips and gets it manufactured in Asia, according to Ahuja.

“That’s very encouraging,” he says. “For Cadence, as more product design and product development happen, we are here to enable them through many different ways. We partner with the government on their DLI program and support startups, as well as through ensuring that there is a good supply of talent.”

Ahuja explains that while they support the government’s Chips to Startup (C2S) program, Cadence also has its own independent university program that supplements the government’s initiatives and goes beyond the number of universities that the government is covering.

“I would say that is on the design side. And then the other part that is equally important is the manufacturing side, which is the fabs and ATMP (assembly, testing, marking, and packaging) and the specialty gases and chemicals and other substances that they need,” says Ahuja.

But he notes that it will not happen overnight. “It is a slow ride. We have seen some green shoots and some interesting startups come up, but it is a long way to go to where we would like to go,” says Ahuja. “I’m constantly in touch with the many entrepreneurs that are looking to enter the fray, targeting some of the sectors that I mentioned earlier and many others. So, I am very encouraged by what I am observing, and I would expect some very interesting players to come up in the next 12 months.”

Ahuja notes that an equally important ecosystem also needs to exist for the design side. “It will include the companies that do the design, like Cadence; IP block providers like ARM, and Cadence, because we also provide IP blocks; and the foundries—because the chips that they design have to get manufactured at those foundries. We play a very vital role in foundry enablement,” Ahuja explains. “There are things that we need to develop, such as PDKs and rule decks, when a company is doing its design. So, we have partnered with the foundry to create the PDKs and the rule decks for these companies to be able to verify that when the design goes to such and such fab for manufacturing, it will work well.”

The next frontier

3D ICs are considered the next frontier for semiconductor innovation. And Ahuja explains why this technology is needed in India.

“If you think about the history of the semiconductor industry and you think about Moore’s Law, what is happening is that as we go below 28nm—up to 16nm, 14nm, 10nm, it was still OK. But, when you really start to get to 7nm, 5nm, and even lower, the economics start to change. It no longer makes sense to move the entire chip to those advanced nodes; they do not get any benefit. The cost of design, the cost of manufacturing, and the cost of setting up the fab is through the roof, therefore it does not make sense to move everything to that. But applications are evolving, and they are requiring bigger and bigger chips,” says Ahuja. “So, how do you put a larger chip into a single package? You must rethink the whole world.”

And then there is the issue of yield. “As the chips get bigger, the risk of yield going down is very high,” says Ahuja. “You do not want to do large monolithic chips anymore, but you will not break them up into smaller chips so that you can get better yield—and it all goes back to economics. At this point in time, a lot of the drivers for these decisions are no longer technology or technology capability, it is purely economics.”

Ahuja continues, “So, you want to do small chips. And if you are going to do that, you need a way to connect these chips. That is where 3D IC comes in.”

With 3D IC, designers no longer need to do 3nm. A lot of the chips can remain at 28nm or 10nm.

“Semiconductors are a scale business,” says Ahuja. “The more scale you have, the better your economics. If you can design a piece of silicon and manufacture it at 10nm instead of doing it for one application, and you can do it for 10 applications and manufacture 10 times the number of those chiplets and use them across those ten applications, your economics will improve again.”

According to him, 3D IC allows designers to take these different pieces of silicon chiplets at different process technology nodes—some digital, some analog, and some RF—and integrate them together in a single package to achieve a small footprint.

“There is enough study and research and data available out there, based on which I can state with confidence that 90 plus percent of the chips designed by 2030 will be 3D IC in nature. It is inevitable,” says Ahuja. “It is heading in that direction. There is no turning back and there is an amazing amount of R&D that is happening in trying to further enhance the packaging technology and how to do these 3D die stacks, and if there are power hotspots, how to manage them.”

The industry is at the early stages of this journey, but Ahuja notes that it is an irreversible trend. “3D IC is the next big inflection point for the semiconductor industry that will have an impact across the board in manufacturing because, obviously, packaging technology must evolve significantly in the industry ecosystem from a point of view of new, independent chiplet providers emerging. It will have an Impact on the design solution side, where we have already a lot of good technology available, but it is evolving very rapidly. The entire industry ecosystem is going to get impacted.”

Now, what is its relevance to India?

“If you are running an independent business and you are starting a new startup, what would you focus on? Would you focus on today’s technology or where the technology is going?” says Ahuja. “In all our consultations with the government, we repeatedly emphasize that point. Ice hockey player Wayne Gretzky used to say that he always skates to where the puck is going to be. Well, the puck is going to be 3D IC in five to six years from now.”

In fact, the 3D IC strategy is already being used today. “This whole chiplet-based design in 3D IC is happening inside of certain companies,” says Ahuja. “If you look at the announcements by Nvidia, Intel, and by AMD, they are already approaching their design in a chiplet-based paradigm with 3D IC packaging. They are already doing it.”

