Insilico Medicine states that INS018_055, its candidate for idiopathic pulmonary fibrosis, advanced from novel target discovery to a Phase I clinical trial in under 30 months - roughly half the time the company estimates the same process would take with traditional drug discovery. The target, the kinase TNIK, was nominated by Insilico’s PandaOmics platform, and the molecule was designed by its Chemistry42 generative-chemistry engine.
The compressed timeline is the central claim Insilico makes for AI drug discovery: that machine learning can dramatically shorten the slow front end of the pipeline, where scientists choose a biological target and invent a molecule to act on it. Speed at that stage, however, does not change the odds in clinical testing - most candidates that reach human trials still fail, regardless of how they were discovered.