It has been assumed that there is a clear linear causal effect of research and development (R&D) on performance. However, recent empirical studies have shown both a positive relationship, which suggests an increase in R&D leading to improved performance, as well as a negative relationship, which suggests an increase in R&D resulting in poor performance. Therefore, the evidence for this assumption is not as clear-cut as might be expected, despite a number of efforts to validate it. This paper proposes a new unified nonlinear relationship that incorporates both effects (optimal and threshold) into the proposed model. In examining a panel data of high-technology industry, the relationship between R&D and performance could be graphically depicted as an inverse S-shaped nonlinearity. Our empirical findings also confirmed not only an optimal effect that a specific level of R&D corresponded to maximum performance, but also a threshold indicating that a minimum level of R&D is required in order for R&D to be effective.