Yokohama, Japan -- Asian South Pacific Design Automation Conference -- February 1, 2001
Semiconductors will be to the 21st century and the transition to an information society as steel was to the 20th century and the industrial revolution. That's the road map that was laid out by Hajime Sasaki, chairman of NEC Corp., in his keynote speech at the Asian South Pacific Design Automation Conference this week in Yokohama, Japan.
Sasaki pointed out that the semiconductor industry has changed in only the last 50 years from an emerging technology to one that itself enables new technologies. And the main benefactor of this metamorphosis as well as the current drivers for semiconductors are the information technologies, which are responsible for the increasing ability to digitize information and disperse that content. The creation of the Internet enabled a new set of structures and the idea of a self-developing and ownerless information system. The next step for information technologies, said Sasaki, will be the mobile revolution, where information becomes pervasive and, at the same time, personal. The ubiquitous nature of information and its access humanizes the information, he said.
The semiconductor industry's evolution in some sense parallels the changes in computers and the Internet, said Sasaki. In the late 1960s computers were mainframes, and the first implementation of the Internet started as the Arpanet, where military and academic researchers shared and obtained information through the first Internet application, e-mail. The development and deployment of e-mail application has become so widespread that information transfer and interconnectivity are becoming universal and expected everywhere.
Sasaki said that the second wave of computers and the Internet started in the mid 1980s with personal computers and the beginning of the World Wide Web. In the late 1990s this generated a new way of transacting business, e-business. The ability to do business without a local physical presence has led to great flexibility and a concomitant reduction in costs, changes that will affect the roots of society as more companies provide information and services through e-mail and home pages instead of through buildings and people.
The third wave, he continued, will be driven by Internet appliances and many e-commerce applications, and the changes in business models will be reflected within societies. Although everything will have computers associated with them, the computers will be as different from PCs as the PCs were from mainframes. Many of the applications will be associated with mobile handsets, and if Japan is an example of adoption rates then the Internet's penetration will see extremely high volumes of wireless connectivity in many countries. In Japan, more than half the people have mobile phones and more than 25 million have signed up for DoCoMo, the wireless Internet service, in the past two years.
These mobile terminals are technology drivers, Sasaki pointed out, with requirements for high performance and low power so stringent that they can be achieved only with higher levels of integration. The increases in bandwidth associated with the latest applications (like music and picture phones) lead toward complete revamping of both infrastructure and terminals, which will spur the semiconductor industry to grow as much as 12.4 percent per year until 2006.
Some of the major trends driving the industry will derive from the increases in home networking, office functions and mobile applications, said Sasaki. Home applications will need system-on-chip levels of integration to meet cost and power requirements; the upshot will be that more than half of large ICs will be SoC devices by 2006. This change in design requirements will create a paradigm shift that will define a new way of thinking. Semiconductor vendors will have to stop thinking of supplying hardware and move toward system applications, with the main factors reduced time to market and a need for low-cost production despite a greater diversity of processes and high added value.
Sasaki predicted that the move to SoC will require a shift in focus from costs per transistor to overall speed, throughput and functionality. More silicon intellectual property (IP) will be needed to fill the chips and provide the required functionality in the time available. The IP will be classified three ways: killer IP, such as a CPU; star IP, such as MPEG decoders; and commodity IP, such as I/O and standard interfaces. The value of the various forms of IP will change, with star cores becoming commodities over time. Sasaki warned that companies will not be able to ignore the star and commodity cores, but will have to change to have better relations with IP developers. The future SoCs will have a combination of best-in-class IP, but no single company can create all the cores itself.
Because of the changes in relations between core developers and silicon vendors, more companies will have to work together on many aspects of their businesses, said Sasaki. Successful companies will focus on their core competencies while cooperating with others in industrial and academic consortia to solve the larger technical issues. There they will have to balance the costs of new fabs with the need to have good cash flow, end the "all-in-house" philosophy to develop cross-border alliances and still overcome the increasing worldwide competition.
Sasaki said that the methods and structures in Silicon Valley, where large universities work with semiconductor companies, coupled with ready sources of money and technical expertise and availability of a good technical infrastructure, is not equaled anyplace in the world. But what are not fundamental parts of the facilities and location are the speed and knowledge within the companies-that, said Sasaki, is a function of the diversity of people in the industry. In addition, strategic planning functions must define clear and simple targets. The increase in collaboration leads to centers of excellence. One key factor is the need to make the business a part of the R&D function.
Semiconductor makers will have to meet many challenges, said Sasaki. For example, the requirements and characteristics of e-commerce can conflict with the need to maintain a clean and healthy environment. Moreover, e-commerce can help reduce the amount of energy needed to create a million dollars of gross domestic product, but can increase the divide between the haves and have-nots. The new technologies such as biotechnology will not be value free, but will require tradeoffs in economies, societies and ethics.
Sasaki said that the information technologies can provide improvements in lifestyles and societies, but an uncontrolled information technology is at the mercy of human desires, which can lead to what he called the total personalization and isolation of the individual. And boundless desires will hit barriers because of environmental limitations, he predicted. So a more human set of information technologies will help to overcome the more difficult obstacles and enhance the feelings of unity within the world.
Finally, said Sasaki, an essential condition for all successful semiconductor companies in the next century will be to have the wisdom to continue activities with passion, coupled with the courage to follow their wisdom and implement the desired tasks or to speak out against the unwise choices of others. And this must be tied to generosity to consider others in a global perspective.