EL SEGUNDO, USA: Concerned about their image as they face the specter of bankruptcy, many memory chip suppliers are attempting to paint a more optimistic picture of the business by talking up a potential market recovery.
However, while overall memory chip prices are expected to stabilize during the remaining quarters of 2009, iSuppli Corp. believes a true recovery in demand and profitability is not imminent.
After a 14.3 percent sequential decline in global revenue in the first quarter DRAM and NAND flash, the market for these products will grow throughout the rest of the year. Combined DRAM and NAND revenue will rise by 3.6 percent in the second quarter, and surge by 21.9 percent and 17.5 percent in the third and fourth quarters respectively.
“While this growth may spur some optimism among memory suppliers, the oversupply situation will continue to be acute,” said Nam Hyung Kim, director and chief analyst for memory ICs and storage at iSuppli.
“For example shipments of DRAM in the equivalent of the 1Gbit density will exceeded demand by an average of 14 percent during the first three quarters of 2009. This will prevent a strong price recovery, which will be required to achieve profitability for most memory suppliers.”
Due to a long-lasting glut of DRAM, the imbalance between supply and demand is too great for this market to recover to profitability any time soon.
“Even if all of the Taiwanese DRAM suppliers idled all their fabs, which equates to 25 percent of global DRAM megabit production, the market would remain in a state of oversupply,” Kim said. “This illustrates that the current oversupply is much more severe than many suppliers believe—or hope.”
Besides cutting capacity, which suppliers have already been doing, they presently have few options other than waiting for a fundamental demand recovery. iSuppli believes that another round of production cuts will take place in the second quarter, which will positively impact suppliers’ balance sheets late this year or early in 2010 at the earliest.
DRAM prices now amount to only one-third-level of Taiwanese suppliers’ cash costs. Unless prices increase by more than 200 percent, cash losses will persist for these Taiwanese suppliers.
While average megabit pricing for DRAM will rise during every quarter of 2009, it will not be even remotely enough to allow suppliers to generate profits in this industry. The industry needs a dramatic price recovery of a few hundred percentage points to make any kind of impact.
iSuppli is maintaining its “negative” rating of near-term market conditions for DRAM suppliers.
Confusing picture in NAND
The picture is a little more complicated in the NAND flash memory market.
Pricing for NAND since January has been better than iSuppli had expected. However, iSuppli believes this doesn’t signal a real market recovery.
Most NAND flash makers are continuing to lose money. The leading supplier, Samsung Electronics Co. Ltd., seems to be enjoying the current NAND price rally as prices have almost reached the company’s break-even costs. However, all the other NAND suppliers still are losing money.
“While the NAND market in the past has been able to achieve strong growth and solid pricing solely based on orders from Apple Computer Inc. for its popular iPod and iPhone products, this situation is not likely to recur in the future,” Kim said. “Even if Apple’s order surge, and it books most of Samsung’s capacity, it would require a commensurate increase in demand to other suppliers to generate a fundamental recovery in demand.”
However, iSuppli has not detected any substantial increase in orders from Apple to other suppliers. Furthermore, Apple’s orders, according to press reports, are not sufficient to positively impact the market as a whole.
It doesn’t make sense for major NAND suppliers Toshiba Corp. and Hynix Semiconductor Inc. to further decrease their production if there is a real fundamental market recovery. This means supply will continue to exceed demand and pricing will not rise enough to allow the NAND market as a whole to achieve profitability.
The NAND flash market is in a better situation than DRAM at least. However, the market remains challenging because fundamental demand conditions in the consumer electronics market have not improved due to the global recession.
One of the reasons why the price rally occurred is that inventory levels have been reduced in the channel and re-stocking activity has been progressing. Overall, memory suppliers will begin to announce their earnings shortly and iSuppli will remain cautious about the NAND flash market until we detect solid evidence, not just speculation, of a recovery.
iSuppli is remaining cautious about the near term rating of NAND market, holding its negative view for now, before considering upgrading it to neutral.
“Production cuts undoubtedly will have a positive impact on the market in the future. However, it’s too early for to celebrate. iSuppli believes the surge in optimism is premature. Supplier must be rational and watch the current market conditions carefully to avoid jumping to conclusions too quickly,” Kim concluded.
The 2008 DRAM chip price dropped more than 85 percent, while the global DRAM industry has faced more than two years of cyclical downturn, and the consumer demand suddenly froze because of the global financial crisis in 2H08.
