SAN JOSE, California – There are an alarming lack of new, leading-edge fabs on the drawing board-a trend that could propel the IC industry into a nightmarish era of tight capacity, higher chip prices and shortages.
Capital spending is on the rise in 2011, but leading-edge fab capacity is expected to be tight for the foreseeable future for various reasons. Fab companies are reluctant to build new 300-mm plants and are only upgrading their existing facilities. And at least three fab companies-Intel, Samsung and TSMC-have suddenly become pre-occupied with 450-mm activity.
It’s unclear if fab companies will continue to build new 300-mm greenfield fabs in the future. Perhaps they will bypass 300-mm and simply take a quantum leap into the 450-mm era.
Analysts have warned that there is not enough 300-mm capacity right now. And costly 450-mm plants are delayed and not expected to go into mass production until 2017 or 2018, according to IC Insights Inc.
What that means is that there could be a noticeable gap in the transition between 300- and 450-mm, leading some to believe there could be possible disruptions and chaos in the IC supply chain.
It also leads some to believe that leading-edge fabs will become endangered species. Only a select few can afford to build leading-edge fabs today. Fewer will be able to afford 450-mm plants. And only a few 450-mm plants will be required to meet future IC demand, leaving the balance of manufacturing power in the hands of a select few.
Times have changed in the IC industry. Several years ago, nearly all chip makers owned fabs. Then, in the late 1980s, the foundry business was born, fueling the growth of the fabless industry. Starting at that time, the integrated device manufacturers (IDMs) began to outsource a large percentage of their production to the foundries as a means to cut manufacturing costs. More recently, fewer IDMs can afford to build new fabs, as chip manufacturing costs have skyrocketed.
Amazingly, there are only three new 300-mm fab projects slated to break ground for 2011 and 2012, compared to eight new plants in 2010, according to SEMI. The three new fab projects include those from Hynix Semiconductor Inc. (M12), Samsung Electronics Co. Ltd. (L17) and Taiwan Semiconductor Manufacturing Co. Ltd. (Fab 15), according to SEMI.
Spending on new fab construction projects jumped 162 percent in 2010, but that number is expected to fall 11 percent in 2011 and another 26 percent in 2012, according to SEMI. ''Currently, there is a bleak outlook in new fab construction plans,’’ warned Christian Gregor Dieseldorff, an analyst with SEMI. ''Companies (are) focusing their spending mainly on upgrading or ramping existing fabs.’’
Indeed, something has changed in the psyche of semiconductor manufacturers, said Malcolm Penn, founder and principal analyst with Future Horizons, at a recent event. They are no longer building wafer fabs in anticipation of demand. "Forget fab-lite, welcome to the fab-tight era," Penn warned.
Fab companies have been reluctant to build new plants-and for good reason. ''At this point, there is little to no excess leading-edge 300-mm capacity and more and more fabless companies are migrating onto the 300-mm network as they migrate to 65-nm process technology,’’ observed Gus Richard, an analyst with Piper Jaffray & Co., in a recent report. ''The foundries over the last five years have been (also) hesitant to add capacity ahead of demand as the tighter the supply, the better the pricing’’ and margins.
Forecasts and more forecasts
''Given the drastic cuts in semiconductor industry capital spending in 2008 and 2009 and a double-digit increase forecast for the IC market in 2010, 2011, and 2012, IC industry capacity utilization can be expected to remain relatively tight over the next couple of years,’’ warned Bill McClean, president of IC Insights, in a report. ''This in turn is forecast to lead to firming IC ASPs, extended leadtimes, and spot shortages.’’
In 2010, a 28 percent surge in wafer starts, coupled with the 6.4 percent increase in IC capacity, caused fab utilization rates to jump to 93.2 percent, up 15.5 percent over 2009, according to IC Insights. ''For 2011, (overall) IC industry capacity utilization is expected to rise to 93.8 percent before falling back slightly to 91.3 percent in 2012,’’ McClean said in the report.