Another sector where 3D IC is expected to offer significant benefits is the memory space.

“And guess who’s in the memory business? Micron,” says Ahuja. “And they are setting up an assembly and testing unit (in India). Now, I have to admit that I do not have full visibility into their roadmap, but I would not be surprised if—Micron is doing 3D IC—whether they will do it in India or not, and when they will do it, that will be a different question. But 3D is inevitable—and because it is at such an early stage of its evolution, much like we have missed the bus on previous inflection points of technology and industry, this time we have an opportunity to play and be an equal participant.”

In short, Ahuja says there are two compelling reasons for India to include 3D ICs in its semiconductor industry roadmap: first, because that is where the puck is going to be; and second, “we have an opportunity to be an equal player and not play catch up with the rest of the world.”

“Strategically, I think it is very relevant and important for India to acknowledge that and to incorporate that into our national strategic roadmap,” says Ahuja.

Very few Indian electronics product companies

Despite all the efforts by the government to boost the country’s electronics and semiconductor manufacturing industry, there continues to be a dearth in Indian electronics product companies.

Ahuja attributes this to the skills gap and the lack of product managers in the country.

“I think, the reason we do not have Indian product companies yet is that there are still gaps, and from a skill perspective, while we have a very large number of extremely proficient designers, we do not have product managers in India because the design community is largely based on multinationals operating in India or services companies,” he explains. “So, product management, product definition, the business side of things, the product architects—we do not have enough of them in India. There are still gaps that we need to fill, much like there are gaps on the manufacturing side. So, while there are areas where we are very strong from a talent perspective as a country, there are also areas where we have gaps. There is still the risk-taking ability access to capital. There are still some challenges on that front that need to be overcome for India to have a flourishing product development ecosystem.”

AI’s impact on the EDA industry

One of the biggest trends happening right now is AI. And according to market analyst International Data Corp. (IDC), the Indian AI market is poised for exponential growth and is projected to surpass $5.9 billion by 2027, driven by a robust compound annual growth rate (CAGR) of 33.7% over five years.

“I think AI will have an impact on every industry over the next five to 10 years,” says Ahuja. “And mostly in a very positive way.”

When it comes to the EDA industry, he notes that Cadence is very intensely involved in leveraging AI in its own tools or leveraging AI techniques in EDA tools—both generative AI (GenAI) as well as machine learning (ML) and other techniques—to try to get the next-level, the next big jump in productivity for its customers’ designers.

“In that sense, the use of AI has a positive impact on the EDA industry, and the EDA industry’s ability to deliver those productivity increases for our customers,” says Ahuja. “In fact, we have been working on incorporating ML and GenAI algorithms in our software. GenAI is more recent, but ML is something that we have been working on for five to six years.”

A lot of Cadence’s tools have ML capability built in already. One example is Cerebrus, which is for digital place and route.

“Traditionally, what happens is when a designer is doing it heuristically, based on their experience and expertise, they will say that to get the optimal design in terms of power, performance, area, they need to explore certain options. But they are limited by how many options they can come up with,” explains Ahuja. “This is a space we would call design space exploration. With Cerebrus, designers can explore hundreds of scenarios by leveraging ML algorithms.”

Outlook for India’s semiconductor industry

Ahuja expects majority of the various announcements that have already been made to be in production by the end of next year.

“I would be very happy to see that,” he says. “That is a very good short term objective. Going a little bit beyond that, I do expect that there will be more such investments that will get announced. I am expecting investments in the compound semiconductor space, in the packaging and ATMP space, as well as in the EMS space for product manufacturers.”

But what is more interesting is where India will be five years or 10 years down the road. “That is the more interesting scenario to look at. I have seen different governments over the years come up with different policy frameworks, with all the right intentions, but the results-oriented or the outcome-oriented focus that I see from this government is unprecedented. Their attitude is not one of saying ‘We have done our work, we have put the policy framework out there, and the industry has to do the rest of the job.’ Their mindset is one of ‘We put something out there. If it does not work, we will change because we want outcomes, we want to get this done right,’ and that, to me, is a huge source of encouragement. If that is how they are approaching it, then outcomes are inevitable.”

It will take time. It may take 10 or even 15 years, but India will stay the course.

“We are going to make it happen. I am extremely optimistic more than I have ever been in the past about the road ahead for the semiconductor ecosystem and the semiconductor industry in India, both on the design side as well as on the manufacturing side,” says Ahuja. “There are challenges and obstacles, no questions about it, but the level of commitment to make it happen is what is going to keep us moving forward and help us overcome those challenges.”

By: DocMemory
Copyright © 2023 CST, Inc. All Rights Reserved