India Semiconductor Mission: India Needs To Do Enough To Sit On The World Table

Even kids today are aware that there are things called semiconductors and chips. Some of them even know where these are used thanks to the semiconductor shortage that has been rattling the world since the onset of the Covid-19 pandemic. But in India, this awareness is not just an outcome of the global chip crisis. The government of India’s recent endeavors to make the country self-reliant in the areas of chip manufacturing has a greater part to play in beginning a conversation around hardware manufacturing in every other household, something very few in the country cared about before. The government has come up with an ambitious `760 billion incentive to build the semiconductor industry of the country. Although the roots of it lie in reducing its dependence on China, the move has been long awaited by the industry, and thus, highly lauded. Better late than never, right? However, no matter how ambitious the plan, the industry which has been neglected for the longest time now needs to buckle up and get a few things straightened out before it can dive headlong into the game. “I think we always talk about three aspects: economic security, cybersecurity, and infrastructure security. The infrastructure in particular needs to be built. It is high time because otherwise we will be in trouble in the future. You need to have a seat on the table,” says Dr Satya Gupta, President, VLSI Society of India.

Building the ecosystem: There’s more than what meets the eye

“I think that if there is any nation that can emerge as a superpower in chip design, along with China, India has to be there,” notes Dr Naveed Sherwani, CEO, RapidSilicon. The road to the zenith, however, is a long and risky one, something Indian businesses are yet to be ready for. In our bid to achieve self-reliance, big corporations are necessary to give our mission the momentum, but the ultimate goal must be reached by homegrown chip companies.

It is a widely known fact that making chips is a precarious business, which needs steep investment. It becomes even more tedious when a new startup wants to enter the space thanks to the high cost of entry. “The first challenge these startups face is the cost of EDA tools which are prohibitively expensive. Almost 25% of the total cost of the entire design chain is attributed to the cost of the EDA tools,” says Vivek Khaneja, Executive Director, C-DAC.

And although this is being addressed to some extent by open source tools and open domain companies, which help provide easier access to EDA tools and IPs, the truth is that their standard offering is beyond the reach of a small startup.

Availability of IPs is another major obstacle that prevents chip startups from moving forward. When you are trying to solve a problem, you will create a solution that has some sort of programmability and peripheral IPs. Getting hold of those IPs, which are silicon proven today, is not easy and, most importantly, it is not cheap.

Open source IPs have made life easier but only on the surface. At the crux of it, the quality of such IPs may be subpar and integrating those IPs in the design is a matter of shooting in the dark with the hope of striking Eureka.

Additionally, building a prototype for a chip startup is extremely difficult due to their limited funding and the fact that, typically, foundries refrain from giving them much leverage when it comes to prototyping.

Mike Wishart, CEO, e-Fabless, points out, “You look at what happened in software, by removing all the upfront cost, they took out the ‘no’ out of innovation. One of the biggest issues you get in silicon is that there are so many things you have to spend money on.”

Funding for chip startups

“Based on my experience of raising money for my own company, a big part of getting companies funded and up and running is, in fact, persistence,” explains Wishart. Hardware startups are notorious for being incapable of catching the eye of investors due to the high investments that they require along with the even higher risks and marginally lower profits associated with them.

When you come back to the life cycle of funding a company, it is a ‘crawl-walk-run’ cycle for every company. The crawl part is when you have to get the proof of concept, and then come the walk and run part when you produce it and then scale it up. “The funding difference between software and semiconductor is not that different in the walk and run part but it’s the crawl, the getting started part, which is the killer,” says Wishart.

The existing semiconductor ecosystem is built and consolidated around creating and making standard product chips for mass markets, where the basic model is to pour tremendous amounts of money into capital equipment and R&D, create a product and sell a lot of them to enable PCs and phones. Industry leaders emerged in areas like EDA, IP, and foundry and they quite understandably protected their investment licenses that must be signed and often paid for upfront. As a result, not only did it get more expensive to start a company, but it also became incredibly time consuming.

The emergence of open source led to a great amount of costreduction along with simplification, sharing and faster time to design. Yet, their delivery for chips that are optimised for quality, power, performance, and cost is often not at the level required for startups to be able to scale their product. It gives the required push, but it may not be sufficient to keep the momentum.

Building a company from the ground up is a tedious task. As such, building a chip startup is difficult on its own. Moreover, the chip ecosystem is an extremely rigid one that leaves very little space for newcomers. How do we turn the tide?

“I think the tide will only be turned if there is a demand for locally produced chips,” says Sherwani. While the government’s efforts to support domestic startups to encourage them to build chips is commendable, there is a need for these chips to come to the forefront, which can only happen when enough demand is created.

Sherwani suggests having provisions that may help incentivize the use of locally-made chips.

“I’m not saying putting a barrier for big players, but indigenous players need to be given more benefits. As part of this overwhelming plan, we must create incentives for startups,” Sherwani adds.

Lack of VC infrastructure

The lack of a proper venture capitalist (VC) or venture capital infrastructure is another obstacle in allowing proper funding to flow in the right direction when it comes to hardware startups. A lot of VCs that exist in the country do so only on the surface level. At the core, these are just US based VCs who operate in India, among other places. The need, in this case, is much more precise—India needs Indian VCs— to understand the local demand and fuel their capital to meet those needs.

With the semiconductor policy in place, there is hope for such an infrastructure to be built. But that needs a lot of push to come from the industry to show how marketable their product is, whether in local or global terms, in order to attract the right set of eyes towards them and let the pendulum of reverse brain drain going.

