There are currently approximately 16 million people taking NNRTI-based first-line treatment in low-income and middle-income countries. Most of these patients are using the combination of tenofovir (TDF), lamivudine (3TC) and efavirenz (EFV). The integrase inhibitor dolutegravir (DTG) has shown an improved safety profile compared with EFV in randomized studies. DTG also has a high barrier to development of drug resistance. New co-formulated tablets with TDF/3TC/DTG are being introduced into LMICs, for a median price of $75 per person-year. The prodrug of TDF, tenofovir alafenamide (TAF) is cheaper to manufacture than TDF. A combined pill with TAF/3TC/DTG is also being launched in LMICs, at a similar low price. However, the clinical development programmes for DTG and TAF did not include extensive analysis of several key populations: pregnant women, people with HIV−tuberculosis (TB) coinfection taking rifampicin-based treatment, and treatment-naive or pretreated patients with NRTI drug resistance. An observational study in Botswana has shown an increased risk of neural tube defects when dolutegravir is used in early pregnancy. In LMICs, only 50% of patients have access to regular viral load testing, and genotypic resistance testing is rarely performed. There is currently no clinical data to support switching patients from TDF/3TC/EFV directly to TDF/3TC/DTG if their viral load is either detectable or unknown. New clinical trials and observational studies will be needed to better understand the consequences of this switch of treatment in LMICs. Clinical trials of new antiretrovirals in key populations should be conducted earlier in their development. This will ensure that new treatments can be introduced into LMICs soon after their launch in high-income countries.