The 450mm Semiconductor. Wafer Size Transition contrasted to Prior Generations

Transcription

1 The 450mm Semiconductor Wafer Size Transition contrasted to Prior Generations April 29, 2013* By G. Dan Hutcheson VLSIresearch inc

2 We all know why some want to make the leap to 450mm wafers. Larger wafers have always brought benefits in terms of an areal gain that is about 2-2.5x, a cost-per-transistor benefit that is roughly equivalent to a node, for a capital cost rise that is about the same as a 1 node or about 30%. In other words, jumping wafer sizes is like getting a free fab for everyone you build and it s great insurance for an EUV delay. But it s no secret that after the 300mm disaster, suppliers fear it. Some believe that 450mm could be as expensive as $40B to develop, which is roughly equivalent to 7 years of equipment industry R&D at current spending rates. In other words, all other R&D must cease between now and 2018 to do 450mm. Clearly that can t happen and moreover, current spending rates indicate a much lower ramp. When I look at the historical spend rates for wafer size transitions, the difference in management so far from 300mm, and assume we don t have to backtrack, I come up with a more reasonable $8B. Admittedly, it s a lot more that 200mm, but the industry is also 5x the size it was then. Equipment Industry R&D Spending on New Wafer Size Platforms I don t think that anyone can deny that we are doing much better this time around. We certainly got the order of development in a much more sequential order, which should prevent the 1998 U-turn. That redevelopment of tools was largely because automation standards were not developed before the tools.

3 300mm versus 450mm Progress Milestones 1998 was a critical inflection point for 300mm wafer development, as this industry stalled. Resistance to it started building early in the year. The industry was heading into a deep downturn, triggered by the Asian Financial Crisis. Intel canceled their 300mm fab in Ft. Worth, which was never built. Then Motorola reassigned it 300mm people and didn t show up to an ISS panel, but sent the message that they were just as committed as ever. Motorola had been pounding the table, threatening to never buy from equipment suppliers if they weren t ready by The unwritten message sent that became true was that they d never buy much of anything again. There was another fundamental problem: Canon was still the only litho tool company in the 300mm game. ASML & Nikon chose to focus on 193nm instead, which proved smarter. This timeframe would mark the beginning of Canon missing 7 generations of 193nm tools. Today, is EUV any different? By April of 1998, Toshiba, NEC, and Hitachi were talking up 300mm, but were not disclosing that they had no funds. By June, Japan started talking up skipping to 400mm and equipment makers began to feel like they d been duped. Applied Materials and TEL had $1.2B in 300mm inventory that would be scrapped due to postponements and subsequent redesigns. Then, Applied quietly threatened to walk out of I300I because the latter had a strategic operative that made them unwilling to change with the times. I then gave my infamous revolution speech and followed with a Chip Insider series that pretty much reset requirements. That said, we are at a critical juncture for 450mm and before my panel at ISS, there were only a few straight talkers. One is Morris Chang, who wants it ready for a TSMC pilot line in

4 2013. The other is Eric Meurice, who says ASML will have their litho tools ready, in all the appropriate beta flavors by 2016, providing that chip makers belly up to the bar with clear commitments (re: orders) for mm fabs. At least Eric is committed and ASML is in a far better position than Canon was. One thing is clear: the period between 2012 and 2015 is very critical. So far, 450mm has been pretty inexpensive for the equipment industry. Our analysis shows they ve spent about half a billion dollars and that s including all the overhead and ancillary costs like airfares and hotels at all the conferences. So far the ramp has been very slow from what we can see, this slow ramp will continue through this year. But given history, there must be a steep ramp in equipment industry spending by 2013, or 450mm will not be in production by Every year this ramp gets delayed, will be a year of delay for the production ramp. Equipment Industry R&D Spending Ramps on New Wafer Size Platforms

5 Even though this slope looks steep, the accumulation is not much different than prior generations of wafer sizes. What equipment suppliers will spend developing 450mm