Leading-edge, 300-mm capacity is expected to remain tight. ''Although in the ‘ramp-up’ stage, capacity utilization for leading-edge <40-nm devices was high in 2010 at 93.8 percent, up from the 90.3 percent figure witnessed in 2009. In 2011, leading-edge capacity utilization is expected to be very tight and register a 96.6 percent utilization rate,’’ he added.
IC Insights also forecasts a 22-point swing in semiconductor market growth (32 percent growth in 2010 and 10 percent growth in 2011), a 2-point difference for electronic system sales growth (11 percent in 2010 and 9 percent in 2011), and a 0.3-point difference in worldwide GDP growth (from 4.2 percent in 2010 to 3.9 percent in 2011).
In IC Insight’s most likely scenario, ASPs will see an increase of 2 percent in 2011, compared to 1 percent in 2010. Unit shipments are expected to grow 8 percent in 2011, down from 29 percent in 2009.
Jim Feldhan, chief executive of Semico Research Corp., said the IC market is expected to grow 8 percent in 2011, down from 31.8 percent in 2010. Capital spending is expected to grow 10 percent in 2011, with fab capacity hovering around 90 percent at the beginning of this year and cooling to 87 percent by year’s end, he said.
Dean Freeman, an analyst with Gartner Inc., sees a glut of capacity in the foundry sector in 2012, as Globalfoundries, TSMC and Samsung are currently increasing their capital spending. The foundry business is expected to see 9.2 percent growth in 2011, down from 40.2 percent in 2010, he said. The IC market is expected to grow 4.6 percent this year, compared to 31.5 percent last year, he said.
For its own forecast, VLSI Research Inc. said the IC market will grow 8.1 percent in 2011, down from 30.9 percent in 2010. And more importantly, 2011 is off to a better start than anticipated.
''Nearly all the chipmaking companies that have reported so far have beaten expectations. They expect Q1 growth to be above normal seasonal growth. This shows that after a weak Q4, the chip industry is roaring back thanks to a strong holiday season and an improving macroeconomic environment,’’ according to VLSI Research.
''The recent trend shows that overall inventories are tight and supply in the channel is dwindling,’’ according to VLSI Research in a report. ''NAND supplies, in particular, are very tight due to strong demand from electronics OEMs. As a result, NAND spot prices jumped higher for the tenth straight week. DRAM is also improving. The decline for mainstream DDR3 has abated.’’
Soaring fab costs
Despite the improvement in business, fewer companies are willing to throw down the money to build new, greenfield plants. One reason is that fab costs are soaring out of control.
At one time, a leading-edge 300-mm fab was about $3 billion to $5 billion. Now, Toshiba Corp. may have set the record for the world’s most expensive fab. The company’s new 300-mm NAND plant (Fab 5), including construction costs, is expected to run $10.5 billion to $11.5 billion, according to SEMI.
At the recent Industry Strategy Symposium (ISS), sponsored by SEMI, Robert Bruck, vice president of the Technology and Manufacturing Group at Intel Corp., said a 450-mm fab could run $10 billion. That figure may not include construction costs.
The R&D costs to develop a 22-nm process is expected to run from $999 million to $1.354 billion, compared to $581 million to $796 million at 32-nm, according to Handel Jones, CEO of International Business Strategies Inc. At 90-nm, the R&D cost ranged from $198 million to $283 million, Jones said.
Fab tool costs are also soaring, especially lithography. In the 1980s, a leading-edge lithography scanner sold for about $1 million. Today, ASML Holding NV’s 193-nm immersion scanners run $40 million each. And a production-worthy extreme ultraviolet (EUV) tool could sell for about $125 million-if or when these machines are shipped.
Photomask costs are also increasing. A 45-nm ''mask set'' runs about $800,000 to $900,000, but a 32/28-nm ''mask set'' could run $2 million or so, according to one photomask vendor. A 22-nm ''mask set'' could double, according to some.
Needless to say, IC design costs are soaring out of control. ''With each successive node the cost of a design goes up,’’ said Piper Jaffray’s Richard. ‘’The cost of a 45-nm SOC chip design is estimated to be roughly $80 million and a 32-nm SOC is $130 million. We estimate that the addressable market of these chips needs to be roughly $400 million and $650 million to make a reasonable return assuming a 50% gross margin.’’
