The AI-Designed Drug That Turned Back Six Clocks

Rentosertib is the first drug whose target was identified by AI and whose molecule was generated by AI to reach late-stage trials. It is being developed for idiopathic pulmonary fibrosis, a fatal lung disease. A paper published in Nature Biotechnology on 7 September 2026 reports something its developers did not set out to test: that patients on the drug looked biologically younger, according to six independent ageing clocks. The claim is carefully hedged in the paper and loudly celebrated by the company, and both readings are worth understanding.

What was measured

The underlying trial, NCT05938920, was a randomised, double-blind, placebo-controlled phase 2a study run in 2023 and 2024 across 21 sites in China: 71 patients randomised to 30 mg once daily, 30 mg twice daily, 60 mg once daily or placebo, for 12 weeks. Its efficacy result was published in Nature Medicine in June 2025. The new paper, by Alex Zhavoronkov, Fedor Galkin, Shan Chen and colleagues, analyses a 42-participant sub-cohort whose serum was measured on 2,841 proteins at baseline and weeks two, four and twelve, then run through six proteomic clocks: ProtAge, two OrganAge variants, PAC, ipfP3GPT and PAOPAC.

The effect, the paper states, was observed most consistently at week four in the 30 mg twice-daily group, by five of the six clocks, with 11 participants per arm at that timepoint and nine in the 60 mg arm. Rentosertib significantly altered the trajectories of 326 proteins across treatment arms against two in placebo; the twice-daily arm had 142 uniquely affected proteins, while 30 mg once daily changed only one. Insilico's release puts the peak effect at roughly three to four years of biological-age reversal and up to six years on one clock.

The caveats, stated plainly

• Sample size: 42 participants, roughly 11 per arm, over 12 weeks.
• Secondary analysis: the trial was designed to measure lung function, not ageing; the clocks were applied afterwards.
• Authorship: Zhavoronkov is Insilico's founder and chief executive and the paper's first author; several co-authors are Insilico employees, as the competing-interests statement records. Named peer reviewers are Tamir Chandra and Samuel Crofts.
• Population: patients with a fatal fibrotic disease, not healthy adults. The New York Times notes the company has not tested anti-ageing effects in healthy people and that approval could be years away.
• Clocks measure predicted age from protein patterns; a drug that changes disease-related proteins will move them, which is not the same as slowing ageing.

The trajectory

The disease result is what carries the programme forward. In the phase 2a trial the 60 mg once-daily arm showed a mean change in forced vital capacity of plus 98.4 mL against minus 20.3 mL for placebo at 12 weeks, according to Insilico. Phase III, GENESIS-IPF-3, was initiated in July 2026; the first patient was dosed at Peking Union Medical College Hospital in the week the paper appeared, with Shanghai Pulmonary Hospital enrolling the same day. Clinical Trial Vanguard reports a design of 320 patients across 47 sites over 52 weeks with the annual rate of lung-function decline as the primary endpoint. Insilico, listed in Hong Kong, reported first-half 2026 revenue of about $106 million and its first profitable half-year.

Ludger Goeminne of Harvard Medical School, a co-author, said the team 'observed significant reductions in predicted biological age across multiple organ-specific clocks' and that pathway analysis showed effects 'far beyond merely reducing fibrosis'. Zhavoronkov's framing was larger still, converting three extra years of lifespan into 'roughly 25 billion life years on the global scale'. The paper's own language is narrower, and it is the paper that will be cited.

Who benefits, who is at risk

Beneficiaries: patients with IPF if phase III confirms the lung result; the ageing-clock field, which now has a trial-embedded method; and Insilico, whose AI discovery platform gains a second published signal. At risk: the credibility of longevity claims if the small, company-led result is oversold before replication.