Archive

Posts Tagged ‘wafer fabs’

Why do we need 450mm wafers?

April 30, 2013 Comments off

Here is a view from Mike Bryant of Future Horizons, taken from the Enable450 newsletter, for which, I must thank Malcolm Penn, chairman and CEO.

This is a question often asked by journalists and others not directly involved in 450mm technology, and indeed was one of the questions that formed the basis of the SMART 2010/062 report Future Horizons produced for the European Commission.

Mike Bryant.

Mike Bryant.

It is also a question every new 450mm project has to answer in its funding request to the European Commission, and whilst working on the Bridge450 submission we realised the arguments have become rather unclear over time. The following gives some insight and clarity into the question.

In 1970, Gordon Moore re-formulated predictions on computer storage by Turing and others into a simple statement that the number of transistors per unit area of an IC will double every two years for at least the next ten years. This became known as “Moore’s Law” and apart from the occasional hiccup has in fact been followed for the past forty years. Note that Moore never suggested a doubling in density every 18 months, this time period coming from a different statement concerning transistor performance.

Of course, doubling the number of transistors would not be that helpful if the price per unit area also doubled. The semiconductor industry has thus strived to maintain the cost of manufacturing per unit area at a constant price, and analysed over time has done a remarkable job in maintaining this number such that the ASP of logic devices has sat at around $9 per square centimetre for this whole period during which the cost of everything else including the equipment, materials and labour used to make the IC have increased, labour costs in particular increasing by a factor of around five times.

The actual cost of processing a wafer appreciates by around 6 percent per annum due to technology cycle upgrades and insertions, for example in the past the replacement of aluminium interconnects with copper or more recently the move to double patterning for lithography of critical layers. Several approaches have been used to maintain a constant area cost, these being:

Improvements in yield – this obviously reduces wastage and vast improvements have been made in this field though yields are now so good that the problem is more maintaining these levels with each new process node rather than improving them further.

Increasing levels of automation – this is still an area undergoing improvement but again we have entered an area of diminishing returns on the investment required.

Introducing larger wafer sizes – this has been performed on an irregular basis over the history of the semiconductor industry. The increase in surface area reduces many but not all of the processing costs whilst material costs tend to stay fairly constant per unit area. Thus at the 300mm transition the increase in area by 2.25 times gave a cost per unit area reduction of 30 percent, approximately compensating for the increased processing costs acquired over the 90nm and 65nm nodes.
Read more…

Is Q1 a cycle bottom? Time for 2009 reality check! Semicon update Apr’09

June 17, 2009 Comments off

Here are the excerpts from the Global Semiconductor Monthly Report, April 2009, provided by Malcolm Penn, chairman, founder and CEO of Future Horizons. There are a lot of charts associated with this report. Those interested to know more about this report should contact Future Horizons.

This will be followed by updates for May and June, and I am speaking with Malcolm Penn to find out more!

Executive overview
“February’s IC sales were up 4.3 percent on January, down 26.5 percent on the same time last year. If March behaves true to historical norms, we can expect to see sales up 26 percent on February, at 14.3 billion (on a calendar month basis), equivalent to plus 0.8 percent, four-week month adjusted.

This would see Q1 sales down around 15 percent on Q4-2008, just slightly ahead of our 18 percent forecast decline. Whilst March’s data point is right now still an estimate, the year to date data and trends give strong guidance on what is actually happening. Time therefore to reflect on our 2009 chip market forecast and growth pattern outlook for the year as a whole.

At our January 2009 International Forecast Seminar in London, we forecast growth for the 2009 market at -28 percent. This was the most pessimistic of all the industry watchers!Source: Future Horizons

The forecast reflected the unprecedented Q4-2008 industry meltdown that started on Sept 16 with the Lehman Brothers collapse. December was an especially a traumatic month, with several firms reporting negative net monthly sales (i.e., cancellations were higher than new orders), with zero guidance visibility on the outlook for Q1-2009.

Our forecast estimated Q4 would decline a 22.5 percent versus Q3, followed by a similar (but slightly slowing) decline of 20 percent in Q1, bottoming out in Q2 (at -2 percent versus Q1) followed by reasonably strong seasonal recover in Q3 and Q4 of plus 12 percent and plus 3 percent respectively.

