The other day, I was engaged in an interesting discussion regarding the Indian semiconductor industry. The obvious question: can fabless semiconductor take India to the top?
Well, it all depends on the definition of ‘top’! Does it mean the role of India-based semiconductor companies as a percentage of the semiconductor market globally? Or, do we take India as a system/gadget maker and thus, as a percentage of that market??
Fabrication is increasingly expensive, much involved and the actual global fabrication players (i.e. those who (also) own a fabrication plant) are declining and will be about three to four companies, and about 10, if we include all off those Chinese fabs.
And, India continues to slip back in having a ((proper) fab!
Now, India’s contributions to global electronics and semiconductors will continue to increase as the MNC subsidiary companies’ hub, and not quite as India-based companies, who are coming out with something that will shake the world in terms of that chip(s)!
If India has domestically consuming gadgets, that are more India specific, that could need devices available less outside. For that purpose alone, a local fab could be essential. However, such requirements appear less each day!
So, yes! Fabless semiconductor could be the way forward for India, in terms of contribution to its economy. However, in terms of India becoming a global player through such chips conceptualized in India, for India and the world, the chance is lesser, for now!
Well, hasn’t the Indian semiconductor industry been shouting ‘fabless’ from the rooftops for some years now? Let us see how India has progressed so far!
One, in terms of having local fab, the answer is NO! Two, in terms of increasing its percentage of contribution to global semiconductors, electronics from India, YES, an increasing role and value (though these are embedded software too).
In terms of having India-based companies working toward developing chips, YES again, in terms of smaller, analog, components that are crucial (like Cosmic Circuits), and YES, in terms of having IP-based companies (like Innovative Logic India for USB3.0) and, YES in terms of increasing service companies.
Many more companies are coming up, and some started directly here in India, such as Apsconnect, Techvulcan, etc. In terms of the actual solutions, YES again, as we have developed solutions in medical, automation, etc.
However, the answer to the question remains NO in terms of having chips come out of India, as yet!
Now, what happens to the fab-lite strategy? Well, it continues, globally. From an India perspective, it is actually in a way, validation of the earlier belief. There is less direct importance to manufacturing from themselves, but more about the actual value add they do OR can do.
Now, given this situation, let us also look at the key growth drivers in Indian electronics, especially, since we are talking about fabless and fab-lite.
The obvious one is to develop solutions for the India market. It is likely that these can be for outside markets as well. This ability will actually make India develop solutions for global markets. Also, these are not semiconductors per se, but, (embedded) solutions based on them.
The above situation can slowly lead to a fabrication and manufacturing ecosystem in India. India should also try to position itself at the higher end of the solutions, markets, services, etc., so that its value contribution can be much more.
Friends, is there a way out of the current situation that India finds itself?
Actually, this is normal process of growth in the chosen path. India continues to think about low end, less (or no) risk options of services. There is only so much growth, revenue, profit possible in those areas unless one goes up the market.
India has not done that as it could be, as an ecosystem in all. India should focus on its own internal requirements. That could mean growth and an increasing role for India, globally, as well!
Besides manufacturing, the big issue lies in marketing of such products. A senior statesman from a leading Indian electronics firm once asked me, “How will India compete in marketing of these products compared to the Chinese or Taiwanese manufacturers, who have more than 30 years of experience in these industries?”
How one wishes that India had at least two wafer fabs by now, what with the technology nodes constantly upping their ante. Even if someone does decide to put up a fab, it will be extremely expensive and has to be cutting-edge. However, as I said, one should never give up hope!
And then, there is the Modified Special Incentive Package Scheme (M-SIPS).
The newly announced M-SIPS is long awaited and much needed. The key is to now turn this ‘gazette notification’ into implementation, by the regulators, and utilisation by the industry.
It is understandable that the government can only do so much, particularly, under the given circumstances. With that kept in mind, this is a yet another good start! Hopefully, instead of just commenting on this policy, the industry sincerely works to benefit from it by properly utilizing it.
