Is global semicon inventory level headed for oversupply in Q3?

Early this month, iSuppli had indicated that semiconductor inventory levels may have headed into oversupply territory in Q3.

It said: "Semiconductor Days Of Inventory (DOI) for chip suppliers are estimated to have climbed to 75.9 days in the third quarter of 2010, up 1.5 days from the second-quarter. DOI in the third quarter also was 4.8 percent higher than the seasonally adjusted average for the period."

iSuppli added that the value of inventory was not been this high since the second quarter of 2008, when semiconductor suppliers’ stockpiles peaked at $35.4 billion.

Thanks to Jon Cassell and Debra Jaramilla at iSuppli, I was able to speak with Sharon Stiefel, analyst for semiconductor inventory and manufacturing for iSuppli on this situation.

Is there really an oversupply?
I asked Sharon Stiefel that given the growth that 2010 has seen so far, why are semiconductor inventory levels heading into oversupply territory in Q3?

She said that semiconductor inventories, overall, have risen both in terms of DOI and dollars for the past several quarters, and not yet achieved pre-recession levels last seen in 2008. "The overly lean conditions of 2009 and early 2010 are giving way to inventory levels, which are more appropriate for the strong growth experienced in 2010.

"Oversupply in Q3 2010 is not a foregone conclusion, but is possible that if the companies are not able to match manufacturing run rates with demand as the year winds to a close," she added.

Which sectors have been witnessing or recording some softness in demand and why?

Stiefel said: "Companies reporting Q3 revenues over the past two weeks have reported a softening in demand, particularly in PC and consumer end markets, attributed to the continued uncertainty in the global economy, leaving consumers unwilling to spend. A company with more exposure to these sectors has more potential of excessive inventories, versus a company with a more balanced product portfolio."

Industry needs to moderate inventories
It is also said in iSuppli's release that: 'The industry will need to moderate inventories at the appropriate time in its growth curve in order to capture current revenue opportunities while they still exist.' So, when exactly is that appropriate time?

Stiefel noted: "The appropriate time is when sales opportunities exist – projected quarters of growth, rather than revenue contraction. Semiconductor revenues are projected to grow in Q3 2010, contract in Q4 2010 and Q1 2011, and then resume moderate single digit growth for the remainder of 2011."

The iSuppli study also indicated that 'Inventory is not increasing at a uniform rate throughout the supply chain, despite the overall expansion during the past four quarters.' Well, does that suggest the industry has been neglecting this situation? If yes, why?

According to Stiefel, inventory management is a high priority for companies, and this statement is reflective of each sector’s unique responses to the supply/demand balance.

What's the situation like in the foundries and how do they stand against the fabless companies?

She indicated that foundries are running at near full capacity and inventory levels have risen to meet this high utilization level. Fabless companies as a group have seen a jump in inventory levels, returning to pre-recessionary levels in Q2 2010. "The days of inventory growth for both foundries and fabless companies are more in line with projected revenue growth," she added.

In what ways will makers seek to balance the "the current situation of long lead times and capacity constraints against concerns regarding of softening demand through the end of 2010'?

Stiefel said: "Companies have adopted different strategies for increasing capacity to bring lead times down to targeted levels, such as outsourcing capacity to front end fabs or back end assembly and test facilities, or purchasing equipment to relieve manufacturing bottlenecks.

"Thus far, in the Q3 reporting period, several semiconductor manufacturers are reporting that lead times are returning to normal, and component shortages are becoming a more isolated occurrence. This is an indication that balance is being realized by those particular companies."

Preventing inventory bubble burst!
Finally, is iSuppli seeing a softening of global demand? Will makers be able to prevent the inventory bubble burst from happening?

According to her, all indicators are that global growth is continuing, albeit at a slower pace than anticipated. "I believe that this data is no surprise to manufacturers, who will continue to keep a tight reign on inventories, and avoid a 'bubble' situation."