A common problem that many tech startups face is the lack of business acumen. No doubt intensive technical knowledge is critical, but the lack of decision making collapses a business even before it begins. And the problem is a deep-rooted one. While a lot of design moved to India, design decisions never moved to India. Big MNCs may have 10,000 people working for them in India but no decision makers are here. The people who are buying chips are mainly in the US and, as such, one key thing that never happened is that chip buyers and people who specify what chips need to be designed and how they need to be designed were ever here in India.

This lack of critical decision making has led to the workforce in India to be unaware of the specific market demands. And when this workforce starts out on its own, it is only natural that their product does not fit market standards. This becomes another hurdle in getting the right funding where the VC’s expectations and the product just do not match.

Leveraging the academia

A large part of Silicon Valley’s success is attributed to the contributions made to innovation and collaboration of the academia and the industry.

Subhasish Mitra, Professor at Stanford, says, “There are three key aspects in which US universities have made a big mark. They are not all Silicon Valley aspects, which is why I believe we can emulate the same in India because I think the expertise that is needed in each of these areas, India has that.”

The first aspect, Mitra reiterates, is the ideas that come out of academia. The way this happens is that they pick a problem that, if solved, would have a big impact.

Second, there needs to be a mindset shift to not just focus on creating specific chips but also create the fundamental technologies to figure out how to create the next technologies that will enable the next generation of chips. Here, too, the formula is the same—you pick a problem that is really important and try to find an end-to-end solution.

Third, it is necessary to give due focus to the electronic design automation (EDA) aspect. The innovation that happened in that aspect was, again, they followed the same formula of problem solving.

Mitra asks, “Why can’t India do all of this? Because if you look at the expertise in India, Indians have creativity that is directly connected to chip design and at the same time Indians are very good at math. So, there is no reason why India should not shine there.”

But while ideas from academia are long-range and industry changing, they also require heavy investments and have long gestation periods to make them see the light of the day.

“In the US, we have been able to fund these ideas and it may take a company 9-10 years before that product hits the market, but it changes the world. We don’t have that kind of patience, which is why I think India needs something different,” says Sherwani.

This, suggests Sherwani, can be solved by focusing on building lowcost, high-efficiency chips, for which the country already has existing capabilities and infrastructure.

The problem, however, lies in striking the right balance. While focusing on such chips would help light that fire and kick that ball, there is a danger of that ball being stuck in a local minima. As such, the focus would only be limited to solving a local problem and may never reach global scale.

This is where it becomes imperative to strike the right balance between encouraging and fueling academic research, while also focusing on keeping India on the right track to achieve reliance and sustainability in chip design.

India first, world second

It is no surprise that the golden period for Indian electronics and semiconductor industry has started to take shape. Sherwani notes, “What we now need is a group of people, Indians who want to make India great. India has a very large number of people with extensive experience in various aspects of chip making. How many nations have that?”

Fostering this ripe talent and providing them avenues to come and build the semiconductor industry is essentially among the best ways to develop the same. Especially since most of these talents have worked closely with companies based in Silicon Valley, they have the knowledge and expertise to emulate the same in India. The product ecosystem in India is already well-developed, what it needs is a little fire to make it shine bright enough for the world to see.

The chip industry is a cutthroat one, and India being in its nascent stage to develop its own niche, needs to be wary of diving headlong into the competition. While the ambition is understandable and lauded, common sense must still prevail if it wishes to make a mark. For that, it needs to turn its head and start looking inward rather than outward. “I think we must start with making simple chips, chips which are relevant to India,” says Sherwani.

Microcontrollers, IoT chips, and LED PCBAs are some aspects that India can aim for, simply because their demand in local terms is extremely high, making it easily scalable while also giving the still-growing market a big indigenous push. Gupta suggests taking a thread out of the big-little approach that many MNCs have followed. This would allow Indian companies the option to choose indigenously produced chips to add to their domestically produced products, which would further increase their value addition.

“For instance, India today makes 700 (types of) LED bulbs domestically for domestic consumption. A small tensor chip goes into every small volt LED. It’s a simple chip but the volume is humongous. And since the product is also made by Indian companies, they can decide what to use. This could be the little part of the big-little. The big part would automatically proceed once we conquer our local ground,” says Gupta.

Additionally, owing to the high investment that chip making requires, not every idea gets elevated. This also means that a lot of great ideas with little marketability, at least initially, have had to go down the drain only because someone decided it would not be worth spending on.

“I think we need to become more open and allow for more people to contribute their ideas in order to see what kind of stars can emerge. It is important to remove the ability of third parties to say ‘no’ to a project as much as we can,” Wishart adds.

Grabbing the seat

The government’s strong foot forward with regards to the semiconductor space was a much-needed move and, although late, it is expected to herald the golden era of the electronics industry of the country. What needs to be done is to adopt a problem-centric approach to think of an end-to-end solution, which can be scaled up from a local problem into a global problem. This is where the academia needs to play its part.

“The real question is how we can get real good ideas. The moment we are able to bring compelling ideas, people will be forced to fund them,” reiterates Mitra.

While sovereignty is the ultimate goal, India needs to focus on getting a share of the pie, no matter how small it is. In the end, India needs to do enough to sit on the table with other countries, because if we control even 3% of the market, we can sit there. If we are zero, we will not even be invited.