6 Areal Development Cost of Fab Capacity by Wafer Size The ratio of Equipment Industry Development Spending on New Wafer Size Platforms versus the Cost of a Wafer Fab should be more favorable than any prior generation of tool sets. That s assuming a $6B fab, which is what they spend today. This ratio is extremely important because it translates into a quicker payback for the wafer size. 300mm was almost 4, so that meant that 4 fabs had to be built before the spending equaled the capex. But the equipment spending in that meant that more like 6 fabs had to be built. Assume an R&D ratio of 15% and 40 fabs had to be built just to cover the R&D. But wait, you have to divide that by 20% to account for the wafer size portion, so a total of 200 fabs had to be built before the R&D spent on 300mm was paid off. That s why this ratio is so important. Another number that is critically important is the relative areal cost of the capacity versus the equipment development cost. 300mm was a disaster, because at almost $35 per square centimeter, it was over 6x the semiconductor revenues that would come off that area. So how would the semiconductor industry make enough in a reasonable timespan to pay it back? That s why the equipment industry got stuck with the bill and the reverberations have been terrible. Since 2000, equipment development spending has been a flat line. The semiconductor industry has had to share an ever larger percentage of the load.

7 Semiconductor Industry RD&E Spending History F When we came up with the $8B R&D bill for 450mm, we assumed that R&D would continue to flat-line between now and But that raises an issue I did not factor into the analysis: if the 450mm R&D spend has to come out of a flat line, then that means it has to come out of process node spending. In other words, it inherently assumes Moore s Law must slow. If we zoom in on the problem, we can see it much more clearly. In the past, R&D spending always bumped up to accommodate the wafer size transition. Then there was a spending holiday, where it ramped down. But that did not happen with 300mm. Spending continued at high levels because of all the exotic new materials and processes we were introducing to stay on Moore s curve. I doubt that these pressures are going to disappear.

8 Equipment Industry RD&E Spending History F Now if you re an old school chip maker you re probably thinking, Why should I care? It s their money and they have to do this to keep my business. I certainly am. But that assumes your business is worth keeping for anything other than survival. If equipment industry R&D has flatlined, it must have done so for a real reason not because R&D is getting cheaper to do. Each new node of Litho tool development costs about a billion dollars today. Applied Materials spends more than two billion advancing its tool sets each node. If anything, R&D has flatlined only because of consolidation efficiencies, which I talked about last year (i.e. Bruck s rule). Otherwise, R&D expenses would be much higher. In other words, the patient is stent for better revenue flow. You see, the R&D flatline has occurred because the equipment industry s revenues have also flatlined. Put on an MBA hat and you say, it doesn t make sense to invest in 450mm if you are an equipment company.

9 Semiconductor and Equipment Revenue History 450mm will make the problem worse, not only because of the cost of its development, but the additional cost of carrying process development for 2-3 nodes across two wafer sizes. We must avoid the 300mm problem of A and B teams and continual push-outs. This is a big worry for equipment suppliers. All these reasons are why 450mm became a battle of the party of Yes versus the party of No. That is changing to maybe but only if you pay me. The equipment industry is far more positive about 450mm this year than they have ever been before. Moreover, I would argue that for all the reasons above, the equipment industry is likely to turn into a growth industry again. The customers who buy most of the tools have a justifiable need for 450mm, which means it will happen. That also means that it must be paid for if the equipment industry is going to supply the tools. Moreover, from a business perspective semiconductors are still very interesting to equipment suppliers. A decade long attempt to find other revenue rivers in LED and Solar has resulted in a few streams and lots of creeks. But nothing comes close to replacing the semiconductor industry. Even if all lights were LED tomorrow, the lighting industry is only a third the size of the semiconductor s. Worse, LEDs last a lot longer so the market will inherently shrink. Solar is a business of far greater government subsidy than

10 semiconductors have ever been. Semiconductors have a lot better value proposition, which means the growth opportunity for equipment makers has to be better over the long run. Finally I have not written this to be on either party s side. I have intentionally steered a middle course, so whichever party you re in means you ll probably see me as being on the other side. I am not. The intent is to lay out the fundamental business and economic issues so the industry can deal with them on a rational plane. Source: White Paper is from wesrch ( used with permission.