Part 2 of my discussion with Sharon will look into whether the global semiconductor industry set for a soft landing in 2011!

Xilinx announces first stacked silicon interconnect technology

Xilinx Inc. announced the industry's first stacked silicon interconnect technology. It proposes to deliver breakthrough capacity, bandwidth and power savings using multiple FPGA die in a single package for applications that require high-transistor and logic density, as well as tremendous levels of computational and bandwidth performance.Xilinx has taken a 3D packaging approach that makes use of passive silicon-based interposers, microbumps and through-silicon vias (TSV) to deliver multi-die programmable platforms. As the interposer is passive, it does not dissipate any heat beyond what's consumed by an FPGA die.

The stacked silicon interconnect technology offers 2X FPGA capacity advantage at each process node. It is a core part of Virtex-7 family. Also, the stacked silicon interconnect technology is supported by standard design flows.

Xilinx has been accelerating FPGA transition to the heart of the system. David L. Myron, senior manager, High Volume Products, Product and Solutions Management, Xilinx, revealed that a lot of Xilinx's customers are doing FPGA starts rather than ASIC starts as that seems more viable. Customers are now asking for much more -- over 2X today's logic capacity, many more high-speed serial transceivers as well as processing elements, as well as much more internal memory to store data. "The challenge is delivering 'more than Moore'," he said.

Myron cited certain challenges. These include availability and capability -- the largest FPGAs are only viable later in the life cycle. Power and bandwidth pose additional challeges. The traditional mitigation techniques are no longer adequate. "One of the trends we have seen is that while gate count has gone up at a certain rate, the I/Os have not," he added. Hence, innovation is the need of the hour to meet capacity requirements.

The stacked silicon interconnect technology is addressing all of these challenges, easily meeting the needs of high bandwidth, low latency and low power. This Xilinx innovation offers massive number of low latency, die-to-die connections. Besides, there is no wasted I/O power.

For applications requiring high-transistor and logic density for high levels of computational and bandwidth performance, these 28nm platforms will deliver significantly higher capacities, resources and power savings than possible in a monolithic die approach.

Suresh Ramalingam, director, Package Design and Advanced Package Development, Xilinx, said that Xilinx's FPGA architectural innovation has been at the heart of the technology. FPGA slices -- which are ASMBL (application specific modular blocks) optimized FPGA slices -- are placed side by side. Use of silicon interposer enables high bandwidth connectivity. He added, "We have enabled direct connections to the logic region."

The stacked silicon interconnect technology harnesses proven technology quite uniquely -- such as microbumps, TSV, passive silicon interposer (65nm) and side-by-side die layout. The technology enables 100x bandwidth/watt advantage over conventional methods. It also delivers feature-rich FPGAs.Ramalingam added: "Going from 40nm to 28nm, there is 3.5X times improvement in logic cells. When we bring 28nm to market, we will have another 2.8X advance in capacity. Xilinx is providing optimized ISE design suite flows for a variety of users.

With software support available from Xilinx in ISE Design Suite 13.1, currently available to beta customers, the 28nm Virtex-7 LX2000T device will be the world's first multi-die FPGA and provide over 3.5X logic capacity of the largest current-generation Xilinx 40nm FPGA with serial transceivers and 2.8X the logic capacity of the largest competing 28nm FPGA with serial transceivers. Initial devices will be available in the second half of 2011.

Xilinx has benefitted by collaborating with technology leaders such as leading fabless and fablite companies, equipment manufacturers, fabs and OSAT, as well as industry consortia such as imec, Sematech, SEMI, etc.

The India launch was held today, October 28th, while the Taiwan launch was two days ago.

Why has the semicon equipment bubble really burst? - II

Here's the concluding part of my discussion with Dr. Robert Castellano of The Information Network.

Repercussions of a deteriorating semiconductor industry
I asked Dr. Castellano regarding the repercussions of a deteriorating semiconductor industry.

