Artificial intelligence-enabled drug discovery: A growing but largely early-stage clinical pipeline

Artificial intelligence (AI) is increasingly integrated into pharmaceutical discovery, yet the extent to which AI-enabled assets have translated into clinical-stage programs remains incompletely characterized. We performed a cross-sectional descriptive analysis of therapeutic assets discovered or developed using AI approaches that had entered at least one interventional human clinical trial through March 1, 2026. We identified 129 AI-enabled therapeutic assets across 69 companies. Oncology was the dominant therapeutic area (74 assets, 57.3%), and small molecules accounted for the majority of the pipeline (103 assets, 79.8%). Among these assets, 72 (55.8%) had completed Phase 1 and 24 (18.9%) had completed Phase 2; only five had entered or completed Phase 3. The observed probability of success was 77.8% for Phase 1 and 19.5% for Phase 2. Forty-eight assets (37.2%) were classified as targeting a novel biological target. Among companies with available data, the median time from founding to Phase 1 entry was 6.5 years (IQR 4.0-9.2), median total funding raised was $186.7 million (IQR $48.1-$595.0 million), and the median employee count at clinical entry was 73 (IQR 32-129), with a median operational intensity of $2.72 million per employee. AI-enabled therapeutics are entering clinical testing at increasing scale, but most programs remain early in development, with few having advanced beyond Phase 2. This dataset provides a baseline for future evaluation of whether AI-enabled discovery ultimately improves regulatory approval rates and clinical outcomes.