Insilico Medicine
Last reviewed
Jun 7, 2026
Sources
17 citations
Review status
Source-backed
Revision
v1 · 1,966 words
Improve this article
Add missing citations, update stale details, or suggest a clearer explanation.
Suggest edit
CtrlK
Last reviewed
Jun 7, 2026
Sources
17 citations
Review status
Source-backed
Revision
v1 · 1,966 words
Add missing citations, update stale details, or suggest a clearer explanation.
Insilico Medicine is a clinical-stage biotechnology company that uses generative AI to discover disease targets and design small-molecule drugs. It was founded in 2014 by Alex Zhavoronkov and Alexander Aliper, and it operates research and development sites across several regions, with a registered base in the Cayman Islands and major operations in Hong Kong, mainland China, the United States, and the United Arab Emirates. The company is best known for rentosertib (development code INS018_055, also written ISM001-055), a treatment for idiopathic pulmonary fibrosis that Insilico describes as the first drug to reach Phase II with both an AI-discovered target and an AI-generated molecule. On 30 December 2025 the company listed on the Main Board of the Hong Kong Stock Exchange under the stock code 3696.HK, in what it called the largest Hong Kong biotechnology initial public offering of that year. Insilico is one of the most prominent companies in the field of AI drug discovery, alongside firms such as Recursion, Isomorphic Labs, and Exscientia.
Insilico Medicine was established in 2014. Co-founder and chief executive Alex Zhavoronkov was born in Latvia, holds a master's degree in biotechnology from Johns Hopkins University and a doctorate from Moscow State University, and was active in aging and longevity research before turning to drug discovery. Co-founder Alexander Aliper has served in scientific and leadership roles including president of the company. Early backing came from Deep Knowledge Ventures, and the company's initial focus was on aging biology and biomarkers before it concentrated on a broad therapeutic pipeline.
By the time of its 2025 listing, Insilico ran a dual co-chief-executive structure: Zhavoronkov leads artificial intelligence and corporate strategy, while Feng Ren, a former oncology executive, leads drug discovery and development. The company reported a research staff of roughly 249 people at the time of its IPO, the large majority of whom hold advanced degrees. Insilico also operates a related longevity venture and has historically positioned aging as a long-term scientific theme, though its commercial pipeline centers on fibrosis, oncology, immunology, and other established disease areas.
Insilico's core asset is Pharma.AI, an integrated software platform that the company updates on a roughly twice-yearly cadence. It combines several engines that span target discovery, chemistry, and clinical analysis. Pharma.AI draws on techniques from deep learning and earlier work by the company on generative adversarial networks and reinforcement learning.
The main publicly named components are:
The company has also described newer foundation-model and multi-agent components, marketed under names such as Nach0 and a research-assistant agent. Insilico operates a robotics-equipped automated laboratory to run synthesis and assays, which it uses to feed experimental data back into its models.
An early and widely cited demonstration of the approach was a 2019 paper in Nature Biotechnology, in which a generative reinforcement-learning system the company called GENTRL designed novel inhibitors of the kinase DDR1; the company reported designing candidate molecules in 21 days and completing preclinical validation in mice within about 46 days total. That work was a proof of concept rather than a clinical program, and it should not be conflated with the later rentosertib effort. The 2019 result, like much of the company's published validation, has drawn both interest and scrutiny over how much of the speedup was attributable to the AI itself versus the surrounding experimental infrastructure.
Rentosertib is an oral small-molecule inhibitor of TNIK (TRAF2- and NCK-interacting kinase), developed for idiopathic pulmonary fibrosis, a progressive and often fatal scarring disease of the lungs. Insilico has stated that PandaOmics identified TNIK as the target and that Chemistry42 generated the molecule, and the company reports that the early discovery phase was completed in about 18 months with fewer than 80 molecules synthesized. The drug received a generic name (rentosertib) from the World Health Organization, and the program has advanced through Phase I into Phase II.
The pivotal data point is a randomized Phase 2a trial, named GENESIS-IPF, whose results were published in Nature Medicine on 3 June 2025. The trial enrolled 71 patients across 22 sites in China and ran for 12 weeks. Patients were randomized to placebo, 30 mg once daily, 30 mg twice daily, or 60 mg once daily. The primary endpoint was safety and tolerability, which the company reports was met. As an exploratory efficacy signal, the 60 mg once-daily arm showed a mean improvement in forced vital capacity (FVC) of about +98.4 mL over 12 weeks, compared with a mean decline of about -20.3 mL in the placebo arm.
