India’s Semiconductor Fab —To Be or Not To Be…

India is facing tough economic challenges today, the hardest being the regeneration of its manufacturing sector, something which has been dramatically crushed over the last decade and a half. The sector-by-sector analysis is dismaying.

Electronics sector is one such sector where the nation is exposed to huge imports (estimated at $400 bn by 2020, larger than India's oil import bill!). The Department of Electronics and IT has prepared a comprehensive National Electronics Policy to kick-start the dormant manufacturing sector and has a clear articulated goal of `25,000 crore of manufacturing investment in the 12th plan period. This powerful and strategic policy aims at both the demand side initiatives as well as supply-side initiatives to set off the program in a holistic manner...and covers the following elements:

• Provides preferred market access to government procurement in a commercially fair manner for domestic manufacturers
• Looks at manufacturing investment grant for selected industries to mitigate disabilities and attract investments
• Takes into account India's semiconductor fab to put in place the non-existent semiconductor sector in the country
• Is considering people capacity building in the sector by supporting creations of the sector skill council
• And looks at creating a self-sustainable cycle of R&D, innovation, and manufacturing via electronics development funds

Each of these initiatives dovetail brilliantly into each other, thus completing the picture and strategy and are designed to send a clear message of the Indian commitment to the global investors, a key need in today's era where manufacturing investments are aggressively competed for at the national level.

This has been coupled with workshops across academia, industry, and government to brainstorm options for indigenizing selected product categories.

Already, some limited success is visible. DEIT has received `4,300 crore of investment proposals-but it still has to pull in an additional `21,000 crore. Obviously, a pray-and-spray opportunity cannot work and it needs to plan meticulously to take some bold decisions. You cannot get another major investment in manufacturing computers at this stage-DEIT will need to look at exploring untouched categories.

Semiconductor manufacturing is a logical step planned in the NEP and incorporated in the 12th plan. There is a strong demand in India, enough to satiate 3 fabs, and growing at a CAGR of 30%. So is the strategy aiming at creating an Intel 14nm resolution 3D chip? Obviously not! The trick is to focus on the mainstream resolution chips, allied market sectors and extend the effective tool life by playing strategically across selected applications. The aim was to work not just on the logic space where Moore's law applied but a select more-than-moore applications.

It is relevant to note that a substantial part of the semiconductor market plays on the 80nm resolution and is migrating to 65nm now and will migrate to 45nm in future. This migration is a natural progression in the semiconductor industry-a result from industry alignment on a technology roadmap approximated by Moore's law (transistor density doubles every 18-24 months).

An empowered committee has done due diligence over two years, evaluating options from a global RFP, and slowly focusing on select consortia in partnership with some of the leading names in the sector. This recommendation is now under government review, but is being buffeted by headwinds driven by the naysayers and lobbies that do not wish to see India enter into this manufacturing arena.

Some of the arguments

• India cannot manufacture semiconductors as our previous attempt was not a success. For a nation which has its own space and nuclear program, this thought is a shame and clearly not true. Our previous attempt at SCL Chandigarh was a government initiative which was impacted by an industrial accident and obsolete capacity that was never upgraded. It is for this reason that the current initiative was planned with consortia of private sector players with global participation.

• Presently, there is global overcapacity. Therefore, India should not enter into this venture. We are surprised at this sweeping and erroneous statement. The attached data clearly shows that the above statement does not hold any veracity. As actual foundry capacity utilization reports show, the overcapacity was in 2008-2009, otherwise, the industry is typically perpetually operating at 80% capacity utilization. This is also a macro picture, the maximum under-utilization is at the cutting edge and the obsolete edge resolutions. In fact, the global industry is increasing capacity in sync with the projected demand and typical capacity utilization is not having any aberration/ overcapacity dip.
• According to the Indian data, there is a strong semiconductor demand and an opportunity for building self-sufficiency and import substitution capability in select areas. Moreover, the players who would setup an India fab would not be restricted to planning for the local market.

• We can divert the resources to developing soft fabless sector. That will be more strategic. Our BPO/ IT services industry is already under pressure. If we do not have an India fab, we miss on a key part of the ecosystem that our regional competitors are trying to develop.

Imagine that a fabless designer sents a chip design to a fab abroad - and there emerges an issue with the production lot. Is it a design issue or a process issue? How will we perpetually expect our designers to respond to a remote customer!

The fabless industry will only get strengthened by a local fab. Its absence is expected to become a major structural weakness for the industry in the medium run!

In fact, the EDF daughter funds will see investments come from the leading global players...when we back away from the fab, we should expect that a number of semiconductor global players will back away weakening the corpus for the fabless sector as well.

• The planned semi-capacity will become obsolete. If it is planned in the project, why should it be an issue? This is an industry cycle that can be witnessed in so many sectors.

• It does not make sense for the fabless industry to design at an obsolete resolution. That is a motherhood fact. However, not all design is at the 14nm 3D chip level. Bulk of it is at lower resolution. No one is asking anyone to design at an obsolete resolution-this is also raised primarily to throw a show-stopper into the decision.

WHY SHOULD INDIAN GOVERNMENT INVEST IN THE FAB

Hi-end technology industry is wooed across the globe by nations. Singapore, a tiny nation, has evolved its own semiconductor program. We are all aware of the success of Taiwan, Korea, and Japan. India has the advantage of a strong demand which has been leveraged to enable a window of opportunity for this project. Why should we procrastinate perpetually?

It makes more sense for India to import than to manufacture. This is definitely true from the perspective of other countries.

So why is the India fab so relevant for our country? Besides these facts, India has to target `25,000 crore of manufacturing investments and it has currently got proposals for `4,300 crore. The fab decision at one stroke will catalyze a manufacturing investment of `17,000 crore-it's catalytic impact will enable India's aspirations.

If a fab is so logical for India, a question of debate does not rise at all? Well, there are vested interests who do not wish to see India get into semiconductor manufacturing. Then, there is the famous Indian crab effect. Unfortunately, several years of work in this arena could be diluted by a chronic decision paralysis and safe play. If we miss on the opportunity now, it is likely that we will lose it for a decade, something our global competitors would love to see.

The time to soliloquy is over...with a little luck, and determination, the benefits of the decision are so obvious and the cost of missing the opportunity so high, India may actually keep it's tryst with the India fab.