Bangalore: The two semiconductor fabs set to come up in India seemed to have encountered numerous roadblocks and are likely to be non-starters.
But as the saying goes: If one door is closed another window opens up.
Cricket Semiconductor - a company founded by two former Texas Instruments executives - created news early this year when they announced the setting up of an analog fab in India with an initial investment of around $ 1 billion. Now, they plan to break ground in 2016 on the wafer fab and begin producing chips for customers in 2018.
It may be recalled that there were two consortia – one made up of India's Jaiprakash Associates Ltd., Israel’s TowerJazz, and IBM and the other with HSMC Technologies India Private Ltd, Malaysia's Silterra, and STMicroelectronics – which were supposed to set up chip manufacturing plants in India. These two projects, with a total investment of around $10 billion, were announced about two years ago.
But it looks like these two projects are also going the SemIndia way. In 2005, there was a false start involving SemIndia, a consortium of Indians living overseas and having a collaboration agreement with AMD. It faltered and died a slow death.
According to an earlier news report these two projects have run into rough weather on account of "several deficiencies", as pointed out by a high-level official committee.
The official-level panel has found the detailed project reports (DPRs) submitted by them “non-satisfactory". The two consortia have not been able to fulfil the conditions required to begin the projects. They are yet to submit 'other documents, which are considered essential to demonstrate their commitment to the projects, the report added.
There seems to be some sort of a jinx when it comes to India and semiconductor manufacturing. Starting with SemIndia, there has been a series of non-starters when it came to the setting up semiconductor manufacturing fabs.
After SemIndia’s 2005 initiative, there were reportedly talks with Intel Corp in 2007 to set up a facility in India that collapsed at an advanced stage. Later, the government formed a new semiconductor policy, which also failed to attract investors. In 2011, when it invited fresh proposals, it did not receive a single proposal from any global chip manufacturing company. Only Jaiprakash Associates and HSMC had bid. In 2013, the government again invited more proposals, but failed.
Says Mark Harward (one of the two founders of Cricket Semiconductor), “We do not want to be painted with a broad brush. From our standpoint we are making very good progress on all fronts. We are talking with customers, investors, state and central governments and are very pleased with where we are right now. That’s the message that we want to convey. We are a different entity from the other two digital fabs and we think that we have a good story to tell. I know there is a lot of scepticism about wafer fabs in India. But we are different and we have actually converted a number of sceptics to supporters. We have shown them our plans about what we intend to do with our fab and why.”
Interestingly, Mark Harward and Lou Hutter (the other founder of Cricket Semiconductor) have been visiting India for over several decades and watching the mounting frustrations in the Indian semiconductor manufacturing space.
Lou Hutter, CEO, Cricket Semiconductor, has a total of 35 years in the semiconductor industry and has spent 29 years at Texas Instruments Inc., until 2007. As director of TI's Mixed-Signal Technology Development, he was responsible for their worldwide analog, power, and mixed-signal technology development. He joined South Korea’s Dongbu Hitek in 2008 as senior executive vice president. He has 47 patents to his name and has co-authored a textbook on silicon analog components.
With over 25 years in the semiconductor industry, and 14 US patents, Mark Harward has spent over eight years at TI, working in the R&D group's VLSI design labs and managed new product development in the ASIC business. Since 1994, he has been a serial entrepreneur. He founded TestChip Technologies in 1994, which focused on automating the development of test chips used in silicon technology development, and which was later acquired by HPL technologies.
“A couple of years ago, Lou and I were talking about the two projects that were announced and we were wondering how that makes sense for India. A digital fab - even at zero cost in India - would not make sense because the cost of developing products for digital leading edge fine geometry technologies is so expensive that only large multinational companies could afford it. We felt that a specialty fab like an analog one, with a much lower cost of development, is more likely to succeed. That’s when we decided to set up an analog fab in the central state of Madhya Pradesh,” he says.
Although Cricket Semiconductor has not officially disclosed the size of the wafer facility, industry experts expect the analog fab to operate on 180-130 nm technology. Also, since large economies of scale are required to drive down costs, this fab is slated to produce 60,000 wafers per month. At this sort of volume, the fab could become cost competitive compared to the digital fabs.
Today, most tape outs are now around 180 nm. While some volumes (like consumer products and few software applications) might be moving to on 130nm the sweet spot still hovers in the range of 180 -130 nm.
“That’s six generation behind digital. But the beauty of analog is you don’t have to depend on the latest tool technology. Analog is based on power voltages and precision. It doesn’t care much about tools – it is more about components. It is about building different types of transistors and you don’t need 14 nm to do that. You need 180-130 nm so the cost is significantly less,” explains Lou Hutter.”
Everything about analog is the lower cost of building the fab, the cost of keeping the fab up to date and technology and product development.
“We believe that this would act as a catalyst for developing the ecosystem in India. There is absolutely no supply chain mechanism in India – almost everything comes is imported and gets assembled here. Moreover, with the rising middle class and the growing consumption India needs to create its own supply chain ecosystem. We think building a fab would be a catalyst for fabless companies, packaging companies and eventually for product systems companies to take root and eventually be able to build products from the concept to packaging stage here,” he adds.
They believe that analog is a great market to be in because the real world is most definitely analog. “Most people today might think the world is going digital but in fact the real world is analog. When I talk to you, it is analog signals that you are hearing. So are motion or humidity or temperature signals,” Harward adds.
Most people think the cell phone is the epitome of digital technology, but in fact there are more analog chips in the cell phone than there are digital chips and the bill of materials for analog is equal to or larger than the digital chips in a cell phone.
In fact, analog is alive and well and is actually a growing market.
“But we don’t talk about it much because it is not at the leading edge. Digital is the kind of flashy and makes a more powerful statement. Analog is less flashy but is a perfect answer to what India needs. So with our background in analog we felt that it is a perfect match for India’s market,” Harward notes.
Incidentally, if one looks at India’s strengths which lie primarily in automotive, industrial market and mobile phone one would realise that analog neatly fits in this scenario.
“Analog is a perfect fit for India and that’s why we proposed this fab to the Madhya Pradesh government. It just seemed like a natural strategy for India,” Harward adds.