He said that the semiconductor equipment industry seems to be in serious trouble. There could possible be little growth in 2011, and the how is that there will be sufficient pushouts in equipment that revenues are moved to 2011 from 2010.

Dr. Castellano said: "We warned two months ago about pushouts, and today, Veeco stated that they “recently experienced rescheduling of tool shipments from the fourth quarter into the first quarter by several customers in Korea and Taiwan”. In other words, pushouts! We will continue to see this more and more.

"Problem is, will the equipment vendors admit it? ASML vehemently denied any customers’ pushouts last quarter, but with tools selling for $35 million each and customers such as Nanya and Inotera announcing losses, there is no way in creation pushouts won’t happen.

"Then, there is the issue of 450mm wafers. The only ones pushing it are the semicons, because they recognized that they could generate twice the number of chips for almost the same capital equipment cost. The equipment industry was dramatically impacted by the 300mm transition, and growth was nearly flat from 2001 to 2009. Not so for the semicons.

"No equipment supplier wants 450mm, it is being pushed by Sematech and Intel, plus a consortium in Europe that feels that perhaps 450mm will knock off competitors and they can make up the vacuum in sales. Only the top 15 equipment suppliers will survive."

How will pushouts benefit the industry?
On the subject of industry pushouts being highlighted time and again, it is also necessary to see whether and how will these benefit the industry in the long run.

Besides the reasons mentioned earlier, semiconductor sales are intimately tied to the economy. There is a direct correlation between semiconductor sales and GDP, as well as the PLIs of The Information Network. If the economy is robust, more money is available to purchase electronic items containing semiconductors. The reverse is true, indicative of the present economic climate.

The Information Network has also indicated that firms will announce lower results, and it'd get worse in the following quarter. Why will this happen and which firms could be likely 'hurt'?

Dr. Castellano said: "This will happen because the crest in the tidal wave was only reached in the past month or so, and it is a long and slippery slope down because it went so high up to begin with.

"The DRAM manufacturers will be hit the hardest. Growth was strongest for them for the first half of the year, where sales grew 135 percent in Q2 2010 compared to Q2 2009."

Is there a way out? If yes, when?
Finally, when will there be some recovery in the semiconductor equipment sales and why? Surely, as with everything, there has to be a way out!

Dr. Castellano concluded: "We see minimal growth in 2011, again depending on macroeconomic factors. We see two years of downturn in the industry – 2012 and 2013."

Why has the semicon equipment bubble really burst? - I

Early today, I'd mentioned about receiving an interesting report from The Information Network — where it said that the global semiconductor equipment industry bubble has burst!

It made interesting use of an analogy around “The Emperor’s New Clothes,” a short tale by Hans Christian Andersen and the global semiconductor industry. So, I got in touch with Dr. Robert N. Castellano, president of The Information Network, New Tripoli, USA, to find out more.

Just why did the bubble burst?
I started by asking him why The Information Network has been pointedly indicating that the semicon equipment bubble has burst?

He said: "If we take a look at the SEMI book-to-bill ratio, bookings were down from $1,837 million in July and $1,816 million in August to $1,616 million in September. Keep in mind that these are three-month moving averages. so that September's numbers were proped up by stronger July and August bookings.

"Additional data come from our proprietary leading indicators (PLI) that we have developed obver the past 15 years. They point to changes and inflections in the economies of the world and correlate with inflections in semiconductor equipment revenues several months out. We plot SEMI's announced billings (revenue) instead of bookings, which are anticipatory. Our PLI has been trending downward for the past three months, signalling an inflection in equipment revenues. We will see this happen this quarter."

Pitfalls of two years of growth combined into one!
The report has also indicated that year 2010 is the same as 2000 -- where two years of growth were combined into one. What are the pitfalls from such a development?

Dr. Castellano added that in 2000, equipment revenues skyrocketed, followed by a severe downturn in the following year. In a typical cycle, we see about three years of growth. But not so in 2000. The reason for the large growth was inaccurate market forecasts, when some 'analysts' kept hyping shortages in certain ICs, particularly DRAMs. This led IC manufacturers to purchase more equipment and build more fabs to meet the anticipated growth.

