Aim Since 2014, the Pepfar has initiated its phase III called ‘Pepfar 3.0’ (2014- now), focusing on ‘Sustainable Control of the Epidemic’, to reach the Joint United Nations Programme on HIV/AIDS’ (UNAIDS) 90-90-90 goal. In this strategic plan, the Pepfar states that it is “pivoting to a data-driven approach that strategically targets geographic areas and populations where [they could] achieve the most impact for [their] investments”. In practice, how were these new targeted HIV screening strategies developed and what were the challenges encountered? Method/Issue We conducted a qualitative study in Côte d’Ivoire in 2015-2018, a country where the HIV program is mainly funded by the Pepfar (73% in 2018). In-depth interviews were conducted with stakeholders in the AIDS public response: CDC/Pepfar (3), Ministry of Health (3), intermediary NGOs (7); and we observed public meetings (nb). In addition, the grey literature of Pepfar (Country Operational Plan – COP, https://www.pepfar.gov/) was reviewed to describe the evolution of Pepfar’s HIV testing strategies from COP 14 to COP 17 (October 2014 – September 2018). Results/Comments Since the COP 14, Pepfar’s HIV testing strategies have been characterized by significant variations: (i) in the targets of the number of people and of HIV positive people to be screened, divided by 2 between COP 14 and COP 15 and multiplied by 4.5 between COP 15 and COP 17; (ii) in the targeted geographical areas, from a regional breakdown to an identification of health districts at high and low impact; and (iii) in the targeted sub-groups, from a focus on key populations to a broader definition of priority populations, including men over 25 years old. A shift was observed in the definition of testing targets, with less room dedicated to programmatic data and feedbacks from field actors and an increasing focus on the use of modelling work to estimate and disaggregate the targets by geographical units and sub-population (even if available data at these fine levels was limited and uncertain); increasingly leading to gaps between targets and results. While the aim of COP 14 and COP 15 seemed to be the improvement of testing efficacy in general and testing yield (i.e. testing positivity rate) in particular, COP 16 and 17 have given priority to fill the gaps in terms of the first 90 (i.e. reducing proportion of PLHIV being undiagnosed). Discussion This study highlights how the Pepfar approach has been changing over the last 5 years as a result of the tension between improving HIV testing yield and achieving the first 90 (90% of PLHIV being diagnosed) in the context of a mixed epidemic, combined with a context of limited resources implying to improve the “value for money” of policies. This study also underlines the limited possibility of very rapidly evolving strategies by the donor - emphasised by the annual system of COPs - in a context where actors have different adaptive capacities. The emphasis on targets to reach tends to underestimate the political and social modalities of implementation of strategies.