Summer 1999 Newsletter

Changing Times
Challenges for the University and Industry

Dr. Richard Dasher, Executive Director

It is no secret that the semiconductor and electronics industries are experiencing profound changes. These include (a) aggressive restructuring via mergers, splits, and new partnerships, (b) greater reliance on outsourcing, (c) commoditization of many chip prices, and (d) lower barriers to entry into the industry for start-ups and other "new players." These trends present great challenges to the university as well as to industry. Mergers, partnerships, and splits are redrawing the lines of industry competition at a frenetic pace. Securities Data Company reports that 1998 saw the most mergers ever in the semiconductor industry, 77 deals with a total value of $6.2 billion.

These numbers were up from 43 M&As in 1997, which had a value of $5.0 billion.1 Although the number of deals is expected to be fewer in 1999, the value of semiconductor industry mergers this year is probably already above the $5 billion mark as of 8/15/99, thanks to a few very large deals.

Chip companies are also aggressively creating cross-firm alliances and other partnerships that go much deeper than traditional marketing and distribution arrangements. For example, this spring Toshiba and Sony rolled out a jointly developed graphics processing chip for the next generation Playstation¨. This project is representative of the partnering that is going on even between companies that are fierce competitors in other markets. Some alliances are yielding basic changes in business models. TSMC and Artisan recently made news with an agreement that delays royalty payments for access to design libraries until after chips are manufactured.

At the same time, one also sees the opposite trend, in which vertically integrated companies are spinning off divisions as separate corporate entities. Several of these splits directly involve CIS partner companies. In place of Siemens and Rockwell, we now have the "new" members Infineon and Conexant, respectively. In the next few months, we will see another major instance of mitosis as Agilent Technologies emerges alongside Hewlett-Packard.

The spin-offs are not simply aimed at narrowing business focus to some core competency; nor are the mergers merely for vertical integration. The new spin-off companies aim to exploit greater freedom to reshape their business models, including their customer bases, products/services, and distribution strategies. Mergers these days appear to be for the purpose of bolstering technical strength in new applications areas or to obtain immediate presence in new markets. The upshot is that the battle-lines of competition are constantly changing location at a rapid pace.

Along with partnerships and alliances, one finds an increasing reliance on outsourcing, especially of chip manufacturing. Motorola reportedly plans to increase its outsourcing of manufacturing from 6% of total output in 1998 to 50% by 2002.2 The "IP business" may still be in its infancy, but already companies cannot ignore the option of outsourcing many types of chip design in order to meet short time-to-market requirements.

In turn, time-to-market pressures are compounded by the commoditization of many chip prices, e.g. DRAMs and CPU processors for PCs, and this commoditization will likely be a feature of major new chip markets as well. Currently, the fastest growth in the chip industry is in communications applications, both Internet infrastructure and consumer applications (cell phones), which are growing at about 30% per year. Beyond this wave, analysts expect huge volumes in a wide range of smart consumer appliances, including everything from sub-$500 PCs and digital televisions to automobiles and microwave ovens.3 All of these new markets demand razor-thin profit margins as well as high reliability and performance improvements in processing speed and power consumption.

Nevertheless, barriers to entry into the industry are lower than before. Increases in chip-targeted venture financing, as well as the availability of outsourcing and partnerships, have made it ever easier to launch a new idea as a new company. According to the National Venture Capital Association, venture investing in semiconductor and other electronics areas totaled $598 million in the first half of 1999, up from $320 million over the same period in 1998.4 (This does not include communications and computer hardware, which they treat as separate categories.) Of course, one major factor that has enabled start-ups in the semiconductor area is the availability of outsourced manufacturing; apparently, all new chip companies for the last ten years have gone the fabless route.

A well-known result of these trends is increased risk and uncertainty. In many instances, new chip development decisions must be made before the characteristics of the target markets can be fully known. Lower barriers to entry require established companies to keep a close eye out for new competitors as well as potential sources of new technologies.

Less frequently mentioned are the effects of these trends on corporate R&D. Because chip designs are increasingly complex, new product development requires greater effort and sophistication. At the same time, there is an overwhelming pressure to achieve shorter time-to-market. Therefore, companies are under great pressure to focus their research more and more into immediate technical problems (e.g. system-level chip integration) and greater methodological productivity (e.g. design for reuse). Conversely, the stringent economics of low-margin markets put pressure on companies to cut back their research in areas that are not immediately critical. Companies that have outsourced their manufacturing may be forced to scale back their research into process technologies.

Industry will thus have to rely even more on the academic sector for fundamental research into the basic devices, systems, and processes of future generation technologies. However, these profound industrial changes may make it more difficult for universities to hold fast to that mission. With the R&D budgets of industry sponsors everywhere being tightly squeezed by short-term needs, it becomes tempting for university researchers to choose projects more on the basis of their immediate attractiveness to a sponsor than on their long-term potential value to the field. Similarly, the need to produce graduates who are immediately marketable may discourage the risky, fundamentally new approaches that provide the students with the deepest, most lasting insights into the underlying principles of a technology.

Industry-university cooperation works best when the demands of both parties are balanced so as to maximize the strengths of each. The faculty and students associated with CIS have benefited greatly from the understanding and long-term vision of the representatives on our industry Advisory Committee. This partnership has turned out great graduates with strong fundamental knowledge as well as real world savvy. As we all move into an era of new industry dynamics, it is good to remind ourselves of this delicate balance. We will all have to explain the benefits of the partnership more often.

Footnotes

1. Cahners Electronic News, February 1, 1999. http://www.sumnet.com/enews/issue/020199/255mach.html.

2. Blaise Zerega, "Real Men Hire Fabs," The Red Herring, March 1999.

3. See, e.g., Craig Matsumoto, "Chip Recovery Seems Solid," Semiconductor On-line, July 3, 1999. http://news.semiconductoronline.com/wires/19990703-600509166.html. See also Gary Smith, "Post-PC EDA Coming into Focus," EE Times July 3, 1999; viewable at Semiconductor On-line: http://news.semiconductoronline.com/wires/19990703-600509552.html.

4. Source: Thomson Financial Securities Data, News Release July 30, 1999; http://www.nvca.com/80299nr.html