Flexible options in semiconductor design

Thesis: S.M. in Engineering and Management, Massachusetts Institute of Technology, System Design and Management Program, 2019Cataloged from PDF version of thesis.Includes bibliographical references (pages 79-82).In semiconductor design, system architects must find the right balance between competing goals [1]: 1. Creating customer value with low average development costs to minimize overhead over a range of products. 2. Fulfilling customer demand with the lowest per unit cost to maximize profit. 3. Exceeding customer expectations to increase market opportunity and drive future growth. As the cost for semiconductor development and investment increases, so does the uncertainty of market success for new products. In this risky environment, a flexible design that increases market opportunity is potentially highly valuable. Flexible designs allow system architects to defer decisions about the exact system configuration and functionality which in turn minimizes the downside effects associated with unforeseen changes in market demand and promotes learning opportunities which lower per unit cost over time [2].In semiconductors, monolithic designs built for a single, specific application, often have lower overall manufacturing cost when compared with discrete, general-purpose designs built to handle multiple applications. This thesis focuses on two dimensions of how flexible designs in semiconductors can meet changing requirements: modularity and generality. Modularity is when we can quickly re-configure elements in a system. Generality is when a single element in the system accommodates changing requirements. When carefully applied to semiconductor design, these flexibilities can provide higher overall value across all possible project outcomes when compared with single application, monolithic designs. The thesis presents a case study describing potential applications of flexible design techniques for semiconductor products.A Net Present Value (NPV) analytical model calculated the expected return on investment for flexible designs combining multiple, smaller dies in a System in Package (SiP) integration. Its goal is to show how proper recognition of uncertainty and the time value of money, generally counter-balances, and sometimes overwhelm the economy of scale benefits from monolithic designs. Ultimately, this case study demonstrates the economic potential of flexible designs of semiconductor products.by Matthew D. McShea.S.M. in Engineering and ManagementS.M.inEngineeringandManagement Massachusetts Institute of Technology, System Design and Management Progra