Can Learning Occur in Reverse? Capturing the Mechanisms, Enablers and Barriers of Reverse Knowledge Spillovers in the Indian Context

Despite sustained interest in open innovation, knowledge spillovers within research and development alliances are still predominantly theorised as flowing from large established corporations to smaller firms. This study examines the mechanisms, enablers and barriers that shape reverse knowledge spillovers, the upward flow of knowledge from young, resource‐constrained start‐ups to large, established corporations, within the Indian open innovation ecosystem. We draw on absorptive capacity theory to capture knowledge as a recipient capability and on the relational view to capture the interorganisational conditions of transfer. The study adopts a qualitative, theory building design based on 60 semi‐structured interviews with senior executives from 28 corporations and 32 micro, small and medium start‐ups across five industries, including 18 matched corporate start‐up dyads, analysed through the Gioia methodology and triangulated against archival (secondary) records. We find that reverse spillovers occur through four predominant mechanisms, namely talent mobility, joint development, supplier innovation and pilot deployment, and that the effectiveness of each is contingent on four conditions: absorptive capacity, organisational openness, relational quality and prior ties. We theorise reverse spillover as the joint product of capability and access, realised only when a recipient able to recognise and value externally sourced knowledge is matched by a relationship through which that knowledge can travel. From this we develop four propositions and a two‐by‐two contingency framework. The study contributes original qualitative evidence from an emerging market that knowledge flows in research and development partnerships are bidirectional and multidirectional rather than unidirectional, and that status inversion activates legitimacy discounting and the not invented here disposition as the binding constraints on upward learning.