<p>This paper critically examines the implementation of a novel affirmative action scheme, the Supernumerary Seats Scheme (SSS), to raise the proportion of female students in undergraduate degree programs to 20% in the Indian Institutes of Technology (IITs). SSS led to the creation of additional seats for females and bifurcated all available seats into two categories: “female-only” and “gender-neutral.” Our analysis of official data and documents shows that SSS has been remarkably successful in meeting its stated objective in a short span of time, albeit with regional variations. We also find that SSS is nudging high-ranking females to apply in greater numbers to highly competitive and historically male-dominated branches in IITs like computer science, electrical and mechanical engineering. There is however an in-built tension in the allocation design of SSS wherein female candidates are unable to compete for gender-neutral seats prior to being considered for female-only seats, in a stark departure from traditional caste-based reservation. In-built heterogeneity also exists in the extent of affirmative action the scheme entails for different branch-institute pairs. Overall, our study presents the first rigorous evaluation of this landmark policy and in doing so highlights the urgent need for improved communication on the design, implementation and impact of the scheme.</p>