Drug Discovery

From Millions of Molecules to One Medicine

A simplified illustration of how AI narrows an enormous search space down to a handful of viable drug candidates.

Candidate Molecules Screened0

AI-Filtered Candidates0

Predicted Viable Compounds0

Laboratory-Tested0

Approved Medicine0

Timeline Reduction

Traditional Discovery Pipeline13 years

AI-Accelerated Pipeline4 years

Illustrative figures for educational purposes — actual timelines vary widely by therapeutic area and regulatory pathway.

Real-World Example — Not Illustrative

Rentosertib (INS018_055) — Insilico Medicine

Rentosertib is a small-molecule drug for idiopathic pulmonary fibrosis (IPF) whose biological target (TNIK) and molecular structure were both identified using Insilico Medicine's generative AI platform, Pharma.AI. The program went from project start to a nominated drug candidate in roughly 18 months, and reached Phase 1 human trials in under 30 months — around half the typical timeline for this stage of drug development. Phase IIa results, published in Nature Medicine in June 2025, showed a dose-dependent improvement in lung function versus placebo, and the program advanced into Phase III trials in July 2026.

It is widely cited as the first case of a drug candidate whose target discovery, molecular design, and clinical validation were all AI-driven, reaching this stage of development.

Verify — Nature Medicine, Phase IIa results (2025)