11 Addendum About the Author G. Dan Hutcheson is CEO and Chairman of VLSI Research Inc. His career spans more than thirty years, in which he became a well-known visionary for helping companies make businesses out of technology. This includes hundreds of successful programs involving product development, positioning, and launch. Dan is a recognized authority on the economics of innovation and has a proven track record of being able to predict trends accurately using the economic models he develops. He is a senior member of the IEEE and a recipient of SEMI s Award for outstanding contributions in marketing for his extensive development of the economics behind the semiconductor industry. He has authored numerous publications on the economic, strategic and tactical aspects of how to succeed in the business of technology, which includes numerous articles and weekly analysis of the memory market. He is respected as a Moore s Law scholar, having published many papers and invited talks on the subject over his career. This includes, most notably, work for the IEEE, the Semiconductor Industry Association, the National Institute of Standards and Technology, SEMATECH, SEMI, and The Electro Chemical Society. He has twice authored invited articles for Scientific American on Moore s Law. He has also been the keynote or invited speaker on many other technology topics at dozens of conferences, including the Robert S. Strauss Center s Technology, Innovation and Global Security Speaker Series. Dan is arguably best known for his forecasting of strategic infrastructure shifts and his early-eighties development of the first factory cost-of-ownership models, which are now the basis for most large-scale capital decision-making. He predicted the shift of the DRAM memory market from the United States to Japan in the 1980 s, then the shift of it to Korea in 1990 s, as well as the driving forces for the rise of Flash Memory. Most recently, he gave the invited talk, The Future of Memory: Challenges and Opportunities, at the Applied Materials Technical Symposium in Japan. His pro bono work has included serving as an advisor on innovation to the White House Council of Economic Advisors, teaching invited courses on Manufacturing Economics and The Economics of the Internet at Stanford University, and serving on the Board of Advisors to the Extension School at UC Berkeley. His work at UC Berkeley helped the Extension School avoid wasting millions on capital acquisitions, as well as developing the Lifelong Learning strategy that was key in turning this part of the University into a profit center for UC Berkeley.

12 ABOUT VLSIresearch VLSIresearch is the leading provider of market research and economic analysis on the technical, business, and economic aspects within high-technology industries. The company is known for its unparalleled accuracy, innovation in market research, and its sharply focused insight into industries covered. VLSIresearch s primary databases and reports cover the semiconductor, flat panel display, PV cell and module manufacturing, and associated high-technology industries. Since the early eighties, VLSIresearch has been the only market research firm to consistently gauge the direction of semiconductor markets. The company has always driven to deliver data and services with the most advanced methods, such as the internet and dynamic databases. Its innovations include being the first to use demand-based forecasting; the first to model semiconductor inventories in forecasts; the first to use cost-ofownership models for evaluating equipment; the first to perform customer satisfaction surveys; the first semiconductor industry research firm to apply capacity models in forecasting. VLSIresearch was founded in 1976, and has locations in Silicon Valley, the United Kingdom, Japan, Taiwan and Korea. PRODUCTS AND SERVICES Every day, top executives in high technology industries rely on VLSIresearch s data and information to guide them through business and technology decisions. VLSIresearch products are designed to provide real-time information to support your decisions. They include subscription services, multi-client studies, as well as consulting and custom projects. SUBSCRIPTION SERVICES VLSIresearch s subscription services are focused on selected high-technology segments. They all share the objective of providing a real-time knowledge for your business and provide key statistics with frequent updates related to each sector. These datasets are supported by timely in-depth analysis and VLSIresearch s proprietary research. The high-technology segments covered by VLSIresearch s subscription services include the semiconductor & electronics supply chain, semiconductor and display manufacturing, and PV Solar manufacturing. MULTI-CLIENT STUDIES Multi-client studies typically provide annual analysis and survey of the market or topics covered. They vary significantly and VLSIresearch constantly evaluates new projects to be covered. VLSIresearch s current recurring multi-client studies are Customer Satisfaction Survey, Test Consumable Market Reports and PV Solar Fab Databases. CONSULTING AND CUSTOM PROJECTS VLSIresearch s team of analysts and consultants constantly engage with major companies and investors in various projects. The projects vary in scope from simple two day engagments and presentations to multi month complex industry reviews and analyses. VLSIresearch can draw from the wealth of experience and expertise within our team as well as from its international network of affiliates.

