Summary: Insilico Medicine began trading on Hong Kong’s Main Board on December 30, 2025 after selling 94,690,500 shares for HK$2.277 billion, the year’s largest Hong Kong biotech IPO by proceeds. Retail demand was about 1,427.37 times oversubscribed and the international tranche drew 26.27 times coverage. Its rentosertib IPF candidate reached preclinical candidate status in 18 months after screening 78 molecules.
December 30, 2025: Insilico Medicine started trading on the Hong Kong Stock Exchange’s Main Board under 3696.HK. The generative-AI drug discovery firm sold 94,690,500 shares globally for HK$2.277 billion — the biggest biotech IPO Hong Kong saw that year by money raised, and the first AI-driven biotech to reach the city’s Main Board via the Chapter 8.05 rules. Morgan Stanley, CICC, and GF Securities carried the sponsorship.
Why it matters: a decade of AI drug discovery talk has produced abundant narrative and scarce public-market validation. The debut supplied the missing data point — Hong Kong’s retail offering came in around 1,427.37 times oversubscribed, locking up more than HK$328.349 billion, while the international tranche drew 26.27 times coverage. For the first time at this scale, the secondary market placed genuine demand behind the question “can AI actually produce drugs?”
The Debut: 1,427x Oversubscription
The book deserves a slow read. Retail made up just a tenth of the global offering, yet it pulled in over HK$328 billion — a record for the year among Hong Kong’s non-18A healthcare IPOs. The 26.27x coverage on the international placement likewise topped that year’s non-18A healthcare category.
Fifteen cornerstone investors signed on, with Lilly, Tencent, Temasek, Schroders, UBS Asset Management, Oaktree, E Fund, and Taikang Life Insurance on the list. Both Lilly and Tencent were making their first appearances as biotech cornerstone investors, and Oaktree was returning to Hong Kong biotech for the first time in the year. Big pharma, tech capital, and traditional institutional money sharing one syndicate is, by itself, a vote on the category.
Rentosertib: An 18-Month, 78-Molecule Proof
What the market was actually buying is the pipeline, and Insilico’s flagship exhibit is rentosertib (ISM001-055), a first-in-class candidate aimed at idiopathic pulmonary fibrosis (IPF). The program raced from discovery to preclinical candidate (PCC) status in 18 months, screening just 78 molecules en route. Set that beside the traditional benchmark — roughly 4.5 years and 60 to 200 molecules per program — and the gap is the product.
Clinically, the China Phase IIa wrapped in 2024, and the results that reached Nature Medicine in May 2025 showed favorable safety alongside a dose-dependent efficacy trend; the company positions this as the first clinical proof-of-concept milestone for AI drug discovery, stacking on discovery-phase work published in Nature Biotechnology in March 2024. The wider portfolio holds more than 30 programs, 10 IND clearances, and 7 assets in active clinical development spanning fibrosis, oncology, immunology, cardiometabolic disease, and CNS.
Why Hong Kong
Picking HKEX over Nasdaq says more about the route than the map. Hong Kong’s biotech IPO traffic has flowed through Chapter 18A, the unprofitable-biotech lane, for a decade; Insilico took the standard Chapter 8.05 path, meaning a conventional financial test absorbed an “AI platform plus pipeline” story. The city’s largest biotech raise of the year landing on an AI company — with 1,427x retail demand stapled to it — hands Asia-Pacific private markets a direct signal: AI-biotech companies now own an explicit exit path and a pricing reference.
Use of Proceeds and the Industry Signal
The net proceeds split roughly as follows: 48% into clinical development of key pipeline candidates, 20% toward early-stage discovery, 15% for developing and validating new generative AI models, 12% to expand the automated laboratories (Life Star 1 live since December 2022, Life Star 2 since September 2025), and 5% to working capital. Pause on the 15%: a listed biotech designating model development as a primary use of proceeds is treating AI capability as a core asset, not a disposable tool.
Commercial revenue already has a shape: three out-licensing deals (Exelixis and Menarini among them) worth up to $2.1 billion in total, co-development arrangements with Fosun Pharma, Sanofi, and Lilly, and software licenses held by 13 of the world’s top 20 pharma companies. Founder and CEO Alex Zhavoronkov frames the mission as extending human “productive longevity”; Co-CEO and CSO Feng Ren insists the value of AI runs “far beyond cost reduction and efficiency gains.” For developers and product teams the takeaway is compact: once the whole stack — generative models, automated labs, clinical validation — carries a public market price, the acceptance bar for AI in science relocates from papers to revenue and pipeline progress.