Why just think of digitalization of TV! The number of set-top boxes required across the country will be huge! Or, think of electrification of roads all over India. The number of LEDs required are likely to be massive. These are just two examples of the many possible. The Indian electronics industry needs to move fast, and now!
Hasn’t all of this been very easy to say, difficult to manage!
According to Dr. Walden (Wally) C. Rhines, chairman and CEO, Mentor Graphics Corp., while fabless startups have declined substantially in the West during the past decade, they are growing in India.
Given the time required to grow large fabless companies in the past, India should not be discouraged by current progress. India has key capabilities to stimulate growth of fabless companies, such as:
* Design services companies.
* Design engineering expertise and innovation.
* Returning entrepreneurs.
* Educational system.
Semiconductor frustrations abound! I recall a discussion in mid-2005 where an industry expert mentioned that fabless was the way forward for the Indian industry! Between then and now, fabs were supposed to come up, but they failed. Nevertheless, one must not give up hope!
As of now, there seems to be too much focus on services, multinational company dominance, perceived lack of progress, perceived lag compared to China, lack of foundry infrastructure, and no clear dominant indigenous Indian company.
Of the top 50 semiconductor companies in 2011, 12 are fabless and four are foundries. Fabless IC revenue has been growing at 17 percent CAGR since 1997 and will continue to grow. Even the fabless market has been gaining in the overall market. However, the fabless revenue is said to be highly concentrated. He added that the leading fabless companies specialize and average ~23 years since formation. Also, the VC funding for fabless semiconductor companies has been declining in the West. As for the number of fabless companies, the GSA put it at 1,200 companies, at the end of 2010.
According to Dr. Rhines, the semiconductor IP market would grow to about $3,707 million by 2015, at a CAGR of 14 percent. The leading semicon IP players specialize and average 22 years in business (similar to fabless).
Now, India is said to be among the top five semiconductor design locations worldwide (SIP + fabless + design services). Also, India is a leading source of semicon IP, accounting for 5.3 percent globally. From the looks of it, India seems to have built a foundation for a fabless future. India can well become the next great fabless incubator! Read more…
The last decade heralded a dramatic transformation in supply chain dynamics, driven by the complexity challenge of staying on the More Moore curve. On the demand side, the high cost of fabs persuaded almost all integrated device manufacturers (IDMs) to use foundries for their leading-edge wafer supply.
The ever-increasing process complexity and its negative impact on manufacturing yields forced the adoption of sophisticated foundry-specific design-for manufacturing (DFM) techniques, effectively committing new chip designs to a single foundry and process.
At the same time, the industry adopted a much more cautious lagging rather than leading demand approach to new capacity expansion, resulting in under-supply and shortages in leading-edge wafer fab capacity. To make matters worse, the traditional oxide-based planar transistor started to misbehave at the 130nm node, as manifested by low yields and higher than anticipated power dissipation, especially when the transistors were supposed to be off, with no increase in performance, heralding the introduction of new process techniques (e.g., high-k metal gates).
Even before these structural changes have been fully digested, supply chain dynamics have been further disrupted by the prospective transition to 450mm wafer processing, to extreme ultra violet (EUV) lithography, and from planar to vertical transistor design.
Since the start of the industry, adding more IC functionality while simultaneously decreasing power consumption and increasing switching speed—a technique fundamentally known as Moore’s Law—has been achieved by simply making the transistor structure smaller. This worked virtually faultlessly down to the 130nm node when quite unexpectedly things did not work as planned. Power went up, speed did not improve and process yields collapsed. Simple scaling no longer worked, and new IC design techniques were needed.