These efficacy numbers are exploratory and come from small arms, and several caveats are important and were acknowledged by the authors and independent coverage. Seven patients discontinued because of liver injury or dysfunction, several of them while also taking the approved antifibrotic nintedanib, and only 12 of 18 patients (about 67 percent) completed the high-dose arm versus roughly 88 percent on placebo. Quality-of-life measures were described as largely inconclusive owing to high variance between arms. The authors also noted the small per-arm sample size, the geographic and demographic homogeneity of an all-China cohort, and the short follow-up. A Phase 2a trial of this kind establishes early safety and a preliminary signal; it is not designed to confirm clinical benefit, which would require larger and longer Phase 3 testing. Insilico has reported progressing rentosertib toward later-stage development and has separately pursued an inhaled formulation that received clearance to begin a direct-to-lung clinical study.
By its 2025 listing, Insilico reported more than 30 internally generated programs, of which around 10 had received investigational new drug (IND) clearance and roughly seven were in active clinical development. Oncology is the largest therapeutic area. The company has out-licensed several assets and signed discovery collaborations, with management citing aggregate deal values well above one billion US dollars across partnerships, though most of that headline value is contingent on future milestones that may never be paid.
| Program / asset | Target or area | Indication | Status / notable deal |
|---|---|---|---|
| Rentosertib (INS018_055 / ISM001-055) | TNIK inhibitor | Idiopathic pulmonary fibrosis | Phase 2a reported in Nature Medicine, June 2025; inhaled formulation in development |
| ISM3091 | USP1 inhibitor | Oncology (BRCA-mutant and related tumors) | Out-licensed to Exelixis (2023), about 80 million US dollars upfront plus milestones |
| ISM3312 | 3CLpro inhibitor | Coronavirus / antiviral | Clinical-stage oral antiviral |
| MEN2501 | Antibody-drug conjugate target | Oncology | Licensed to the Menarini Group and Stemline; deal value reported above 500 million US dollars; a 5 million US dollar milestone followed first-in-human dosing |
| Sanofi collaboration | Multiple targets | Multiple | Announced 2022; upfront and nomination fees up to about 21.5 million US dollars, with potential milestones cited up to roughly 1.2 billion US dollars plus royalties |
| Qilu Pharmaceutical collaboration | Cardiometabolic | Multiple | Drug-development collaboration reported near 120 million US dollars |
Other named partners over the company's history include Fosun Pharma and Saudi Aramco's venture arm. As with most early-stage biotech licensing, the very large totals quoted for these deals are biocontingent, meaning they depend on hitting development, regulatory, and sales milestones.
Insilico raised more than 400 million US dollars across private rounds from investors including Qiming Venture Partners, Warburg Pincus, and Prosperity7 Ventures (the venture arm of Saudi Aramco), among others. In March 2025 it announced a 110 million US dollar Series E round led by Value Partners Group.
On 30 December 2025 the company listed on the Hong Kong Stock Exchange Main Board under the stock code 3696.HK. The offering priced at HKD 24.05 per share for about 94.69 million shares and raised roughly HKD 2.28 billion, equivalent to about 290 to 293 million US dollars, which the company and trade press described as Hong Kong's largest biotechnology IPO of 2025. It was also reported as the first company to list under the exchange's Chapter 8.05 route for pre-revenue technology companies in this category. Cornerstone investors reportedly included Eli Lilly, Tencent, and Temasek. The Hong Kong public tranche was heavily oversubscribed, and the stock opened above its offer price on debut.
Reception of Insilico has been a mix of enthusiasm and caution. Supporters point to rentosertib as the most concrete clinical evidence to date that an end-to-end AI workflow can produce a drug with a plausible efficacy signal in patients, and to the company's deep partnership roster. Skeptics note that the Phase 2a results rest on small arms in a single country, that the liver-related discontinuations warrant careful monitoring in larger trials, and that AI-originated compounds have so far cleared Phase 2 at roughly the same rate as conventional drugs, with several high-profile AI-led programs from other companies (for example Exscientia's EXS-21546 and BenevolentAI's BEN-2293) having failed. The broader question of how much credit belongs to the AI versus standard medicinal chemistry and trial execution remains debated. Insilico's place in AI in healthcare is therefore widely cited as a leading test case rather than a settled validation, and its long-term significance will depend on Phase 3 outcomes and approvals that had not occurred as of mid-2026.