Need for logic foundries and hybrid strategies

It’s been over six quarters since the much-awaited Indian Semiconductor Policy from the government was announced. It was a winning formula for success. The relatively unheard of public-private partnership in the Indian business circles ensured a ‘Special Economic Zone’ status for fabrication units along with a 20% of the capital expenditure coming from Indian government. Moreover, the policy covered manufacturers of all types of semiconductors along with test and assembly facilities.

However, the overall response to the policy has been somewhat of a mixed bag. On the positive front, the policy has already attracted close to $8 billion in investment. These investments have largely come in the form of solar photovoltaic (PV) cell manufacturing facilities. Looking at the current scenario in India, these facilities hold tremendous potential. In a country with more than 70% of the population living in rural areas, energy availability has always been a problem with only 40% of the rural households having access to electricity. Availability of new renewable energy source for farms and other rural applications is likely to have a positive impact on the agricultural sector.

Something amiss?
Worldwide, the $270 billion semiconductor industry has been the key in driving the larger ($1.3-$1.5 trillion) electronics market. Semiconductors have over the past 20 years become the most important part of every electronic product. Completely driving the computing and consumer electronics not only have the semiconductor foundries enabled complete test and package ecosystems around it but also played an important role of promoting the larger semiconductor ecosystem of complete product design and manufacture.

Looking at the investments the semiconductor policy has received, the focus seems to have largely missed the mark with the electronics related industry. With the exception of SemIndia, there is little mention of creating logic ICs for providing the necessary fillip to the Indian electronics industry.

The apprehension though, is justified. With increasing number of fabless firms combined with more semiconductor companies moving to a fab-lite model, the consensus emerging is to leave the manufacturing to the experts who can also justify the volumes required to breakeven on the multi-billion dollar investments.


Rise of fab nations

Looking closely at a sample of three nations with fabs near to India, we can see the focus areas differ in different fabs.

TSMC (Taiwan): The largest pure-play fab in the world, TSMC has almost managed to carve out a ‘made in TSMC’ branding in high volume semiconductors. Established in 1987 by Philips jointly with the Taiwanese government, TSMC has been continuously investing in the cutting edge process technology. TSMC’s presence enabled Taiwan to become the first country in the world to offer professional foundry manufacturing and package process.
Based on the foundry expertise, Taiwan rapidly realigned and developed other parts of the value chain (design, manufacture, packaging and testing) to achieve an extremely competitive position.

SMIC (China): A relatively new foundry started in 2000, the Chinese government’s focus on SMIC has led to formation of a large number of small design houses. SMIC’s focus has been on mainstream technologies that enable it to extend its services to smaller IDMs also. SMIC also runs a ‘shuttle service’ that enables its customers develop and prototype new design development whilst spreading the cost of manufacturing.

Tower (Israel): Started in 1993, Tower takes pride in being a specialty foundry focused on advanced non-volatile memory solutions, mixed-signal and CMOS image-sensor technologies. Looking at Niche markets, Tower provides significant value for clients that are looking at optimal costs and have products that do not require the cutting edge 65nm/45nm process. This also allows Tower to address their customers’ needs without investing in the newer processes.


Follow a hybrid strategy?

What would be the ideal strategy for an Indian fab? Should they go for the state of the art process technologies while looking for large volumes?

Going for the ‘glamorous’ model of pushing the nanometer limits combined with looking at very large volumes might not be the best strategy for a new fab. The established process of the leaders combined with the high costs increases the overall risk of such a mega project.

Additionally, with an increasing number of IDMs embracing the fab-lite policy, the overall investment in manufacturing is likely to reduce. For the fab-lite model to work, this reduction should have been compensated by increased investments by many pure-play foundries. As the overall volume of semiconductor devices increases, there is a possibility of limited capacity with these pure-play foundries. While the larger players would have the volumes to hold the production lines of these fabs, the lower volume niche devices makers could find it increasingly difficult to get their devices manufactured.

Perhaps a hybrid business strategy is the best suited for starting a logic fab in India. While the niche strategy from Tower is perhaps the best suited for established product segments in India, an additional shuttle service would enable greater number of designs to emerge out of the local fabless houses. Perhaps a partnership with a IDM looking to go asset light would be the perfect strategic fit enabling quick transfer of technology.

Possibility of a domino effect:
A hybrid model for the foundry holds the potential to affect more than just address the capacity issues. With the accessibility, it is likely to open up an option for developing ICs for low to mid volume products that are produced for local consumption. The possibilities for lower cost ICs are endless. From measuring devices in public utilities such as electricity meters to even installing level indicators for cooking gas, lower cost ICs can enable solutions for many of the existing problems. Having ICs made specifically for the million unit car market and the multi million unit two wheeler market is likely to become a possibility. With the product developers moving from discreet components to semiconductors, there is a very real possibility of local fabless houses.


You can reach Mohanakrishnan at mohan.krishnan@wipro.com
and Anuj Sharma at Sharma.anuj@wipro.co