The first drug whose target and molecule were both discovered by AI reaches the final stage of human trials
Insilico Medicine's rentosertib is now being tested in about 320 patients with idiopathic pulmonary fibrosis, a serious disease that scars the lungs. In the previous phase, the highest dose improved lung capacity while it worsened in the placebo group.
Until now, artificial intelligence has mostly sped up parts of drug development. Rentosertib is different: AI at Insilico Medicine chose which protein to target (one called TNIK, never before used against this disease) and also designed the molecule to hit it. In July it began its phase III trial, the last step before seeking approval. According to the company, it is the first drug born this way to get there.
What it’s for
It is aimed at idiopathic pulmonary fibrosis, a disease in which the lungs fill with scar tissue and lose the ability to breathe. Current treatments only slow it down.
What we know
- In phase 2a, published in Nature Medicine in 2025, 71 patients at 22 hospitals in China took it for 12 weeks. It was safe and well tolerated.
- On the highest dose, lung capacity improved by an average of 98 ml, while in the placebo group it fell by 20 ml.
- Phase III involves about 320 patients at 47 hospitals in China over 52 weeks. It will measure whether the drug slows the loss of lung capacity and delays disease progression.
What’s still missing
The drug is not approved anywhere. Phase 2a was small and short, and many treatments that look promising at that stage later fail. The figures come from the company. Even so, it is the most advanced proof that AI can find new targets and design medicines that reach patients.
