Early oral language interventions boost children’s language skills, yet we know strikingly little about whether these gains endure. The handful of long-term follow-up studies available suggest that even high-quality language interventions show substantial fade-out. This gap in our evidence base has real consequences for families and for policy, especially as demand for language support continues to rise. We suggest that long-term impact might depend on three levels: characteristics of the intervention (e.g. breadth, instructional approach, and fidelity), features of the learning environment (e.g. classroom ethos, continuity of support, and language resources available) and child-specific factors (e.g. children’s cognitive profiles). We call on funders and researchers to prioritise the routine capture of long-term outcomes and to invest in identifying the mechanisms and tools that could drive sustained improvement (e.g. regular booster sessions). We feel this is a critical priority for future research, as it would help us design support that genuinely shifts developmental trajectories.