Once December’s results were published in February, we modified this profile slightly to reflect December’s actual 24.2 percent decline (versus our 22.5 percent estimate), reducing the first quarter decline slightly (from -20 percent to -18.5 percent) thereby maintaining the overall year-on-year 28 percent decline.

Ironically, despite having the most pessimistic overall year-on-year forecast, we were widely criticised at the time for predicting a ‘V-shaped’ recession. Yet, to achieve the more favoured ‘U-shaped’ recovery would have meant a very low single digit quarterly decline in Q1, something we did not believe was realistic or likely. Our most optimistic (rose coloured glasses) scenario pegged Q1 growth at -8 percent, yielding an 18.7 percent annual 2009 decline.

Despite December’s worse than forecast results, February’s data, both in its absolute value and underlying momentum, added credence to our ‘V-shaped’ scenario, despite the emergence of a new popularist theory of a W-shaped recovery.

As such, we are sticking to both the shape of the recovery — V, not U or W — and profile; there is even some indication that the recovery is happening slightly earlier than we estimated. Now that, if true, would soften the depth of the 2009 decline.

Based on January and February’s WSTS data, March now looks like coming in at $17.019 billion, which would see Q1 reach $43.642 billion, down 16.4 percent on Q4-2009, which is 2 percentage points lower than the 18.4 percent we were forecasting. If this is the case, Q1 will mark the recession cyclical bottom.

Industry capacity
The Q4-08 total MOS IC capacity was up just 3.1 percent versus Q4-07, which in turn was up 13.6 percent on Q4-2006. Quarter on quarter growth was -1.7 percent, compared with +1.1 percent for Q3-08, plus 2 for Q2-08 and +1.7 percent for Q1-08. This dramatic slowdown in net new capacity is in direct response to the slowdown in capex that has been gaining momentum since the second half of 2007.

It should be remembered that there is a ‘nine-month delay’ between a capex spend and saleable units out, so capex in year ‘n’ drives capacity expansion in year ‘n+1’. As such, the capex spend is now growing much slower than the underlying unit demand, and the impact is an eventual increase in the capacity utilisation rates.

Do not be misled by the sharp falloff in Q4-08 utilisation; this was the direct result of the September financial crisis near-term inventory purge driven demand slump and not representative of the underlying trends. We expect this to bounce back quite rapidly once orders readjust during the first half of 2009.

Although due to timing, inventory and seasonality issues, supply and demand will never identically track, utilisation rates have been straddling the 90 percent level since mid-2003. We expect this trend to resume by the end of this year.

With capex spend averaging around US$8 billion per quarter between Q3-06 and Q1-08 spending plunged dramatically in Q2-08 reaching under half this average in Q4-08. Given the current front-end capex book-to-bill trends, this spend will shrink still further at least through 1H-2009. Interestingly, this cutback happened well before the Q4 market meltdown, the impact of a premeditated strategy to dramatically tighten supply and thereby increase wafer and IC average selling prices.

The level of new front-end capital equipment orders has now been sizeably lower than sales for 31 consecutive months, the last four at unprecedentedly low levels, aside from two short-lived incursions into positive territory circa Q4-06 and Q4-08. The 2008 capex spend was down 30.6 percent on 2007’s level, with the outlook for 2009 looking to be at least a further 30 percent lower. That would put 2009’s capex spend at well under half 2000’s peak.

No amount of productivity gains can offset this slowed investment, especially now the one-off 300mm conversion gain has been absorbed. Net new capacity addition is thus condemned to shrink even further during 2009, the effect of which will be masked in the near-term by the current inventory /demand adjustment process.

With near-term demand having shrunk in Q4, this strategy has essentially been blocked but not reversed. Unlike 2001, when recession hit during a period of capex expansion, the bounce back from the current dip will be quite sharp and sudden. It is only a matter of time before capacity gets squeezed and wafer process rise. We expect this trend to bit hard in 2010, possibly even leading to shortages and allocations, just as the economic recovery starts to gain momentum.