Little did they realize that the Dell Computers of the world were also reading the same 'erroneous' forecasts and purchasing twice the number of ICs they needed for fear of shortages.

The IC companies, not realizing the customers were double dipping, thought that the phenomenon was real and kept expanding. In 2001, IC manufacturers were left with about $10 billion in excess inventory. The year 2000 coincided with Y2K. Later that year, the Internet bubble also burst. So, growth came from anticipated applications, rather than real demand.

He said: "The same thing is happening in 2010. The world is mired in poor economic conditions, and even in a good year with healthy GDP growth, semiconductor growth of 30-50 percent is almost unheard of.

"And, for 2010, in which equipment purchases were meant for technology advancements to the next node, and not for capacity increases because of demand by consumers for new gadgets, growth of 100 percent in the equipment sector was virtually impossible with purchases made in a sane process.

"However, when the forecasts coming out of SEMI and other 'analysts' pointed to hyper growth, IC manufacturers, probably for fear of not being able to secure 'minimally needed' equipment, placed orders. The problem was compounded with companies such as Samsung, that went all out and practically locked in the supply of ASML’s immersion tools. Other memory manufacturers placed orders for fear of not getting any. We warned months ago that growth in DRAMs over 100 percent was not sustainable. Sure enough, Nanya and Inotera both reported losses for 3Q10."

The emperor's new clothes
There's this interesting analogy about 'The Emperor has new clothes' with the semicon industry.

Dr. Castellano elaborated that in this scenario, they see the Emperor as the executives of the DRAM manufacturers. The rest of the cast remains the same – tailors are 'analysts' and the child is The Information Network that warned the Emperor of problems with no success.

Here, the citizens are the mainstream press. They continue to publish forecasts from these 'analysts' and never call them to the carpet about their results or how they seem to change monthly.

Well, as a member of the technology press, I have tried my best, and sincerely. I really don't know with how much success, but I do try and call or email each analyst, seeking more insights, which are covered in my posts.

Industry overlooking repeated warnings?
How can the industry afford to not overlook repeated warnings? This is peculiar!

According to Dr. Castellano, certain 'analysts' -- right or wrong, get covered in all press releases. They are referenced by semis and equipment manufacturers as a matter of fact.

He gave an example: "Several years ago, I spoke with someone I knew in the PR department of Applied Materials. They had just published a release and quoted numbers from Dataquest. I mentioned to her that the numbers were not correct, and she owed it to her readers and investors the correct numbers. She admitted my lower numbers matched their internally derived numbers. Some months later, Applied issued another release with the same inflated numbers. I called her and she said the IR department told her that, right or wrong, she had to put a Dataquest forecast in the release.

"This is the problem in the industry. And companies get bigger because of the support, which exasperates the problem. SEMI always partners with Dataquest at ISS and Semicons. It can’t be changed."

I'll add more in a second part a bit later.

What's Farnell (element14) up to? And, semicon equipment bubble burst? Whoa!!

I’ve received two interesting reports today. One, from Premier Farnell on a name change and another from The Information Network on the troubles of the semicon equipment industry.

First, Premier Farnell plc announced the launch of a new, market leading customer proposition in Asia offering a significantly expanded inventory of products relevant to electronic design engineers, next-day delivery in Asia Pacific and an enhanced range of services, under the name element14.

I have not quite understood this move!

I wonder how is this going to help Farnell! It has had element14 for so long now. Hey, I even provided Farnell feedback on the element14 website back in July 2009 (in Bangalore), and even discussed with its executives in Singapore when they sought further queries on the feedbacks provided by me.

What’s Farnell (or element14) really up to?
At least, I don’t see any newness in Farnell’s element14 release. Unless, there’s a move to dilute the Farnell brand itself, which is and will be really strange!