In 1Q09, the DDR2 667 MHz 1Gb chip price rebounded to an average of US$ 0.88, which fell between the material cost and cash cost level. Still, the DRAM vendors encountered huge cash outflow pressure. Not only were capacity cut conducted, the process migration schedules were also delayed in the wake of respective sharp CAPEX cuts.
According to the survey of DRAMeXchange, the worldwide DRAM CAPEX of 2009 has been revised down to US$ 5.4 billion, sharply down by 56 percent, in contrast to the US$ 12.2 billion in 2008.
WW DRAM 50nm process migration schedules all deferred one to two quarters
From the roadmaps of DRAM vendors, the adoption schedule of DRAM mass production using the 50 nm process have now been delayed one to two quarters. DRAMeXchange estimates that by the end of 2009, the DDR3 will account for 30 percent of the standard DRAM.
Regarding the new DDR2 and DDR3 process migration, all DRAM vendors still own different types of strategies of density and types. For example, the Korean vendors’ 50 nm process migration schedules of DDR 3 are earlier than DDR2 and the 2 Gb DDR3 mass production schedule is earlier than the 1Gb chip.
As for the US and Japanese vendors, according to their DDR3 roadmap, the 50 nm process will be introduced between 3Q09 and 4Q09, which is later than the Korean vendors, and also firstly with mass production of 2 Gb DDR3. Therefore, in the DDR3 era, the density will mainly be 2 Gb which is a lower cost driver with more stimulating incentive to the market demand of higher density chips. The Taiwanese vendors are under the high cash pressure and are falling behind in the 50 nm process race. They are mainly focused on “pilot production”.
Gross die increases 40-50 percent as 50nm process drives down cost
According to the Moore’s Law, the number of transistors on an integrated circuit doubles every 12 months. After the process shrinking became more difficult in the recent decade, it increased to 24 months. With new process migration, the closer the line distance is the larger gross die number a single wafer gets, meanwhile the cost is lower and the vendors gain more competitiveness.
The average DRAM output increased about 30 percent during the process migration from 70nm to 60nm. With improvements of process design and die shrink in the same generation of process technology, the output can once again increase 20 percent. In the 50nm generation, the output will increase almost 40-50 percent, compared to 60nm process and the number of gross die increases to 1500-1700 per 12 inch wafer with another 30 percent cost down.
Cost of immersion lithography tools major capex of 50nm process migration
The major challenge of 50nm process migration is the lithography technology. The newest immersion lithography equipment is required and the older exposure equipment at the wavelength of 193nm is no longer suitable under 65nm process, due to physical limitations.
Traditional dry lithography uses air as the medium to image through masks. However, immersion lithography uses water as the medium. Immersion lithography puts water between the light source and wafer. The wavelength of light shrinks through water so it is able to project more precise and smaller images on the wafer. This is the invention that enabled the semiconductor process technology to migrate from 65nm to 45nm.
The current major immersion equipment vendors are ASML, Nikon, and Canon. The largest vendor in the market is AMSL, which is now mainly promoting its XT1900Gi, a tool that is capable to go lower than 40nm and is the most accepted model in the industry. Nikon still promotes its NSR-S610C, which was launched in 2007 and is able to go down to 45nm process. Canon launched its FPA-7000AS7 in mid 2008 that supports the process under 45nm.
This is a continuation of the previous post based on the recent India visit of Hanns Windele, VP Europe and India, Mentor Graphics, where he met key industry figures in a session organized by the India Semiconductor Association. Windele is standing sixth from left, and Poornima Shenoy, president, ISA is standing fifth from right.
Multimode, multicorner tools
Windele mentioned that in every likelihood, another new routing tool would be coming in once the industry enters the 45nm/32nm space. “There is an increasing static timing analysis signoff complexity. The explosive growth in complexity requires multimode and multicorner tools,” he said.
Multicorner and multimode (MCMM) and manufacturing variability will drive the next generation place and route technology. Even in the low-growth markets, technical discontinuities create opportunities for market share changes. For instance, 65nm brings along more than 21 corners/modes scenarios; while 90nm has 10 corners, and 130nm only has four corners.
Therefore, another place and route tool will cover the upcoming MCMM problem. Even in low-growth markets, technical discontinuities create opportunities for market share changes.
Companies cannot afford the growing cost of EDA. Even the cost of design is growing exponentially, especially, verification, as well as embedded software development costs. Even the EDA revenue has been a flat 2 percent of the IC revenue. However, productivity has been growing as the number of engineers don’t seem to be multiplying in a great way. For example, the transistors produced per electronic engineer has been hearly four-orders of magnitude since 1985.
Showing optimism in recession
Turning to the ongoing recession, which has impacted the semiconductor industry, Windele said that 2009 will be most likely turn out to be the worst recession in the history of the global semiconductor industry.
