Two multi‑partner consortia won ARPA‑H awards to build next‑generation drug safety models that reduce reliance on animal testing. Inductive Bio leads a project (DATAMAP) with up to $21 million to develop AI models trained on organoids, ex‑vivo human tissues and microphysiological systems to predict drug‑induced liver injury and cardiotoxicity. The consortium includes Amgen, Baylor College of Medicine and Cincinnati Children’s. Separately, a consortium led by Ginkgo Bioworks and Deep Origin secured up to $31.7 million from ARPA‑H to create PREDICTS, a program to generate structured datasets—DRUG‑seq, cell painting and perturbation profiles—to train computational models predicting human toxicity. Both programs will engage FDA for potential regulatory qualification and focus on translational endpoints. The awards underscore federal momentum to build mechanistic, human‑biologic datasets paired with AI to lower late‑stage safety failures and accelerate safer therapeutic development.