Creating a Better Roadmap

Technology, Business Factors Must Both Be Incorporated

Bijan Moslehi, PhD, is chief technology officer and senior vice president, semiconductor technology research, for The Noblemen Group, a boutique investment banking, strategic advisory, and business development firm. Moslehi has 20 years' experience working in the semiconductor and semiconductor equipment industries. He can be reached at bmoslehi@noblemengroup.com.

Technology roadmaps play a critical role in the semiconductor industry and have become invaluable for the development of future technologies and products. Driven by ever-increasing device complexities since the early 1990s, a monumental industrywide effort has been focused on developing documents such as the International Technology Roadmap for Semiconductors (ITRS).

Equipment, subsystems, and materials companies rely heavily on these roadmaps as the cornerstone of their strategic marketing, product planning, and new product investment decisions. Suppliers need to work with their customers and leverage their specific technology roadmaps as well, an important but challenging exercise that remains a critical component of product planning, particularly for the large vendors. Many smaller companies are at a bit of a disadvantage in this area, since they may not have access to a broad customer base and may not be able to allocate sufficient investment to an effective infrastructure for strategic marketing and product planning. Consequently, many suppliers rely primarily on the industry roadmaps, which are widely seen as representing the broader consensus.

The ITRS has made a tremendous impact by providing guidelines on future technology needs, and it continues to fill a major gap. However, these roadmaps and the way that they are used by the industry have some serious shortcomings and problems.

The unrealistic timing of several major technology transitions has underscored the weaknesses inherent in overoptimistic roadmaps. Two prime examples of this roadmapping flaw are the failure of true low-k integration at the 130-nm node and the premature introduction of 300-mm fab technology. The implementation of low-k materials has been pushed out by several technology nodes compared with the path
plotted in earlier roadmap plans. The 300-mm "false start," propagated in part by inaccurate roadmapping, delayed the transition to the larger wafer-size technology for several years. These two cases alone have resulted in unnecessarily higher costs and lower overall return on investment for suppliers, as well as disrupting any meaningful planning strategies.

The stated objective of the ITRS is "to ensure advancements in the performance of integrated circuits." The document is "an assessment of the semiconductor technology requirements," which "identifies the technological challenges and needs facing the semiconductor industry over the next 15 years." This technology-centric focus has been the primary cause of many recent roadmap difficulties.

In order to enhance the tremendous value that the document provides, this narrow focus must be expanded to include a strong business perspective. The issue is not whether the technology requirements articulated by the roadmap advance IC performance—clearly they do. For instance, the engineering benefits of the move to low-k materials and other similarly important technology transitions are widely recognized and accepted. The real issues are "proper timing," "realistic assessment of the actual needs of end-markets and end-users," and "better understanding of potential overall economic and business implications" of these major technology transitions throughout the supply chain.

As I noted in the first Reality Check in the January 2004 MICRO, "The intertwined and inextricable nature of technology and business, combined with the realities of the marketplace, must guide the development of aggressive yet practical technology roadmaps along with optimized, business-savvy economic implementation strategies." In other words, in addition to the technologists who do a superb job of highlighting the technical issues, marketing and business people need to get involved and help with issues such as timing, markets, applications, and economics. Furthermore, a management perspective would ensure a more practical and realistic approach. Finally, contributions by systems providers will further improve the roadmapping process, particularly in relation to end-markets and timing. With such a balanced approach, artificial insertion points for major technology transitions would be avoided, and the industry would be driven primarily by end-market requirements and sound business principles, not just technology-inflection points.

Another factor that has not been carefully considered is the notion of latency. By latency I mean issues such as time to technology insertion, time to manufacturing readiness, and time to market adoption. Experience tells us that it takes a long time to successfully introduce a major, complex new technology into manufacturing. In many cases, the time it takes to start serious work on a new technology insertion and then overcome the various obstacles to its successful early adoption may range from 7 to 10 years.

Timing can also be affected by end-market adoption as well as the rate of adoption. In some cases, the technology trend is out of synch with the market trend. For instance, in contrast to just a few years ago, consumers want low-cost, portable PCs with a longer battery life, not just faster systems. Also, many users will no longer pay a premium for speed and memory. These lower-average-selling-price pressures supersede the traditional quest for performance.

Roadmaps such as the ITRS primarily identify the technological challenges facing the semiconductor industry and assess technology requirements, as they should. However, the problems start when the documents also specify the "solutions" that at first glance seem necessary. Because of the enormous complexity of most of these technological challenges, it is quite difficult to endorse a particular solution with enough certainty to preclude other possible approaches. This methodology potentially undermines innovation by either the established equipment, subsystems, and materials companies, or aspiring entrepreneurs. It may be unwise, risky, and costly for a supplier to work on alternative ideas that are different from the solutions promoted by an international industrywide consensus.

The lithography space offers a clear example of an alternative technology that has taken the industry by storm: immersion lithography (IML), which now appears to provide a practical solution over the next several technology nodes. IML was not even seriously considered as a potential mainstream process technology until 2003. This "wet" solution appears to offer a successful evolutionary change, one that will be less expensive, involve fewer technological showstoppers, and take less time to implement than 157-nm and other lithographic methods under consideration. According to International Sematech, 193-nm IML seems to be on track for 65-nm manufacturing in less than three years, with likely extensions to 45 nm and perhaps 32 nm. Many major chipmakers have already inserted IML into their own roadmaps.

Providing solutions in the roadmap is quite tempting. But in some cases, particularly in the early stages, it may be counterproductive. The proper approach would be to avoid restricting or constricting innovation and let the market forces provide the suitable practical answers. When needed, the industry consensus can be achieved outside of the roadmap itself through various international symposia, research centers, and industry consortia and groups, such as the meetings sponsored by Sematech or industry-sponsored academic research centers. Moreover, individual companies or alliances and partnerships can effectively promote certain ideas, perspectives, and solutions.

The technology perspective of the roadmap needs to be complemented by a suitable business perspective from suppliers, chip manufacturers, and systems providers. This would result in a practical, realistic, and properly balanced roadmap, which would better address the issues of timing, actual end-market needs, applications, and economics. One must be extremely cautious not to prematurely endorse particular solutions in the roadmap, especially when a high level of uncertainty or insertion risk exists. Potential solutions, where needed, can be listed as an analysis of the existing situation in such a way that supports innovation and encourages out-of-the-box thinking.