“It seems to be heading that way. There is also a lot of reason for optimism. I feel that 2009 will be a lot milder than 1985 and 2001,” he said. Even the electronics indsutry’s growth rates have been slowing, decade by decade as well.
Therefore, with this ongoing global recession, why should we remain optimistic? Simple! A crisis translates into opportunities!!
Betting on India
No prizes for guessing where the most opportunities lie — India! Significantly, the ‘middle class’ in urban India becoming a majority. There is likely to be $3 trillion of discretionary spending by 2010. “People who can afford electronic and consumer goods will be growing further,” he added.
Windele cited ISA’s figures, which says that India’s electronics consumption is headed toward $300 billion by 2015. India’s electronic equipment consumption will likely grow at a CAGR of 30 percent through 2015. It was around $28 billion in 2005, and is likely to increase to $127 billion by 2010, and to $363 billion by 2015.
Yet another reason is the growing number of new cell phone subscribers in China and India, which will be 2x larger than the total US subscribers until 2011. Asia is, by far, the most attractive market for new cell phone sales. India will grow fastest, he added.
Comparing the downturns of the recent years, Windele noted that 2008 and 2009 look different than the other downturns. “There is hardly any inventory left in the industry. One prediction is: as the price upswing comes, prices in the semicon industry will go up very quickly,” he noted.
Seeds already being sown for recovery in 2010. Already, the industry has experiecned two years of severe price declines in memory. Further, systems will be re-designed to take advantage of lower bit prices of FLASH and DRAM.
There will be consolidation and reduced investment in semiconductor capacity in 2008 and 2009. Ramp-up of new system designs will likely happen in 2010 during the period of reduced semiconductor supply.
Concluding, he added that Mentor Graphics became the number 1 EDA company in Europe as the company managed the crisis better than some of our competitors.
Connecting with new friends from all over the world is one of the best things that I have experienced while writing a semicon and electronics blog. One such gentleman is Ingo Guertler from Europartners Consultants. He is based in Munich, Germany — a city I have frequented several times.
Guertler has been part of my LinkedIn network as well. He has spent 30 years in leading positions in the electronic components market, mostly with semiconductors, at General Instrument, QT Optoelectronics and Vishay. Since 2005, he has been the senior partner at Europartners Consultants, a network of independent consultants, mostly located in Europe.
Beside individual projects, Europartners analyze the worldwide distribution market for electronic components each year and publish the results in its Worldwide Distribution Report.
Additionally, the company also organizes a two-day conference in Paris every two years, with high-level speakers out of the electronic component industry, discussing actual topics with top managers from component manufacturer and distributors.
This year, the headline will cover the world economy crisis, its effect on the electronics industry, and how companies can successful manage the crisis.
Naturally, our conversation revolved around the global semiconductor market, the memory market turmoil, and how European companies view the Indian semiconductor industry.
Semicon to decline 20 percent in 2009
Ingo Guertler expects a decrease in the global semiconductors market of approximately a minimum of 20 percent. Europartners does not see a recovery before the middle of 2010.
Guertler said: “Maybe, there can be a light recovery in the second quarter 2010 in some regions. Everything will depend on how fast the funds of the governments to the industry will draw. We have to specifically watch, the American and Chinese markets.”
Regarding the memory market, he added that everybody had expected that Vista will stimulate the markets. However, that did not happen. “For the time being, the industry has no direct killer application for memories available or in the design stage. I expect a further price erosion on memories, especially on DRAMs,” he cautioned.
Europe’s interest in Indian semicon
Again, it was natural for me to query Ingo on how European companies view the Indian semiconductor industry.
According to him, for the time being, collaboration between the European and Indian companies is limited in the most cases, or on a one-way service base, using the excellent skills and resources of the Indian software companies and engineers that also includes EDA. “I don’t think that will change in the near future,” he added.
Guertler said: “Personally, I see India as a new market challenge in the next 10 years or more for the European companies, because the local demand will grow faster in India, than we today see in China. Also, I believe and have seen it already, that a lot of companies will likely shift their production bases from China to India the next time, simply because of lower costs, availability of good, graduate engineers and a more Western orientated politics of the Indian government.”
However there is one handicap for the Indian continent! That is: the current infrastructure and the political situation between India and Pakistan.
According to Guertler, India has to set up huge investment programs to invite more investors in the country. Very importantly, is there any feeling that overseas companies’ interest in India is slowing down?
“Definitely not, as far as the European companies are concerned. If India meets the industry’s requirements, I believe their preference will be for India in comparison to China,” he said.
