Decreasing research and development (R&D) can impair the ability of firms to remain innovative in the long run. CEOs have been accused of curtailing R&D investments as they approach expected retirement, yet received findings on R&D investment behaviors of late-career CEOs are mixed. We argue that one reason for these inconsistent findings could be that traditional approaches overlook the fact that CEOs are not isolated agents in making R&D decisions. We build on the premise that CEOs interact with their top management team (TMT) when shaping R&D strategy and advance a contextualized view of CEO dispositions in their late career stages as being constrained or enabled by their TMT. We hypothesize that some TMT attributes (e.g., tenure and age) may amplify, whereas others (e.g., functional experience and education) may mitigate inclinations to reduce R&D. Our findings, based on a longitudinal sample of 100 US manufacturing firms from 1998 to 2008, provide nuanced insights into how different TMT characteristics influence CEO-TMT dynamics, with TMT age and TMT tenure playing particularly pronounced roles. We discuss implications of our CEO-TMT interface approach for theory and practice.