Sub-Saharan Africa (SSA) has a scarcity of research on autism spectrum disorder (ASD) and available early interventions, as most of what is known about the disorder is from highincome countries. Early detection and intervention methods were found to have positive effects on developmental delays and to alleviate symptom severity in children with ASD or at risk of it. There is a need for scalable interventions in low-resource settings, which are characterised by a lack of highly-trained specialists, infrastructure and funding. This study explored the barriers and facilitators to implementing and sustaining a caregiver-coaching ASD early intervention, informed by the principles of the Early Start Denver Model (ESDM), adapted for South Africa and for delivery by non-specialists. The study also identified some changes that could be made to improve intervention adoption and sustainability. Nine multilevel stakeholders involved in the implementation of the caregiver-coaching intervention were purposively sampled, individual in-depth interviews were conducted, transcribed verbatim and thematically analysed. Major implementation facilitators included: ECD worker baseline ASD knowledge and experience; skills gained from the training received and coaching; clear in-session caregiver-coaching structure; value of strong team relationships; clear video illustration of intervention concepts; and the mastery and generalisation of skills by the ECD workers, ECD supervisors and caregivers. Implementation barriers included: the complexity of the intervention and coaching concepts; misalignment of ECD teacher training with the caregiver-coaching approach; logistical challenges; and mismatch of the video content with the South African context. Facilitators to sustain the intervention included: child outcomes; caregiver ‘buy-in;’ and competence; and the need for ongoing live supervision. Barriers to sustaining the intervention included: structural issues of poverty, transportation and unemployment. Positive child and caregiver outcomes could be offset by larger contextual and system-level issues such as poverty and the need for ongoing support, supervision and local coaching materials in South African languages. The results will inform tailoring of the intervention training and supervision approach for a larger pilot study.