On July 7, 2026, Insilico Medicine announced it had launched a Phase III clinical trial for rentosertib, an oral small-molecule drug the company says was substantially designed using generative AI, for the treatment of idiopathic pulmonary fibrosis, a progressive and often fatal scarring disease of the lungs with no cure. The trial matters beyond one company’s pipeline: as of this year, not a single AI-discovered drug holds full FDA approval anywhere in the world, and Insilico’s program is now among the furthest along in trying to change that.
The Drug and the Disease It Targets
Rentosertib is designed as a first-in-class inhibitor of TNIK, a kinase enzyme implicated in the fibrotic processes that scar lung tissue in idiopathic pulmonary fibrosis, or IPF. IPF gradually stiffens the lungs, making breathing progressively harder, and historically carries a prognosis often compared to aggressive cancers, with few treatment options that meaningfully slow disease progression rather than just managing symptoms. Insilico describes rentosertib as its lead asset precisely because IPF represents the kind of high-unmet-need indication where a genuinely novel mechanism could have outsized clinical impact.
Inside the Phase III Trial Design
The newly launched study is a randomized, double-blind, placebo-controlled, parallel-group trial expected to enroll 320 patients with IPF across 47 clinical centers in China, evaluating a once-daily oral dose of rentosertib over 52 weeks. The trial is registered under identifiers CTR20262475 and NCT07687459. A year-long placebo-controlled design for a progressive lung disease is a significant commitment for both patients and investigators, reflecting the kind of rigorous, lengthy evaluation regulators expect before a novel mechanism can be considered for approval.
The Regulatory Trail That Got It Here
Rentosertib’s path to Phase III did not happen overnight. The U.S. FDA granted it Orphan Drug Designation for IPF back in February 2023, a status meant to encourage development of treatments for rare diseases through incentives like extended market exclusivity. In May 2025, China’s Center for Drug Evaluation added rentosertib to its Breakthrough Therapy Designation list for IPF, a designation intended to speed regulatory review for therapies showing meaningful early promise over existing options. Those two designations, from regulators in different countries, gave the program credibility before it reached the far more expensive and consequential Phase III stage.
Why the AI Origin Story Matters to the Industry
Insilico has built its identity around using generative AI models to both identify novel drug targets like TNIK and generate candidate molecule structures, compressing a discovery process that traditionally takes years into a fraction of the time. Industry trackers estimate that roughly 117 AI-enabled therapeutic assets across 63 companies have entered human clinical trials so far, but only about 8 of those, roughly 7 percent, have completed Phase 2, and none have reached full approval. Rentosertib’s advance to Phase III places it in a very small group of AI-originated drug candidates that have cleared that bar, making it a closely watched bellwether for whether the AI drug discovery industry’s multibillion-dollar investment thesis can produce an actual approved medicine.
Optimists and Skeptics Are Both Watching Closely
Backers of AI drug discovery argue that rentosertib’s progress validates the core premise of the field: that machine learning models can identify viable disease targets and molecules faster and cheaper than traditional discovery pipelines, and that regulatory designations from both the FDA and Chinese authorities show outside experts already see credible clinical promise in the molecule. Critics of the sector’s hype point to the wider numbers as a caution, noting that with $8.9 billion in cumulative investment into AI drug discovery and still zero FDA approvals as of this year, one promising Phase III readout would not prove the model works broadly, while a failure would reinforce doubts that AI-generated candidates perform any differently in the clinic than those found through conventional methods.
What Happens Next
The 52-week trial duration means a primary readout is unlikely before mid-to-late 2027, so the industry’s attention in the near term will focus on enrollment progress across the 47 participating centers and whether Insilico pursues parallel regulatory pathways outside China, including a potential future FDA submission given the existing Orphan Drug Designation. Whatever the outcome, rentosertib’s Phase III launch is likely to be cited for years as one of the clearest real-world tests of whether AI-driven drug discovery can convert its computational promise into a therapy that reaches patients.