While every attempt was made to prolong the life of the classic planar transistor structure, out went the polysilicon/silicon dioxide gate; although this transition was far from plain sailing, in came high-k metal gates spanning 65nm-28nm nodes. Just as the high-k metal gate structure gained industry-wide consensus at 28nm, it too ran out of steam at the 22nm-16nm nodes, forcing the introduction of more complex vertical versus planar transistor design and making the IC design even more process-dependent (i.e., foundry-dependent). Dual foundry sourcing, already impractical for the majority of semiconductor firms, will only get worse as line widths continue to shrink. Read more…
I received an interesting news alert from the Global Semiconductor Alliance (GSA), formerly, Fabless Semiconductor Association, which spoke about how fabless companies, only, were funded in July 2011. Well, it also led me to this feeling that each time there is any new electronics or related segment being talked about globally, it seems that the Indian semiconductor industry is slowly losing the plot! One surely hopes not!!
May I take your attention back to Pravin Desale’s speech during Mentor Graphics’ U2U conference of December 2009. He had cited some numbers during his speech, borrowing heavily from GSA. According to the GSA Dec. 2008 figures, distribution of fabless IC companies is: Canada 29, USA 606, Europe 151, Israel 61 and Asia 510 — China 222, Taiwan 196, Korea 47, Japan 16, Taiwan 16, Singapore 7, Malaysia 4, and India 2-28 (two fabless companies and 28 design services/IC providers).
Agreed that these numbers should have changed a bit, if not, a lot.
Now, when the India Semiconductor Association (ISA) was formed on Oct. 28, 2004, members at the BangaloreIT.com that November, dwelt upon the need for fabless companies.
Somshankar Das, e4e, had said that for building fabless semiconductor companies in India, the country had a major advantage, as Indian talent was a large part of the global semiconductor industry. Some other advantages in favor of India at that time (Nov. 2004), were: local IC design service firms, who were creators of selective IP as well. Development of smart chips with embedded software was ongoing. Next, the US-funded cross border semiconductor firms were setting up development centers in India. Dr. Bobby Mitra, TI, had cited the need for microelectronics as the national agenda.
Well, where are we today? Why hasn’t the fabless semiconductor industry grown in India? Just two months ago, I wrote on how, China’s fabless market was set to double by 2015. Is any such movement even happening in India? At least, I am not aware, in case it were!
The establishment of fabless semiconductor companies is one good way to drive the growth of the semiconductor industry in India.
I still have the photograph of the founding members of the ISA, which is pasted above. The original participants were: Dr. Ananda, Dr. Madhu Atre, S. Uma Mahesh, Rajendra Khare, Dr. Sridhar Mitta, Dr. Anand Anandkumar, V. Veerappan, S. Janakiraman and Dr. Satya Gupta. Today, the ISA only has Dr. Satya Gupta as a representative. One hopes the others have not been left behind in the run of events following the ISA’s formation!
By the way, why am I referring to the original ISA, and fabless companies? Perhaps, there is a very deep significance!
I received a report from IHS iSuppli, which stated that China’s fabless market is likely to double by 2015! Well done, China, and all kudos.
According to the report, there are 3Cs - China, Consumer and Convergence — that China has been focusing on. However, there are three more Cs — Culture, Content and Contribution – that the report urges China to focus on.
The report states: “The companies must accommodate and adjust to the differing cultures of overseas customers. They must learn more about end-content sectors that drive the growth of technology markets. And China’s fabless firms must take advantage of government contributions to the industry, given that Beijing has instituted a range of policies designed to improve the fabless industry in areas including investments, tax rates and capital investments.”
Having taken its own sweet time, the Chinese fabless industry is coming up well, and fast! However, what is the state of the fabless market in India?
If one goes back to Dec. 2009, Pravin Desale, VP and MD India operations, LSI Corp., while speaking at Mentor Graphics’ U2U conference, had said that India has only two (2) fabless firms. If this number, or even a number below double digit is considered to be correct, then India surely has a lot of catching up to do!
Now, who are the leaders in the fabless semiconductor business globally? That’s Qualcomm, Broadcom, even AMD (as per 2010 reports), MediaTek and Marvell. Four of these companies are based in the USA, leaving MediaTek as the only Asian (Taiwan) representative.