I for one felt element14 was (and is) doing magnificiently as a standalone entity, and has done a great job in building a community of design engineers around itself.

The release also says that the Group’s businesses located in Australia, China, India, Malaysia, Singapore and New Zealand will be known as element14 and will offer customers and suppliers a multi-channel, multi-currency, multi-language way to source their product needs, quickly and easily.

So, I checked Farnell’s India site. Now, the India website of Farnell is confusing me. There’s a message on the site that says ‘Farnell Electronics (India) Pvt Ltd name change in progress’.

Where is the need to call Farnell by its original name in the UK, and why rename it elsewhere, especially in Asia? Is it that the brand name is not well known across Asia? What’s wrong in the first place with having two separate standalone and successful identities such as Farnell and element14?

It would be prudent on Farnell’s part to maintain element14 as the community it has been originally meant to be. For instance, why call Farnell India as element14 India all of a sudden?

I hope to hear from my friend, Ravi Pagar at Farnell soon, hopefully, in the morning. Have sent him the query. Name change or not, best wishes to element14/Farnell in its future endeavors.

Semicon equipment industry bubble burst?
Now to the second one! I also received an interesting report from The Information Network — where it says that the global semiconductor equipment industry bubble has burst!

It has also made interesting use of an analogy around "The Emperor’s New Clothes," a short tale by Hans Christian Andersen and the global semiconductor industry.

The report has set me off thinking. Just how can the global semiconductor industry afford to not overlook repeated warnings? This is peculiar! What could be the repercussions of a deteriorating semiconductor industry? The Information Network also mentioned industry pushouts. How will these benefit the industry in the long run?

I hope to speak with Dr. Castellano regarding the study and get some replies on my queries.

What’s very interesting is a prediction made at the end of this report. The Information Network says that firms such as AMAT, AMD, APD, ASML, AVZA, CAJ, FJTSY.PK, HPQ, INTC, MU, NVLS, OMG, TOELF.PK, TOSBF.PK and other IC, equipment, and material suppliers have or will be announcing their results soon. “Nearly all will point to a slowing in the next quarter in their releases and conference calls. It will get worst in the following quarter!”

Wow! That will be significant! Hey folks, I’m going to watch all of these firms very closely as they annouce their quarterly results. Suggest, you do the same.

Rankings of various semicon market watchers' sales growth expectations 2010

This is a continuation of my coverage of the fortunes of the global semiconductor industry. I would like to acknowledge and thank Mike Cowan, an independent semiconductor analyst and developer of the Cowan LRA model, who has provided me the latest numbers.

Mike routinely tracks how the Cowan LRA Model sales growth forecast (each month) compares to a wide range of major semiconductor industry market watchers in order to monitor how the model's latest monthly update result "stacks up" against the competition, that is, 15 other prognosticators he routinely monitors.

Table 1 summarizes the latest sales growth forecast expectations of these 15 other semiconductor industry forecasters, thus comparing Cowan's forecast number to the rest of the crowd.Source: Cowan's LRA model.

Cowan routinely updates his forecast numbers each month immediately following the WSTS release of its monthly actual sales numbers in order to "dynamically" predict the industry's sales posture. Consequently, the sales and sales growth output of the monthly model run does NOT "sit still" as highlighted below:

Cowan LRA model for forecasting global semicon sales --– 'divining' (mathematically) future from past
* It should be emphasized that each month’s actual global sales number published by the WSTS is a "lagging indicator" since it is released a full month after the fact.

* The Cowan LRA Model, however, "turns" this lagging actual monthly sales into a "leading indicator" by virtue of its near-term forecasting capability looking out over the next five quarters.

* This is the "beauty" of the model and, therefore, makes it dynamic in the sense that it can be run each month utilizing the most recent actual global semiconductor sales number published by the WSTS. Thus, it allows “rigorous tracking” of the near-term sales forecast outlook for the global semiconductor industry on an “almost” real-time basis.