The interim period of ‘plentiful capacity in 2009’, will feed the perceived wisdom of a sense of supply security, those with an inkling of a medium-term plan need to tie down their supply positions whilst the going is good. Today’s era of cheap and plentiful wafers, like the discredited ‘debt is cheap and free’ era, are number and counting down.

Just to make the point deeper we tracked the book-to-bill ratio against future capacity adjusted for the three quarter lead-time delay. It shows both a good trend correlation and the depth of the problem. Once again, it reinforces all of the other anecdotal and hard evidence that net new capacity growth is condemned to slow even further. When the unit demand recovers, the capacity simply will not be there, especially at the leading edge technologies.”

Definite need for rethink on India’s fab strategy!


I am intrigued to see lots of great things happening in the Indian semiconductor industry, and equally frustrated to find certain things that I feel should happen, not really going the way they should!

Yes, India is very strong in the semiconductor related chip design services. However, do keep in mind folks that design services have been impacted a bit by the recession as well! There have been calls from several quarters for India to now start thinking beyond its chip design services. Therefore, are there any areas that India can look into within the semiconductor space?

Leverage strength in software
Certainly, value add in products are heavily influenced by the embedded software in addition to features of the chips, says S. Janakiraman, former chairman, India Semiconductor Association (ISA) and President and CEO-R&D Services, MindTree. “The Google Android is a great example of that. India should leverage its strength in software to enhance its value add to semiconductor companies,” he adds.

Innovation is now shifting from the development of new technologies to the creation of unique applications.

“Mobile browsers and management of remote appliances to save power at home/office are examples. We need to innovate new applications that can drive the need for more electronic gadgets, and in turn, the need for more semiconductors,” notes Jani Sir.

Rethink on Indian fab strategy?
One of my earlier posts focused on whether an Indian investor could buy Qimonda’s memory fab, and somehow kick-start the India fab story! I did find support from many quarters on this idea, but till date, I don’t think anyone from India has made a move for Qimonda. At least, I haven’t heard of any such move.

Nevertheless, some folks within the Indian semiconductor industry and elsewhere have called for India to rethink on its fab strategy.

What should it be now? Or, shall we just discard this and go on, as India has been doing fine without fabs so far? Perhaps, the last option is easier!

According to Janakiraman: “Perhaps, we should consider where semiconductor technology will be after five years from now, and prepare grounds for that through encouragement of fundamental research, as well as shuttle fabs to enable prototyping. We should skip the current node of technology and make an entry into the one that will be prevalent after few years.” Now, that’s sound advice! Will it be easy to achieve?

“That may not be as easy to achieve for the private enterprises considering the cost involved,” adds Janakiraman. “It has to be a mission of the nation to create that infrastructure and later privatize.”

According to him, it is not unique to India. “Every country, be it Taiwan or China, have done it. The only other way is to heavily subsidize and support fabs like those in Israel or Vietnam, but it will be tough to choose a partner in a democratic country like ours, wherein every investment and subsidy is seen with a colored vision,” he says.

To sum up, a fab for our country will be fundamental to gain leadership and self reliance. It cannot be ignored totally, although we can take our own time to reach there. Janakiraman adds, “We don’t have a choice other than paying a price to reach there, now or later!”

Indian silicon wafer fab story seems dead and buried! Should we revive it?

February 12, 2009 1 comment

Now then, this will make a very interesting read! Back in October 2007, I had discussed the timing and the need for a silicon wafer fab in India, in-depth, with Anil Gupta, managing director, India Operations, ARM.

We have come a long way since then! There was all the hype last year about SemIndia’s fab, which never really did happen, and eventually, BV Naidu moved on! Then came the rush to solar fabs. Recently, when I blogged on how a Qimonda buy could be good for India, I am told that it is really outrageous. No problem, it is merely a suggestion.

At times, I have got the feeling whether the Indian semiconductor industry is losing its way! However, when I see all around, it is hale and hearty, and business as usual — fabs or no fabs!

It was interesting to meet up again with Anil Gupta of ARM, and to find out what he thought about what I thought!

