S. Uma Mahesh: India's semiconductor industry has lots to catch up!

We are all probably rejoicing at India's very late entry into semiconductors. Agreed, the starting point is at a lowly 28nm node. Have we also asked the question: Why hasn't India managed to get into nodes higher than this? 

First guesses are that a Rs.76,000 crore ($10 billion) outlay may be very less for any significant investment in semiconductors. A latest fab may cost around $25 billion alone. I am not an expert, and leave the experts to decide on the actual cost of the latest fab. Next, 28nm is perhaps the lowest entry point right now. We have to stay at this node for some time. There is no point in speaking about what higher node entry points can mean.

S. Uma Mahesh, Co-founder and Director of Innatera Nanosystems, had more to say about India's journey in semiconductors. India's journey in semiconductor, software, etc., had begun because some Indians were of prominence in the parent ecosystem, who either inspired or pushed 'India as an option', TI in India, is one huge example.

This was initially at the entry level, and then come 'managers'. The highest level decision making always remained(s) at the HQ -- even for the so-called 'HR" where its a norm that low-qualified personnel lead HR from HQ, with XLRI and Management Institutes who have far superior graduates reporting to them.

Guarding shepherds?
Now, the senior posts have to remain in the HQ, and low posts continue to increase n India, with the middle being the next in line to be reduced, naturally. So, the attempts of the 'middle level' at HQ is to do everything to ensure to keep these roles there.

Hence, they never really encouraged decision making, rather, taking the ownership at the mid-level. They are more of managers who guarded the 'shepherds', so that the mid-level posts remain in the HQ.

This is reflected, across the decades, in decisions/roles on architecture, and later, design was far more than required or fair or right, remaining in the HQ.

Strikingly visible in EDA
One area where this is strikingly visible is in the decisions about the EDA flows, design methodologies, what tools to chose, experimenting with and identifying new tools, etc. These are schematics to HDLs to ESLs to formal verifications, to trying new tools to, and the list goes on.

Hence, continuing this time-tested, well-established trend, will be obvious in this aspect of 'python/GPT 'included' design/verification' methodologies, processes. Decision making remains at the HQ. Very few, if at all, will want to take the ownership, drive the initiative, to make India play a leadership role in ensuring this. This is not just the US, but also from companies from all nations/locales.

Considerable innovation is now happening from India. There are some illustrations for sure. TI did the entire innovation from here decades back (digital lens projector being one). Intel developed some of its biggest processors from here. There are such, in other companies too.

What's the value add?
But, it remains few and far between, sort of an exception, than a norm. After almost four decades, what is the value add percentage by the Indian semiconductor industry overall? Could (and should) it have been higher? There are many reasons, which can include the true limitations of Indian counterparts, in not having more done from India, other than 'verification', regression types.

One such is the 'cost' and the 'value' companies can get, particularly for latest and game-changing technologies, systems, architectures, etc.

It is quite possible that the Indian ecosystem is not quite 'direction setting' in latest of technologies, as other nations do. Hence, that is bound to map to the lesser levels of contributions from India-based teams, as well. The extent to which they are held back (to execute implementation) is more than that can be attributed to they 'not being ready enough'.

Even now, chances that the latest of design methodologies, the EDA tool environment choices, all are to be 'controlled' more by the HQ teams. This goes across the nations!

Essentially it is all about:
* Levels of value add / contributions from India for global products -- are they in proportion with the overall growth seen in India over the last four decades?
* If they are deemed to be lesser than what they could have been, what are the possible reasons, and what can one do about them ?

Particularly, in this era, when semiconductor is now known to the grandmother and grandchild as well, and when India is making determined and concerted steps for highest value add through manufacturing, we need to ask: is India ready for actually making products defined, and developed in in India for its very own fabs, as has happened in all economies?

Need to introspect!
If not, why? Should this be introspected upon to take large, game-changing strides ahead?

India has grown more because of the Western influence. Hardly any new idea or product has happened from India.  How will the Indian companies work on 28nm, when they are so dependant on the others for everything?

S Uma Mahesh said: "Precisely, and why? Is this trend beginning to reverse? Can it? What can we all do? By those, like us, who have had the honor of being with it from the earliest days, how can we inspire and guide the current and future generations. This is the time to reflect and introspect."

Indeed! There should be lot of introspection done for semiconductors in India. Hardly anything has been done from India, so far. All that the current Indian government has done was to give money for some fabs. The trend is not reversing, yet!

Where are the R&D institutes that can handle the challenges posed by semiconductors? If at all, are they even having investments done? What is the status of new materials for semiconductors? Are we even starting to talk about materials, for that matter?

 It will take lot of effort from stakeholders. We also need to look at EDA, advanced packaging, photonics and silicon photonics, semiconductor materials, and so on!