That indeed, is great to hear! The Indian interest is very much intact, in Europe and elsewhere. Now, it is time for India to get some work started on semiconductor product development companies.
I had ended one of my previous blog posts by saying whether the Indian semiconductor industry was hitting the right notes?
In a continuation to that specific thought, it is necessary to first examine where India stands in the global industry. We are very strong in embedded design and design services — our traditional strengths. While these will hold good for a long time, these are probably not enough to really help India make a serious mark at the global level.
The Indian semiconductor industry, in its current state, needs a rethinking as far as strategy is concerned. Maybe, it cannot survive on chip design alone. Especially in times of downturn, the global semiconductor industry players would be looking for new markets and even customers, rather than low-cost production centers.
Consider these points: In the current economic environment, is the interest in developing new business relations with India really a top priority for overseas companies? Probably not, at this very point of time!
India is also seen more as a source of resource; and the extra resource is the last thing firms need at the moment, given the recessionary climate. What global firms are looking for are new markets and customers, and these points, along with its infrastructure, have been the areas of Indian weaknesses. Maybe, all of this will change, but definitely not overnight! And it needs some more planning.
That leads me to an interesting comment from a reader of my article on CIOL, who went on to suggest that an Indian investor could consider buying Qimonda!
Now that is some serious thought and vision as far as mid- or long-term planning is concerned. However, will there really be any takers for this? If this really happens, fabs can be built in India for memory production. If these fabs perform well, it just might turn out to be a good investment in the mid-term future of the Indian semiconductor industry. Definitely, it will make the world sit up and take notice. The other players would surely give India a look-in thereafter.
Quite a thought! This suggestion of investing in Qimonda is indeed a vision. Can the Indian semiconductor industry develop the courage to show and work toward making this kind of a vision a reality?
I asked him: Does India have the capability to sustain or even build a product development ecosystem? What needs to be done?
He 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, and this will happen as the economy improves any way, and
* Deep enough technical/technological knowledge/know-how to put reasonably competent end products together.”
According to him, all of these qualities exist in India, and he cited examples of companies such as Sukam, Tejas, etc.
Well, there you have it!
We need enterprising entrepreneurs in India who are committed toward product development and willing to take that risk, especially in semiconductors. We need investors who can believe in things like even buying Qimonda, or some other company. After all, isn’t this what everyone’s been saying: this is the time to buy!
Dream big, India!
DRAMeXchange has recently released its rankings for the top NAND suppliers of the world. I am producing bits of that report here, for the benefit of those interested in NAND and the memory market.
Be aware, that this segment has been hit particularly bad. We have heard of Qimonda’s problems, as well as Spansion’s. They are trying to battle it out, gamefully, and best wishes to them.
The global semiconductor industry needs the flash memory segment to recover, and fast, to bring the health back in the industry, as well as the missing buzz!
Getting back to DRAMeXchange’s report, NAND Flash brand companies released their total revenue of 2008. Samsung’s annual revenue was $4.614 billion and it gained 40.4 percent market share, to maintain the number 1. position.
Hynix’s annual revenue was $1.727 billion, with 15.1 percent market share. Though it stayed at the number 3 position, its market share declined 4.1 percent, compared to 2007.
Micron’s annual revenue was $897 million. It had a 7.9 percent market share, which enjoyed a 1.8 percent increase when compared to 2007. Micron was number 4. Intel was at number 5. Its annual revenue was $660 million with 5.8 percent market share, which increased 2.1 percent, compared to 2007.
Numonyx’s (STMicro) 2008 annual revenue was $295 million. It was at number 6 position with the market share of 2.6 percent, which remained the same as 2007.
According to DRAMeXchange, the 4Q08 total revenue of worldwide NAND Flash brand companies was $2.227 billion, which dropped 19.3 percent from $2.761 billion in 3Q08. Under the continuing impact of global recession and the influence of declining worldwide consumer confidence, the 4Q08 revenue of NAND Flash brand companies showed signs of decreasing.
The overall demand and expenditure for consumer electronics declined. Although bit growth in 4Q08 increased 18 percent QoQ, the overall average selling price (ASP) dropped 32 percent QoQ, says DRAMeXchange. A big thanks to DRAMeXchange.
“Let’s start from the very beginning! A very good place to start!!”
Hope you all remember this lovely song sung by Julie Andrews in The Sound of Music!! So, what’s the connection?
Right! Last week, I blogged about how the global semiconductor industry is likely to drop by 28 percent in 2009, while the Indian industry should grow by 13.4 percent during the same period, and that, we should not get carried away by these statistics!