Starting with an old question, whether India has the capability to sustain or even build a product development ecosystem? Gupta said: “We need the following for this:
* Entrepreneurs committed to product development and willing to take that risk.
* Investors willing to take risk on product development companies.
* Consumption (this will happen as the economy improves any way).
* Deep enough technical/technological knowledge/know-how to put reasonably competent end products together (It exists. Examples like Sukam, Tejas and other are there).

Indian fab story dead and buried
Turning focus on fabs, is the Indian silicon wafer fab story completely dead and buried now? Gupta notes: “When TSMC says they are running at only 38 percent capacity, one can imagine what the rest of the fabs must be going through. In any case, the Indian fab story was a longer term story and the current economic climate actually makes it further and further remote. So yes, it is dead and buried now!”

Wow! India probably flattered to deceive! However, I am an optimist, and hope that one day, India will have its own silicon wafer fabs!

Gupta adds: “What worries me now is the glut of the solar/PV fabs. By the industry estimates, solar/PV is a viable option only when the price of oil is >$100 per barrel (oil is at $40 per barrel now). This means, there would be challenges for the solar cell industry too! One can only hope that the economy picks up growth soon enough and sends the price of oil higher so that solar becomes a viable option.”

Again, this is a concern I have as well. The rush toward solar is good, but then, is this what the Indian semiconductor industry really needs? Where’s all that talk of developing silicon and product companies? You simply cannot equate the two — semicon and solar! You can’t have a policy, and then ignore the main crux either, and simply go for the ones that are easily attainable! It does not project a good impression, or maybe, I am somehow wrong in my assessment. Hence, my feeling that the industry could be losing its way somewhere!

However, Gupta feels that’s not really the case! What has been working until now, still continues to work!! “Our strengths are design and verification. We will continue to be in demand for that. The other pastures we explore, there are a lot of uncertainties,” he adds.

“The challenge is to pick the right pasture where the grass remains green even in the summer. This is not easy to find and does require that we bet on some of them and learn through the experience,” he advises.

How can India really buzz?
What now needs to be done to get the semiconductor industry in India really buzzing? Surely, local consumption is key. Local consumption would hopefully foster electronic product innovation just like products by two-wheeler manufacturers and the Tata Nano.

“The current initiatives in the industry for rural applications are also quite interesting. I am optimistic that some good offerings will come out of this. While these may not be specifically from a “semiconductor” perspective, at least at the “system” level these would make sense,” says Gupta.

What India NOW offers to semicon world?
What does India NOW offer to the semicon world, in these times of a global recession?

The Indian economy is still mostly internal consumption oriented, as opposed to exports oriented. This is very different from the economies of island nations like Taiwan, Korea, and Japan, which are very heavily export oriented.

In a recession like the current one, these predominantly export-oriented economies experience a far greater crunch than the others. Thus, as long as products are being sold in Indian markets at the right price points, there would be consumption.

Gupta says, “This time around, the world would come out of recession mainly driven by Asian countries, India being one. People in the industry that I talk to tell me that as the worst is over in this crisis, and as things begin to pick up, India will once again be the beneficiary of a lot of work moving here. However, my personal view is somewhat different.

“I believe that the last round did witness this phenomenon mainly because it was the honeymoon period. But by now, the honeymoon period is over and the India centres of these companies are working hard to reach a level where they become “mission critical” to the businesses of their companies.

“The journey hasn’t been very easy for multiple reasons. And by now, the cost differentials also do not look as attractive as they did before. Hence, what work comes here would come only after a careful assessment and very selectively (not by leap of faith).”

I did blog about how Qimonda could be a good buy for starting a memory fab in India. You have all the facts in front of you! My question to the Indian semiconductor industry is: should we revive the call for having a silicon wafer fab in India, post SemIndia and post recession?

India fab story not disappearing: SemIndia

June 24, 2008 Comments off

According to B.V. Naidu, managing director, SemIndia Systems, and Vice Chairman – India Semiconductor Association (ISA), the Indian fab story is well on track. Here, he speaks about the FabCity, the status of fabs in India and SemIndia’s initiatives.

On the status of setting up of IC wafer fabs in India, he says that a couple of companies are in the process of raising the money for setting up the wafer fab in India. We still do not know the timelines for the same.