Do Indian companies have the necessary experience of designing complete chips from scratch? Perhaps, some may have. Can the Indian companies get the silicon manufactured easily (local manufacturing) and later, debugging it? Here comes the first major gray area! Do the Indian companies have easy access to tools? Perhaps, some, most of them, have that access today!
Now, guess what? The last para — I had written about 7 years ago! ;)
Brilliant! There’s no other word to describe the first part of this headline!
As per IC Insights’ forecast of 2010 billion-dollar fabless IC suppliers, excerpted from a ranking of top 50 fabless IC suppliers in its ‘ 2011 edition of The McClean Report’, as many as 13 fabless IC suppliers are tipped to cross the $1-billion mark in sales in 2010! As per IC Insights, this is a significant step up — from 10 companies in 2009 and eight in 2008.
Just sit back and admire this table. There are nine firms from the US — Qualcomm, Broadcom, AMD, Marvell, Nvidia, Xilinx, Altera, LSI and Avago, three from Taiwan — MediaTek, Novatek and MStar, while ST-Ericsson is Europe’s lone representation in this stellar list.
In this august club of IC billionaires, no surprises, but Qualcomm retains the top place for the third consecutive year. Broadcom moves up a place. AMD should become the world’s third largest player.
Broadcom at 53 percent, Marvell at 34 percent, Xilinx at 39 percent, Altera at 63 percent, Avago and Novatek at 40 percent each are top performers. However, MStar of Taiwan steals the show with an estimated 75 percent growth in 2010.
Qualcomm, Nvidia and LSI have performed well, especially the last two – coming pff a difficult 2009. Taiwan’s MediaTek has seen the biggest slip — down to 3 percent in 2010 from 22 percent in 2009.
There is no representation from Japan in the fabless IC billionaires club. IC Insights has indicated that the fabless/foundry hasn’t caught on in Japan and is unlikely to do so in the near future. However, Taiwan and China based firms should sooner or later find their way into this club.
I will now come to India! Read more…
VLSI manufacturing: The government of India has to give appropriate incentives to this industry and suitably phase these to allow an ecosystem to build up. The fabless model is fine for companies to transact their business. But, India needs to have this technology or some semblance of it to start with or else it would never be able to count herself amongst the leading nations of the world. How can we achieve that?
a) One of the quickest ways to get started is to create an organization (with equity wholly or partly owned by the government of India) with the charter that that company invest in/fractionally own some of the leading semiconductor manufacturing facilities in Taiwan. TSMC may be very difficult to own even modest percentages, but simultaneously buying reasonable stakes in UMC, Chartered Semiconductor, the IBM multi-platform consortium may be possible with (my guess) at under $1 billion that gives adequate leverage. This organization then makes available fab capacity (that is available to them now due to their fractional ownership) to fabless product companies with intended product sales in India.
b) Repeat the above for packaging, assembly, and VLSI test operations. (The challenges are less here though, but these are just as important elements of the ecosystem).
c) With this, India will “own” leading edge VLSI technology although in a circuitous manner, but given that we missed the boat by over 20 years I think this is a reasonably low cost but effective way to get a starting foothold.
d) Next, and in parallel, the government ought to give incentives to companies to create semiconductor manufacturing capacities on Indian soil but buying re-furbished and lower cost equipment at previous generation technology nodes.
Some policies are required here that allow easy transport back-and-forth of sub-assemblies for repair and replacement.
e) Eventually, the organization(s) in (d) above will “catch” up with those in (a) above. There will need to be phased modification of the incentives, reducing some, changing some such that over a six-seven year period all incentives can go away and the eco-system will be self sustaining.
f) Well, not quite, we will also need to address (i) the semiconductor manufacturing, assembly, test equipment manufacturing and (ii) the semiconductor manufacturing raw materials (silicon wafers, silicon grade pure chemicals, etc etc). The focus to these technologies have to be phased to the decade after the steps outlined in (a) to (d) have yielded some measurable results.
Friends, please feel free to share your thoughts, comments, etc., and add value.