A moment to ponder: isn’t this drop of 28 percent too high for the global semicon industry? Or, is the situation really that bad? So, let’s start from the very beginning, and go straight to the source — Malcolm Penn!
Revival likely by 2010?
Here’s what Malcolm Penn, CEO and founder of Future Horizons, had to say: “Fraid not! It could even be lower, but remember that this is a year on year number. It is based on the following assumptions: Q4-08 down 22.5 percent vs. Q3-08; Q1-09 down 20 percent vs Q4-08; Q2 down 2 percent vs Q1; and Q3 up 12 percent vs Q2, and Q4 up 3 percent vs Q3! And, if this pattern runs true, 2010 will be up 28 percent vs 2009!”
Voila! The global semiconductor industry could well be in for a major revival next year itself! Why, even Bill McClean, president of IC Insights, took a more optimistic look at the state of the industry in light of the current global economic situation at the recently concluded SEMI ISS 2009 conference!!
Continues Penn, “The actual Q4 results (released this Sunday) were down 24.2 percent, slightly worse than our estimate.”
How to get the buzz back in semicon?
It has been said that the current situation the global semiconductor industry finds itself in was fueled by greed and short-term business goals. So, who were the culprits? Weren’t they warned earlier?
Adds Penn: “It was more complex that that! The woeful state-of-the-world economy was a consequence of debt, greed and irresponsibility; political self interests and short-term business goals, aided and abetted by compliant governments; ineffective regulators; imprudent institutions; incompetent management; irrational self delusion and vested self-interests! No one is blameless for this crisis! Concerns were raised, but the human nature is often irrational, and the ‘easy option’ always the one of choice.”
So true! Perhaps, the ‘easy option’ factor seems to be affecting the Indian semiconductor industry as well, but more of that later!
The key issue today is: what needs to be done to get the buzz back in the global semiconductor industry? The answer probably lies in the following: in the short-term, it involves rebuilding the industry confidence, and in longer term, it involves a radical return to ‘old fashioned’ business and political values.
On another note, I was curious to know how the EDA segment is doing? Penn said, “No better, no worse than normal, technology marches on, new designs accelerate in a downturn.”
Memory is another segment that’s been hit hard. In fact, the other day, someone asked me why Qimonda’s story was so important!
Another could not understand what Spansion really did, and why it had announced this January 15 that the company was exploring strategic alternatives for a sale or a merger! Doesn’t matter! Memory is a very tricky business, and semiconductors is the mother of all such tricky businesses! Perhaps, isn’t that why they once said in jest: “Real men have fabs!” Anyhow!
Coming back to memory, when can the industry expect some recovery in NAND? More importantly, will the various government interventions help? Qimonda also recently petitioned for the opening of the insolvency proceedings.
Penn is clear: “NAND will recover when the excess capacity abates, and that will take several more quarters. The government intervention won’t help, rather the opposite, and it will exacerbate the excess capacity issue.”
Fab spends to move up only by Q1-2010
Earlier, Penn predicted a recovery in 2010 with the resumption of growth in Q3 2009. What will make this happen? He says, “A recovering world GDP growth, plus a return in business confidence.”
However, those keen on fabs, do not expect the fab spends to look up any time soon! In fact, Penn estimates fab spends to start moving north not until Q1-2010 at the earliest.
The Chinese impact!
Interestingly, China is set to see negative growth of 5.8 percent during 2009. It will be worth noting how much of this this impact the global semiconductor industry.
Point one, compared to a global semicon fall of 28 percent in 2009, Penn considers a fall in China’s semicon fortunes of 5.8 percent to be ‘darned sight better!’ So, China should still be a high growth market (relatively speaking).
Like I mentioned earlier, the Indian semiconductor industry is perhaps getting affected by the ‘easy option.’ Design services continue to do well, hopefully, but when it comes to real semiconductor product companies, those are far and few.
And, I haven’t seen any real activity in the recent past that could tell me more such initiatives are in the pipeline. Nor do I think there are many attempts to even incubate such companies. On the contrary, there’s a mad rush toward solar!
No harm there! Solar is great for India and the need of the hour. However, India should not forget its semiconductor priorities as well! Indian simply cannot bank on chip design services and solar gains, and then proclaim that it has a very successful semiconductor industry! Real action is still quite far away.
I think, India needs to rethink its semiconductor strategy! It cannot survive on chip design alone.
“When you know the notes to sing, you can sing most anything,” concludes the song from The Sound of Music!
So, is the Indian semiconductor industry hitting the right notes? That’s going to be my next blog post, friends.