* Consequently, the model’s monthly sales forecast does not “sit still” but “evolves” with each succeeding month’s latest published sales number. Since conditions change rapidly and unexpectedly in the semiconductor industry, industry market forecasters are hard pressed to keep up with these changes. How can the industry management be sure that a sales forecast issued two, three, or more months ago is still valid and relevant to what's happening in today’s global semiconductor market?

In order to illustrate this "does not sit still" principal embedded in the monthly update to my forecasting approach, here is a graphic showing the month-to-month evolution of year 2010 sales and sales growth forecast estimates as determined by the model from January to the present.Source: Cowan's LRA model.

Global semicon market set for slowdown
Before I conclude, here's a wonderful piece of information sent to me (and to several others) by Dr. Robert Castellano of The Information Network. Two months ago, I had done a post with Dr. Castellano where he'd maintained that the global semicon market is set for slowdown.

The Information Network had, on September 21,issued a release similar to the report I had done with Dr. Castellano to all of its readers. This was at a time when the semiconductor market was humming along and many 'analysts' continued to up the ante in the market.

It seems that nobody believed them! Well, maybe not nobody, but few (including yours truly!). Fast forward another month and The Information Network reiterates the slowdown. Dr. Castellano says, "This time, we have a lot of company and it's somewhat vindicating."

Dr. Castellano had actually pointed out problems in the DRAM sector. "Isn't it interesting that Nanya and Inotera both reported losses for 3Q10. Again vindication for being slapped around by managers and shareholders in denial," he added.

Dr. Castellano notes: "I also pointed out that we would be seeing problems in the semiconductor equipment market, particularly lithography. Isn't it interesting that ASML went out of its way in its conference call that they were not seeing any pushouts in purchases. Interesting that ASML never bought our lithography report in the 25 years we are in business. It reminds me of when I pointed out last year that Intel was losing money on the Atom, and every conference call after that they said they were making money on the Atom (maybe Monopoly Money!), yet Intel never purchased our report on ARM versus Atom (ARM did!)."

The question now is: Has the Internet really become such a marketing tool that multibillion dollar sized companies want to save a few thousand dollars and only read headlines instead of buying insightful marketing reports? Or maybe they are just buying from the wrong market research firms.

Or well, are such companies even able to really 'understand' what's actually happening in the industry?

PS: Hope that most of you have also read that SEMI's latest North American semiconductor equipment industry Sept. 2010 book-to-bill ratio is 1.03, which is 11 percent lower than August 2010.

Enhancing reliability in medium voltage power line communications

Semitech Semiconductor is solving current problems faced in communications over the power grid with its line of innovative semiconductors designed to address power line noise. Owing to noise and variations in equipment and standards, communications over power grid have been historically difficult. Semitech’s proprietary technology design produces chips that reliably operate in this challenging environment.

Semitech CEO Matt Rhodes recently spoke at the Smart Grid Electronic Forum 2010, in San Jose, California on October 18-20, where he presented on "Enhancing Reliability in Medium Voltage Power Line Communications,”

Rhodes examined smart grid characteristics observed from field trials of medium voltage power-line communications and suggesting solutions that enable reliable communications. The presentation also explored using the OFDMA scheme as a means to enhance communications reliability.

Here's a snapshot of his presentation, thanks mainly to Kevin Mayberry, Lages & Associates.

Semitech Semiconductor, in conjunction with a Chinese equipment provider, has implemented OFDM to deliver robust, easy to deploy communication over the medium voltage grid.

The SM2200 is Semitech's high-speed powerline communication solution, a next generation OFDM PLC modem. The SM2200 adds OFDMA to provide most networking flexibility.

The SM2200 implements OFDM technology, and adds adaptation and multiple access features to provide the most flexibility for the networking application. OFDM and the SM2200 are proven in a broad deployment on the medium voltage grid in the Hebei province of China.

The 2000-node system has been operating reliably for nearly two years. Based on these results, a broad deployment across China’s medium voltage grid is in progress.