JURA builds foundation models for therapeutic design that generate the proprietary training data they need to continuously improve, closing the loop between AI and the wet lab at a scale no public dataset can match.
What is the problem?
AI-driven therapeutic discovery is constrained by a fundamental data problem. Public databases lack the functional annotations that matter: binding specificity, off-target activity, and developability. Conventional lab workflows cannot generate experimental data fast enough to meaningfully train frontier models, and the result is AI that fails for the hardest targets and newest modalities.
What is their solution?
JURA builds the infrastructure for frontier models to self-improve. JURA's platform VISTA screens AI-designed therapeutic candidates at a scale that even automated high-throughput systems cannot reach by integrating AI directly with DNA synthesis chemistry, enabling physical construction and functional screening of libraries spanning over 10¹⁶ AI-designed candidate sequences. Each experimental cycle feeds back into JURA's foundation models, compounding into proprietary functional data and continuously improving candidate quality across antibodies and cell therapies, regulatory promoters, and emerging modalities.
