<p>This study investigates whether equity concentration exerts a nonlinear threshold effect on the relationship between private capital participation and bank risk‑taking. Using panel data from 171 Chinese commercial banks over 2013–2022 and employing threshold regression models, we find strong evidence of a double threshold. Private capital significantly reduces non‑performing loans only when ownership concentration lies within moderate ranges of 29.75–35.68% for the top‑five shareholder ratio (CR5) and 36.60–56.79% for the top‑ten ratio (CR10). The effect becomes insignificant at both low and high concentration levels. These findings are robust to alternative risk measures (Z‑score), different sample periods (2013–2019 and 2014–2021), and a control‑function 2SLS approach that addresses endogeneity. Heterogeneity analyses reveal important boundary conditions: the threshold effect is present in small banks but absent in large banks; it is also present in private‑owned banks (using CR10) but absent in state‑owned banks. Our study contributes to the corporate governance and banking literature by demonstrating that the governance role of private capital is inherently nonlinear and regime‑dependent, challenging linear assumptions in prior research. From a policy perspective, the results suggest that banks introducing private capital should maintain equity concentration within a moderate range to maximise monitoring efficiency, and that size‑differentiated and ownership‑differentiated regulatory guidelines may be more effective than one‑size‑fits‑all policies.</p>