The cost to move from 300-mm to 450-mm fabs is also expected to be enormous. Every 15 years or so, the industry has moved to a new wafer size, ostensibly to stay on Moore’s Law. The transition to a new wafer size has already been bumpy, especially in 300-mm, when fab companies announced and then delayed a plethora of plants several years ago. The fab tool makers developed the first round tools for these plants, but were left holding the bag-and suffered massive losses-when the fabs were delayed.
For some time, four chip makers-Intel, Samsung, TSMC, and, to some degree, Toshiba-have been pushing the industry towards the 450-mm era. IDMs wanted 450-mm pilot lines as early as 2012.
In the beginning, most fab tool vendors refused to invest in 450-mm for several reasons. The return-on-investment remains unclear and no one is sure who will pay for the R&D. And fab tool vendors do not want to repeat the same mistakes in the early days of the 300-mm era.
Then, suddenly, Intel recently announced the D1X fab in Oregon, a plant that is capable of producing 300-mm wafers and is ''450-mm capable.’’ The startup time for D1X is 2013.
Last week, TSMC officially disclosed plans that it will build a 450-mm fab, according to an analyst. The first 450-mm line is planned for Fab 12 Phase VI; it will be for their 20-nm production ramp, according to VLSI Research. TSMC’s 450-mm pilot line is expected to start around 2013 to 2014, with production due in 2015 to 2016, according to the firm.
''That means that you can expect Samsung and Globalfoundries to put 450-mm into their plans immediately. I would bet Samsung will be first or it may well come as a Common Platform alliance deal,’’ said G. Dan Hutcheson, CEO of VLSI Research, in a report.
Now, after dragging their feet for some time, the fab tool community has suddenly warmed up to 450-mm, possibly due to the anticipated announcement from Intel and TSMC. Simply put, the fab tool vendors must get on board in the 450-mm era-or else. Intel, TSMC, Samsung and Toshiba represent a huge percentage of today’s capital spending budget and fab tool vendors can’t afford to miss out on the order window.
''If you are an equipment supplier, it means you can no longer ignore 450-mm. It’s as simple as that. In other words, we all have to get cracking to hit a 2013 target,’’ Hutcheson said.
Intel, Globalfoundries, Samsung, Toshiba, TSMC and others have new 300-mm fabs in or near production. The question is whether those companies build new 300-mm fabs beyond those plants or will they hold out until the 450-mm era?
''Strategically this is an immediate problem for every chip maker planning a fab. They immediately must decide whether TSMC is committed (to 450-mm). After all, at this point it’s all cancellable or can be delayed,’’ Hutcheson wrote in the report.
''That’s a gamble Japan took with 200-mm and they lost their leadership to Korea and Taiwan. Japan had built in a cost disadvantage that they couldn’t recover from. You can’t be a fast follower with wafer size transitions because fabs have such long lives,’’ he said.
''Complicating matters is that you can’t develop a leading-edge process on one wafer size and easily port it to another wafer size. TI tried to do this with 300-mm, where they developed on 200-mm to save money. The result was they shut down research in 1997 and stopped being on the leading edge of process technology. It’s more reasonable to choose to stay behind on wafer size,’’ he said.
''You can bet TSMC, being the pioneer, will be the one with the arrows in their back. TSMC was early with 300-mm and that created problems because the automation standards had not been worked out. But that’s done now,’’ he said. ‘’Now the only real barrier is tool availability and process development. (450-mm) should go smoother than 300-mm.’’
Still others are urging the industry to build more 300-mm fabs in the near term. ''In the late 80s and early 90s the chip industry under invested, helping to create a super cycle in the late 90s. The over investment in the late 90s created a lost decade in the 00s,’’ Piper Jaffray’s Richard wrote in a report.
''We believe excess capacity has been wrung out of the supply chain and chip makers need to increase spending. In addition, the lack of investment in process technology is slowing Moore's law and 450-mm fabs are in the distant future. Thus, more fabs and equipment will be needed to supply the industry,’’ he said.