Chips, Wireless Technologies and All..

This comes from a misconception that owning chips means “manufacturing chips”. It does not take too much of googling to know that most chips in the world are owned by fabless companies. Semiconductor Fab is a printing press and Fabless Semiconductor Company is the newspaper. We need both. While the fabs focus on creating transistors from the base material, fabless companies go and understand the verticals like 4G, 5G, wifi, positioning, IoT, compute etc. in terms of technology and market. They develop Intellectual Property in these domains.  They create the circuits and come up with all the patterns necessary to directly enter the fabrication process. They pay the fabs to make these chips. They own the chips that are so manufactured by the fabs. They put the chips in larger systems and create reference designs for equipment manufacturers to whom they sell the chips. 

While it is acknowledged that building a fab requires a large investment, what is sometime missed out is that, with the complexity of chips crossing hundreds of millions of transistors, designing a chip costs a lot of money as well. On an average, semiconductor companies spend more than 100M$(INR 800Cr) to make a complex System on Chip (SoC) before they start mass manufacturing it.  Ultimately, since customers for the fabs are only the fabless companies, the cumulative capital requirement of the fabless companies that use a fab is in fact even more than that of the fab. They need to design the chips, fund their own R&D and also pay the fab (on which the fab has to sustain itself). Once established, the fabless ecosystem needs to be very robust to sustain the fab efforts as well. Otherwise, we will have huge expensive clean rooms sitting idle. 

Notwithstanding the requirement of filling our fabs in the future, promoting Design Led Innovation in fabless semiconductor companies right away is very important for a secure future. Communication, automotive, compute, entertainment, almost every application of the day involves usage of semiconductor chips. Many of these technologies are crucial strategically and almost all are important from an economic perspective. Most of these technologies take very long time to design, test and prove. Also, with the current geopolitical situation, there is a large requirement for chips and equipment from trusted sources. This has opened up a new set of opportunities. Being the world’s largest democracy, India can tap this situation to place itself as a trusted supplier of chips and equipment for key sectors like telecommunication, networking, data centers, AI and many more that are fast emerging. 

One option is to let the industry take its course and leave it to the entrepreneurs to figure it out. However, if quick results are to be achieved with certainty, a conducive and supportive policy framework is necessary. A while back I wrote about semiconductors being the forgotten future of India. Very soon after that, a flurry of activities started, almost as if to prove me wrong. We are seeing a definite shift in the focus on Semiconductors with policy interventions from the Government. The recent semiconductor policy looks at creating fabs and also the fabless ecosystem in India. Many incentives have been notified by the Government of India (GoI) for the fabless chip industry. The Design Linked Incentive program (DLI) from MeiTY is a very good scheme to support start-ups trying to build chips from India. We also have the DCIS scheme by DoT which is supporting technology development for telecom, chips included. Also, DoT is supporting Indian companies to do field trial of their chips/systems in live environments through pilot projects. MeitY has a Chips2Startup program targeting new chip companies. More such schemes are needed. Also, there is a requirement to acknowledge the huge scale of investment that is necessary for a Fabless Chip Design Company to do what it takes. As I mentioned earlier, on an average, semiconductor companies spend more than 100M$(INR 800Cr) to make a complex SoC.  The current DLI scheme reimburses 50% of the R&D expenses up to a limit of 15Cr. While this kind of support has never happened in the past, the quantity might need to be increased for complex chips. There still remain some barriers like minimum revenue requirement for eligibility to participate in Govt programs etc. that needs to be addressed.

Also, achieving all this will require us to make the mighty Indian design power work for India, which leads me to the next thing I would like to touch upon. Chip design is a complex “super-skill”. This is a field that requires the top talent in the country. While there is no dearth of smart engineers in India, there are very few who choose to work for an Indian start-up. We need to get the best talent pool work for Aatmanirbhar Semiconductor Chips. There is a need to motivate the top talent to contribute to this. Schemes to incentivize top talent to work for Indian companies could go a long way in India realizing its Aatmanirbharta.

All said and done, we keep moving along with what is there. And what is there today feels better than what was there yesterday. For the few who are trying, there are more voices supporting them. Slowly, the “shabhashi” is turning into real help. As an eternal optimist, I would like to think tomorrow is going to be even better. The momentum that has picked up will thrust us into a new orbit. What started in 1890s as the contribution of a radio receiver to the world can become “Aatmanirbhar Bharat’s chips for a Samyukta Jagat”, Indian chips for a connected world!