VLSI: the next ten years

An examination is made of the various scenarios for the semiconductor industry over the next ten years. The following are anticipated. (1) Scaling will continue through the next decade, providing density improvements. (2) Significant structural material changes will be needed in building sub-half-micron ICs in order to continue performance and reliability improvements, especially if monolithic integration of multifunctions is envisioned (linear, power, nonvolatile functions with digital). (3) The cost of scaled silicon is increasing rapidly. Unless there is a fundamental change in this trend, the cost will be prohibitive. (4) A 100*-1000* improvement in design productivity is needed in the next decade to continue to cost-effectively design applications that utilize state-of-the-art technology. This may need a paradigm shift in the approach to IC design. (5) As scaled silicon becomes more expensive, advances in packaging, perhaps using older generation silicon technology, will provide the system-level cost improvement that will keep the application revolution going for the next decade.<<ETX>>