BANGALORE, India – Several weeks back, a Palo Alto-based venture capitalist specializing in semiconductors said partnering with Israel rather than with U.S. or Taiwanese companies would be a great idea.
According to Renu Raman of Tallwood Ventures, partnering with Tower makes sense since Israel has many tech startups, fabs and a highly developed IC ecosystem. Tapping into Israel’s infrastructure would ease development of a similar ecosystem in India rather than just building a fab. “So I would start looking at partnering with Israel rather than the US or Taiwan,” Raman said. “Israel needs another market. It is a technology source and not a manufacturing place.”
Tower’s plan to build a 300-mm fab in India is the latest twist in a long and so far fruitless effort to develop an Indian chip infrastructure. The SemIndia consortium plan for establishing a Fab City in Hyderabad went nowhere. Talks with Intel failed when neither the Indian government nor the chip giant could get what they wanted out of a deal.
Next came talks between the Indian and Taiwanese governments. According to sources here, there was a plan to send an Indian delegation to Taiwan to solicit ideas for a fab joint venture in India. The Indian government promised access to it huge markets for defense, transportation and mobile phones. Talks collapsed, sources said, because New Delhi declined to specify the amount of Indian business the proposed fab would gain.
That’s a problem since one of the first things global investors want to know is the size of the potential domestic market. That figure is something government officials have yet to put it down in paper. The talks with TSMC and others in Taiwan ended there.
That was yesterday. Today, the entire landscape here seems to have shifted as the government gets serious about boosting domestic chip companies. Sachin Pilot, Union Minister of State for Communication and Technology, warned the India Semiconductor Association (ISA) recently that if India doesn’t set up a fab and electronics ecosystem within the next five years, it never will. Pilot, along with senior minister Kapil Sibal and a well-meaning team of bureaucrats, are working to get the latest initiative off the ground since the cost of entry will be out of reach five years from now.
Still, many here remain skeptical about yet another attempt to set up an Indian fab. “As an Indian, I would be proud to have a fab in India,” said one observer. “A few things have to be set in the right context. Planning to set up a fab in India is like wanting to set up a five-star hotel in the middle of a desert.”
Others are more upbeat. “It is just not the manufacturing of chips but a means to an end to realize value-added manufacturing in the country to generate intellectual property, product design companies and a complete ecosystem,” noted former ISA Chairman Rajendra Khare. “The government t has to stand firm behind this decision for the next ten years. Only then will this take off.”
A senior government official said there would likely be a mandate for large companies like Broadcom and Intel to commit to using a portion of a proposed Indian fab in exchange for preferential market access.
Total Indian demand for electronics could reach $400 billion by 2020, while domestic production would account for only about $104 billion. Current consumption of electronics is about $45 billion, with semiconductor components accounting for as much as 20 percent of the total.
So, is an Indian fab really akin to a five-star hotel in manufacturing-barren India?
“If somebody has a vision of setting up a five-star hotel in a desert and creating a whole new city around it with the idea of someone using [the five-star hotel], then the analogy fits well,” agreed UmaMahesh, CEO of product startup Indrion. This is precisely the scenario facing India today in setting up a fab, he added, but the vision is to create an entire ecosystem around the fab just as a city might bloom around a new, five-star hotel.
Moreover, wasn’t Israel built on a barren strip of land? Yes, but then it grew to become a homeland for the Jewish people that has prospered far beyond what many would have guessed.
It’s clear the Indian chip market is big and getting bigger. The $5 billion question is whether the government can finally get its act together and close the fab deal with Israel’s Tower Semiconductor.
India also has special need to have fab for its defence and space program. This may be prime reason for government involvement. Apart from this, it may be very difficult for fab to be successful in India.
Why would India even want a Fab? It is a huge capital investment and if you manage to get enough volume of customers to justify the investment, it is time to re-invest in the next technology node. It is also getting harder and harder at lower geometries to scale in volume - take a look at the 32nm problems at GlobalFoundries.
The Indians have perfected the S/W outsourcing model and are making money hand over fist. It is a low investment, quick ROI business, why mess with success?
That is true! India has a huge pent up demand in almost everything. Same was the case with telecom when people had to wait for 10 years to get a telephone connection- Came Sam ( Not Uncle Sam!) and just in a matter of couple of years all the scenario changed.
There is money with the government. Infrastructure can be created and the huge project such as setting up of a fab can well be executed on schedule = we just need another SAM!!
IF we have a mission man like Sam Pitroda ( who was responsible for the telecom revolution in India and who enjoyed 100% support from the then Indian prime minister) behind this project then this project will definitely succeed.
Call me a cynic... I don't believe a fab will ever come up. What will happen is somebody at consortium will get paid off, Indian politicians will have a new revenue scheme, a few poor souls would have a cushy job for sometime and that will go away.. and in the end there will be a new empty fab city, with whole lot of folks owning useless land around it (these will be the suckers who bought the land).
David Patterson, known for his pioneering research that led to RAID, clusters and more, is part of a team at UC Berkeley that recently made its RISC-V processor architecture an open source hardware offering. We talk with Patterson and one of his colleagues behind the effort about the opportunities they see, what new kinds of designs they hope to enable and what it means for today’s commercial processor giants such as Intel, ARM and Imagination Technologies.