AI-Discovered Drug Reaches Phase 3 Trial, Shows Aging Reversal

Insilico Medicine has become the first company to advance an AI-discovered drug to Phase 3 clinical trials, marking a watershed moment for artificial intelligence in pharmaceutical development. The company's lead candidate, rentosertib, is being tested in China for idiopathic pulmonary fibrosis, a fatal lung-scarring disease.

What sets rentosertib apart is that AI was used both to identify the disease target and to design the molecule itself. The first patient in the Phase 3 trial received treatment on September 10, according to details first reported by Fortune.

The drug has generated additional excitement after Insilico published findings in Nature Biotechnology showing blood samples from 42 patients exhibited signs of reversed biological age. Company executives cautioned the effect is small and may not persist long-term, but it represents one of the first clinical hints that AI-invented drugs might influence aging processes.

Why it matters

Insilico's progress tests three critical assumptions simultaneously: whether AI can produce drugs that survive rigorous Phase 3 trials, whether AI drug discovery can be commercially viable, and whether China has become a viable alternative to Western markets for biotech innovation. The outcome will shape how pharmaceutical companies worldwide approach both AI adoption and geographic strategy.

China as innovation hub

Co-CEO Alex Zhavoronkov attributes the company's rapid progress to Chinese regulatory sophistication. After 2015 reforms that expanded the reviewer corps tenfold, average drug review times in China dropped from 900 days to 300 by 2019. Zhavoronkov told Fortune that Chinese regulators "knew not only my drug, but all the other drugs" during the approval process.

Insilico is pursuing what it calls a "China-for-China" strategy, using Chinese resources and talent to develop products for the domestic market. The company has secured deals with Chinese pharmaceutical firms including a $120 million partnership with Qilu Pharmaceutical, alongside larger Western partnerships with Eli Lilly ($2.75 billion) and SK Biopharmaceuticals ($2.5 billion).

Profitability milestone

The startup reported $35.5 million in net profit for the first half of 2026, making it one of the few AI drug discovery companies to reach profitability. Zhavoronkov, who took over business operations last year, restructured the sales approach by putting scientists in charge of business development and building an automated inquiry portal.

Beyond rentosertib, Insilico has one drug in Phase 2 trials for ulcerative colitis and eight more in Phase 1 trials, mostly targeting various cancers.

Expanding the AI platform

In January, Insilico launched MMAI Gym, which fine-tunes large language models from OpenAI, Anthropic, and Alibaba on molecular language spanning medicinal chemistry to clinical development. The company also released LongevityBench on September 17, a benchmark for testing how well AI models understand aging biology. In a study featured on the cover of Cell, Insilico reported that smaller, aging-trained models outperformed several frontier systems on this metric.

Zhavoronkov acknowledged the company made overly ambitious claims in its early years, particularly his regretted assertion that "AI was going to replace medicinal chemists." He described the 2014 cohort of AI drug discovery startups as "now a graveyard," with Insilico among the few survivors.

The details in this report were first published by Fortune.