On India’s fab story is disappearing, Naidu feels that it is always difficult to raise the money for such large capital-intensive projects, which are happening first time in the country. “One of the large industrial groups also announced their plans to set-up the fab. This shows that the fab story is not disappearing,” he says.

The Andhra state government had recently sent notices to SemIndia and NanoTech over FabCity. Pertaining to the status of the project, he adds that any capital-intensive new projects of this kind will take some time. It is also quite natural that the government generally issues such notices more to put pressure so as to expedite the project implementation. It is good that the governments are closely monitoring the progress of the project implementation.

According to him, SemIndia’s ATMP project construction has started. The FabCity has already come-up and many solar PV fabs are being set-up in the FabCity. This shows that efforts of SemIndia, ISA and the government of AP have yielded the successful results to make FabCity a successful project.

The India Semiconductor Association (ISA) can only indicate that such government communications are common for wherever there is government support. The ISA will continuously put in their efforts for attracting the new investments to India and work along with the governments to make sure that their efforts are fruitful.

“SemIndia’s ATMP project construction has started and we are still looking for the investors for their fab project,” says Naidu.

So, what sort of planning is now required from the Indian semicon industry? As per the SemIndia managing director, the Indian semicon industry should continuously work with the government to make sure that the government of India’s semicon policy is successful and efforts are various state governments in attracting the new investments in this area are fruitful.

The ISA will continuously strive for creating the balanced eco-system for the semicon design industry, high-tech manufacturing and talent nurturing.

Has the Indian silicon wafer fab story gone astray?

June 23, 2008 Comments off

The recent news of the Hyderabad Fab City giants — SemIndia Fab Pvt Ltd and Nano-Tech Silicon India Pvt Ltd — being served notice by the local state government and to explain the reasons for their delay in setting up the Fab City in Shamshabad on the outskirts of Hyderabad, does not come as a surprise at all!

Setting up of a silicon wafer fab takes up a lot of time and money, and I am not sure how this bit is perceived by many. Also, the rate of return is not exactly immediate! Maybe, it is time for everyone to realize that semiconductor is a very different industry from any other, and there is a need to understand how it really functions! Besides, one needs to keep an eye on the global semiconductor industry and associate movements there with what kind of value would a fab in India bring to the world.

This May, I’d done a reality check on where the global semiconductor is placed. Several folks have contacted me since, pointing out my accuracy. While it is good to be spot on with the assessment of the global (and Indian) semiconductor industry, the assessments should serve as a warning for the global (and Indian) semiconductor industry — that it is not going to be an easy ride ahead!

On the same note, I had earlier questioned whether this was the right timing for setting up fabs in India. Perhaps, there is a need to examine whether we started on the fab path a bit too late! If we are found to be wrong or hasty in our assessment, let us feel no anguish in accepting that! This is not the first time such a thing will happen in the semiconductor industry, nor will it be the last. Having said that, if a wafer fab or two do start functioning in India later in 2009 or beyond, that would be exemplary!

Let us hope that the Indian silicon wafer fab story does not go astray for the overall benefit of the Indian semiconductor industry. There is a need on part of the Indian semicon planners to integrate clear vision with careful planning.

Yes, several solar fabs are coming up globally, and investments in solar/PV are rising as well in India, but that was along expected lines.

It was also pointed out earlier that investments in photovoltaics (PV) had somewhat eased the pressure on capital equipment makers and spend. In fact, 2007 is now well documented as the year when the PV industry emerged as a key opportunity for the subsystems suppliers and provided a timely boost in sales for those actively addressing this market.

Perhaps, here lies an opportunity for India, and I’m repeating this to the extent of sounding boring!

Further, even though it has been quite a while since the Indian semicon policy was announced, some feel that India should continue to focus on design services and embedded — its well known strengths, rather than go after something as mature as wafer fabs. We don’t have to ‘force ourselves to believe’ that we are good at product development? We are not!

Yes, like most things, it can change, but that would need great effort on part of all industry stakeholders. The question is: are we ready to bring about that change?

%d bloggers like this: