This study investigates the mechanism by which a novel starch-based admixture named Temperature Rise Inhibitor (TRI) modifying cement hydration and limiting heat evolution. This admixture is designed to avoid thermal cracking of concrete. Depending on the initial state, TRI has two distinct effects: Solid TRI has a "depressing effect" that modifies the maximum heat flow and reduces the 1d cumulative heat by up to 89.4% with less than 0.15% addition. Addition of TRI pre-dissolved in the mixing water has a "retarding effect" with a delay in heat release, but little impact on the shape of main hydration peak. The nucleation and growth of C-S-H was semi-quantified by SEM. It was found that, irrespective of adding mode, TRI has a negligible influence on the growth of C-S-H "needles". Simulations based on the C-S-H "needle model" support the hypothesis that the modification of hydration kinetics can be caused by affecting C-S-H nucleation.