This paper empirically examines the impact of academic research on high-tech manufacturing growth of 28 Organisation for Economic Co-operation and Development (OECD) and emerging countries over the 1991-2005 period. A standard research and development (R&D) expenditure based measure is found to be too general to capture the input in high-tech research. To overcome this problem, a novel proxy for high-tech research investment--the supercomputer capacity--is proposed. Empirical evidence strongly supports this choice of variable. It is also found that academic R&D exerts a larger growth effect on high-tech output than its industry and government counterparts, but only the impact differential between academic and government R&D is statistically significant.