Insilico Medicine announced on September 10, 2026 that it had dosed the first patient in GENESIS-IPF-3, a Phase III trial of rentosertib, which the company describes as the world’s first generative-AI-discovered drug to reach a pivotal Phase III trial. The drug targets idiopathic pulmonary fibrosis, a progressive and ultimately fatal scarring of the lungs with no cure and few effective treatments.
What makes rentosertib an AI-native drug
Rentosertib — previously known by its development codes ISM001-055 and INS018_055 — is a small-molecule inhibitor of a protein called TNIK, discovered and designed using Insilico’s Pharma.AI platform. Professor Zuojun Xu of Peking Union Medical College Hospital, the trial’s leading principal investigator, said TNIK had never previously been linked to fibrosis in prior research: “TNIK, the target driven by AI, had never previously been linked to fibrosis. This perhaps indicates that AI is carving out a path distinct from traditional research paradigms.” That claim — that the AI system identified a viable drug target humans hadn’t previously connected to the disease — is the central argument AI-drug-discovery companies have made for years to justify the approach, and rentosertib is one of the furthest-advanced test cases of whether that argument holds up in the clinic rather than just in a research paper.
The trial itself
GENESIS-IPF-3 is a prospective, randomized, double-blind, placebo-controlled trial enrolling 320 patients with idiopathic pulmonary fibrosis across 47 centers in China, according to Insilico’s announcement. Patients will receive once-daily oral rentosertib over 52 weeks, with the primary endpoint measuring the annual rate of decline in forced vital capacity — a standard lung-function measure used across IPF trials because the disease’s defining feature is a steady, irreversible loss of lung capacity over time. A key secondary endpoint tracks time to first disease-progression event. The trial is led by Professor Xu alongside co-leading investigators Academician Nanshan Zhong of the Chinese Academy of Engineering and President Chang Chen of Shanghai Pulmonary Hospital, giving the program significant institutional weight within China’s pulmonology research establishment.
What the earlier data showed
The Phase III launch follows a Phase IIa trial whose results were published in Nature Medicine, showing dose-dependent improvements in lung function over 12 weeks among IPF patients on the highest dose arm. That publication was widely cited in 2025 and 2026 as the industry’s first proof-of-concept clinical validation that a drug candidate discovered primarily through generative AI could show real efficacy signals in human patients, rather than only in preclinical models — a distinction that matters because AI drug-discovery companies have faced years of skepticism about whether computationally generated molecules would actually perform in the clinic the way animal and cell-based models predicted.
Why IPF is a meaningful test case
Idiopathic pulmonary fibrosis carries a prognosis often compared to aggressive cancers, with median survival historically cited in the range of three to five years after diagnosis, and the two currently approved antifibrotic drugs, pirfenidone and nintedanib, slow disease progression without reversing it and come with significant tolerability issues that cause many patients to discontinue treatment. A new mechanism of action — targeting TNIK rather than the pathways existing drugs address — would matter clinically even if rentosertib’s benefit turns out to be modest, simply because the treatment landscape for IPF has been stagnant for so long.
The skeptical read on AI drug discovery broadly
Rentosertib’s advance to Phase III arrives alongside a broader industry reckoning about whether AI-discovered drugs can actually deliver at scale: multiple companies have candidates in various stages of clinical testing, but Phase III is historically where drug candidates fail most often and most expensively, regardless of how they were discovered, because larger and longer trials expose safety and efficacy problems that smaller early studies can miss. Carol Satler, commenting on the program, said she believes “rentosertib will be transformative to patient care in IPF, and AI has enabled this breakthrough with higher speed and reduced cost” — but that view remains a prediction until the 52-week Phase III data reads out, not yet a confirmed outcome.
What’s next
With patient dosing now underway across 47 Chinese trial centers, the GENESIS-IPF-3 results — expected after the 52-week treatment period plus follow-up and analysis — will be one of the most closely watched readouts in the AI-drug-discovery field, since a positive Phase III result would be the strongest evidence yet that generative-AI-designed molecules can clear the same late-stage bar as conventionally discovered drugs, while a disappointing result would sharpen the argument that AI’s contribution so far has been mostly to speed and cost of early discovery rather than to the fundamental odds of clinical success.
Photo: PublicDomainPictures / PIXABAY via Pixabay