13 NOTICES, TERMS OF USE AND DISCLAIMERS, ETC. Copyright 2013 by VLSI Research Inc. All rights reserved. By accepting this report, opening it, or using it you are agreeing to these terms. This report contains valuable proprietary information developed or acquired by VLSI Research at great expense. You have a limited license to hold these materials but do not become the owner of any material. The materials provided are protected by copyright, trade secret, and trademark law. This presentation has been approved by VLSI Research for public release with attribution. The information in the materials may be used by you on a limited basis in your own documents, provided that those documents are not-for-sale, VLSI Research is clearly referred to as the source of such information, and you obtain written approval prior to use. This report is provided on an "AS IS," "WHERE IS", "WHERE AVAILABLE", WITH ALL FAULTS basis. VLSI Research does not warrant these materials or the information provided therein, either expressly or impliedly, for any particular purpose and VLSI Research specifically disclaims any express or implied warranties, including but not limited to, any express or implied warranties of TITLE, ACCURACY, NON-INFRINGEMENT, MERCHANTABILITY or FITNESS FOR ANY PARTICULAR PURPOSE OR USE. The sources of the information in this report include numerous individual reports, memos and bulletins from various segments of the industry, annual reports, financial reports, interviews, questionnaires, surveys, technical symposia, trade journals, technical journals and individual assessments by knowledgeable company or industry representatives as well as our own analysis and judgment. Some companies are more cooperative about providing information than others and some companies decline to provide or validate the accuracy of any information. Although the information provided is obtained or compiled from sources VLSI Research believes to be reliable given the oftentimes difficult circumstances under which it is collected, VLSI Research cannot and does not warrant or guarantee the accuracy, validity, truthfulness, timeliness, or completeness of any information or data made available to you for any particular purpose. In no event will VLSI Research be liable to you or any third party, whether in contract, tort or under any other legal theory, for any direct, indirect, special, consequential or incidental damages, or any other damages of any kind even if VLSI Research has been advised of the possibility thereof. We receive letters and s on current topics covered in our services and/or reports that are of interest to our subscribers, as well as comments on our reports. We value that subscriber input and like to use it. By submitting such material to us, unless you tell us specifically not to publish it, or except to the extent that you give us an embargo date before which you instruct us not to publish it, you authorize us to publish and republish it in any form or medium, to edit it for style and length, and to comment upon or criticize it and to publish others comments or criticisms concerning it, as the case may be. This report may contain information concerning stocks that is obtained from the opinions of industry analysts. Quoted past results are not necessarily indicative of future performance. None of the information should be seen as a recommendation to buy or sell any securities. We are not stock analysts or investment advisors. You should contact a registered investment advisor as to the nature, potential, value or suitability of any particular investment action. No information provided is investment advice and any such information is just an opinion and is not tailored to the investment needs of any specific person. Certain statements in this report, other than statements of historical fact, and other written or oral statements made by VLSI Research may be forward-looking. In some cases, you can identify forward-looking statements by terminology such as "may", "will", "should", "expects", "intends", "plans", "anticipates", "believes", "thinks", "estimates", "seeks", "predicts", "potential", and similar expressions. Although VLSI believes that these statements are based on reasonable assumptions, they are subject to numerous factors, risks and uncertainties that could cause actual results and outcomes to be materially different from those stated or projected. Those factors, among others, could cause actual results and outcomes to differ materially from the results and outcomes stated or projected in, or implied by, the forward-looking statements. You should understand that forward-looking statements are not guarantees of results or outcomes. New risks and uncertainties arise from time to time, and VLSI Research can not predict those events or how they may affect you, the reader. VLSI Research Inc does not have any intention or obligation to update forward-looking statements after the date of this report. No part of this report may be used in any legal proceedings nor may any of these materials or the information contained therein be disclosed to any third party, including investors or affiliated firms belonging to investors, outside directors or to your affiliated companies, or reproduced or transmitted to any third party, in any form or by any means mechanical, electronic, photocopying, duplication, microfilming, videotape, verbally or otherwise without the prior written permission of VLSI Research. The Chip Insider, The Industry Pulse Pro, and the CSS 10 BEST logo are registered trademarks of VLSI Research Inc. All other trademarks, service marks, and logos are the property of their respective owners.