So, IC Insights has revealed the top 25 list of fabless IC suppliers for 2009! No surprises, Qualcomm still leads!
However, AMD is the surprise runner-up, for now. The reason being: AMD became a fabless company by including its Dresden, Germany fabs as part of GlobalFoundries spin-off. IC Insights included all of AMD’s sales for 2009 in its study.
Some other interesting points
First, as many as nine fabless IC companies — all of the top nine companies — had sales of $1 billion or more in 2009. These are: Qualcomm, AMD, Broadcom, MediaTek, nVidia, Marvell, Xilinx, LSI Corp., Altera and Avago! And you still believe there was a recession in H1-09?
Movers and shakers
So, who are the leading top movers and shakers?
The list comprises fabless IC suppliers from the USA — which has 17 representations, including nine suppliers in the top 10! Taiwan has six representations, with one — MediaTek — figuring among the top 10, well, top five actually! Europe and Japan have one representation each — in CSR and MegaChips. Read more…
Anil Gupta, Technovation 2010 and UK CIG Convener, India Semiconductor Association (ISA), also needs no introduction. As former managing director, ARM Embedded Technologies Pvt. Ltd, he has been a prominent figure in several industry events. Here, he presents his views on what needs to be done for the Indian semiconductor industry.
An interesting fact being brought up time and again within the industry is the requirement of a robust entrepreneurial spirit and the need for much more sources of funding for semiconductor product companies. Also, renewable energy, healthcare and security are some of the verticals where the Indian industry believes there is a lot of value to be added from the Indian market/need perspective.
Further, local products/systems design and development activity needs to be encouraged and kick-started in a big way in India, for the industry to really succeed big time!
What does Indian semicon need?
We have discussed several times in the past regarding what needs to be done with the Indian semicon industry. So, what really needs to be done, given the current slowdown? What can be done boost chip designing activities in India.
According to Anil Gupta: “The Indian story has always been a story of a lot of potential, but most often this potential is never realized.
“The software industry has done well and has gone far, perhaps, somewhat farther than the hardware or chip-design industry in India. However, you still don’t see a software product conceptualized, designed and developed in India that is worth mentioning.
“The Infosys’es and the I-Flex’es can do a phenomenal job of executing software projects for their customers, but they are all a far cry from the league of the top consulting firms that define the problem to be solved and the software solutions to be built.
“The Indian software industry is still plagued with the “revenue per head” model and is unable to grow beyond it. The Indian software companies clearly bring a significant value to their customers but this is NOT strategic value, it is merely an execution value.
“Compared with the software industry, the embedded systems industry in India is puny today. However, the opportunities are phenomenal because there is so much automation potential in so many verticals. However, once again, the lack of significant products/solutions development in India is a very big hindrance.”
From a technical expertise perspective, there is a lot of engineering talent available, but the expertise is in general quite shallow. Even in the open source space like Linux, there aren’t many noteworthy contributions to date from the Indian engineering community.
Is the Indian fab story truly dead and buried?
In the past, we have extensively discussed whether the Indian fab story was dead and buried. Do you see any change in the current situation?
Well, it is dead for now! Gupta added: “Its day will come ONLY when the economics works out in its favour. Today, it doesn’t!”
He said: “It is interesting that many point out to the success of the solar fab investments. However, it should be noted that there is no solar wafer manufacturing activity worth mentioning. Only modules are being assembled in India as there seems to be a global glut in wafer production. Thus, wafer fabs in India is a pipe dream for now since the economics doesn’t work out.”
Does India have entrepreneurs committed to product development and willing to take that risk? How can they be encouraged?
Gupta said that there are not that many who are willing to come out and take the risk, and the lack of funding is a very big handicap. The lack of prior successes that could be emulated is probably the biggest handicap.
In that case, what needs to be done in India to move up a higher level, beyond design and verification?
He said: “Clearly, a willingness to take risk, strong stomach to face failure, and strong will to learn from that failure to rise again from the ashes!”