Abstract:Under the circumstances of profound shifts unseen in a century and a "dual circulation" development pattern in the post-epidemic era, R&D and innovation performance of sci-tech innovation enterprises should be paid more attention from academic and business circles both.Based on the panel data of the STAR market listed companies from 2017 to 2019, the fixed-effect panel threshold model is constructed to test empirically the nonlinear relationship between R&D intensity and enterprise growth in the current and first lag phase. The results show that: (1) there is an inverted V-shaped nonlinear relationship between R&D intensity and current enterprise growth, and the optimal R&D intensity is or less than 4.42%; (2) there is an lag-effect on enterprise growth from R&D and innovation input,and only when R&D intensity more than 6.83%, R&D input promote enterprise growth in the next phase,and exists U-shaped nonlinear relationship between R&D intensity and enterprise growth in the first lag phase; (3) in the optimal R&D intensity span respectively,the promoted performance of R&D input on the current enterprise growth is better than the next phase. Results of this paper can directly assist sci-tech innovation enterprises to define the optimal range of R&D investment, rationally balance short-term risks and long-term benefits, and finally reach the optimal R&D intensity.