Potholes in the silicon design

Date:   Friday , February 02, 2007

Though the semiconductor industry in India is close to twenty years old, in ways more than one, it is still in its infancy. There is no denying the design capability of Indian firms, but there is a huge opportunity to further improve the maturity of semiconductor companies in India. I see a lot of them doing chip design as part of a large geographically distributed organization. Complete and end-to-end ownership of the design cycle in India is still very rare.
While we are good at design and related fields, we could be a lot stronger in gaining an end-to-end business perspective, and also in being able to architect. For most of the chips that are designed and developed by professionals in the country, the architecture is defined overseas. One of the primary reasons for this is because the end market of the silicon is not present here.

First, for us to architect, we need to understand the end market, the needs of the users and come up with the required specifications, which can then be factored into the conceptual space. Towards that end, we could look at sending personnel to advanced markets such as the U.S., Europe and Japan. This exposure will give them a precise idea about the needs of the market, which can help in pushing India up the silicon design value-chain.

Secondly, one can look at a chip from the standpoint of digital and analog components. While we are good at the former, and a lot of start-ups are also being launched in that arena, there is a need to steer the competency development vehicle into the analog domain. This will help in developing competencies on a uniform basis and aid in companies doing end-to-end chip design in the country.

Lately, there has been a buzz about setting up a semiconductor fab in India. Though I am not a proponent of this idea, there are a few tangible benefits that catch the eye. With the setting up of a fab, one hopes that the ecosystem would improve considerably. One example of such a benefit would be that the semiconductor component industry will take hold in India.

There is a lot of dependence on hardware for certain kinds of software design. One develops software for a certain board and every time there is a modification to the board, the software change in tandem. If the board design and modification is made overseas, as is often the case, it could mean several weeks before the board returns to India with the modifications. This whole process, iterated at least a few times, results in significant delays in the development cycle and hence has a huge bearing on getting the product out in a timely manner. Importing components like resistors and capacitors result in further delays.

If there were a manufacturing facility in place in India, along with a strong component infrastructure, the time lag would be minimal.
At the same time, setting up a fab would call for investments to the tune of $2-3 billion. Depending on the technology used, it would call for massive investment in infrastructure, and significant technology upgradations every 18-24 months. Considering the stage the semiconductor industry is in India at present, it would be desirable instead to set-up fabs only for testing purposes. This can speed up the development of the final chip, which can then be shipped to other countries for large-scale manufacture.

The author is Vice President and Managing Director (India), Ikanos.