NeuroAI NEUROAINEUROAI.SITE
China AI, reported clearly
1,300+ verified stories on models, robotics, BCI and the business of AI.
Browse stories
ESC

Insilico Medicine's Hong Kong Debut: An AI Drug Developer Raises HK$2.28 Billion With No Medicine Sold

AI drug developer Insilico Medicine (英矽智能) raised HK$2.277 billion in Hong Kong's largest biotech IPO, with no marketed product yet and losses widening.

2026-11-14 · 1077 words · NeuroAI
Insilico Medicine's Hong Kong Debut: An AI Drug Developer Raises HK$2.28 Billion With No Medicine Sold

A drug discoverer that has never sold a single approved medicine is now worth tens of billions of Hong Kong dollars.

Investors just handed Insilico Medicine (英矽智能) more cash in one morning than most biotech firms see in a decade.

The wager is simple and strange: trust the software to invent the molecules faster than the wet labs can.

Key takeaways

  • Insilico Medicine (英矽智能) listed on the Hong Kong Stock Exchange (港交所) on 30 December 2025 under code 3696.HK, becoming the first AI-driven biotech to list on the Main Board under the standard Chapter 8.05 route rather than the un-profitable 18A track.
  • The IPO raised a total of HK$2.277 billion (≈US$292M / RMB2.09B), the largest Hong Kong biotech offering by proceeds in 2025, at an issue price of HK$24.05 per share across 94,690,500 shares.
  • Demand was extreme: the Hong Kong public tranche was oversubscribed about 1,427 times (locking up over HK$328 billion of subscription money), and the international placement about 26 times.
  • Fifteen cornerstone investors (基石投资者) anchored the deal, including Eli Lilly (礼来), Tencent (腾讯), Temasek (淡马锡), Schroders, UBS Asset Management, Oaktree, E Fund and Taikang Life — Lilly and Tencent's first-ever cornerstone bets in a biotech.
  • The company remains unprofitable: 2024 revenue was about US$85.8 million against a net loss of roughly US$17.1 million, and it has no commercialized drug.

What actually got listed

Insilico is not a typical pharma story. Founded in 2014 and headquartered in Hong Kong, it runs a generative-AI platform called Pharma.AI that tries to compress the slow, expensive front end of drug discovery — finding a disease target, designing a molecule, and nudging it toward a clinic-ready candidate. The headline asset is rentosertib (formerly ISM001-055), an idiopathic pulmonary fibrosis candidate that has completed a Phase 2a study.

What makes the listing notable is the route. Most young drug companies in Hong Kong list under Chapter 18A, a carve-out that lets pre-revenue biotechs raise public money. Insilico instead qualified under the ordinary Chapter 8.05 profitability pathway by leaning on its software-licensing revenue — selling access to Pharma.AI to large pharmaceutical partners — which let it clear the regular revenue bar. That is a structural signal: the "AI tooling" side of the business is doing real, booked work, not just promises.

The money and who showed up

The numbers from the prospectus are striking. At HK$24.05 a share and just under 94.7 million shares, the deal pulled in HK$2.277 billion (≈US$292M / RMB2.09B). The Hong Kong public portion — only 10% of the offer — was subscribed roughly 1,427 times over, tying up more than HK$328 billion of retail orders, a record for a non-18A healthcare listing that year. The 90% international slice was about 26 times covered.

The cornerstone roster is the more interesting clue. Fifteen institutions committed, spanning global drugmakers (Eli Lilly), internet and gaming capital (Tencent), a sovereign fund (Temasek), big asset managers (Schroders, UBS AM, E Fund) and insurers (Taikang Life). Oaktree returned to Hong Kong biotech as a cornerstone for the first time in years. When a pharmaceutical incumbent like Lilly writes a cheque as a cornerstone, it is both a commercial vote and a relationship: Insilico and Lilly had already struck an AI drug-discovery collaboration worth over US$100 million in potential upfront and milestone payments earlier in 2025.

Why the profits are still absent

The caution lives in the same filing. Insilico booked about US$85.8 million of revenue in 2024 but posted a net loss near US$17.1 million, and the first half of 2025 widened that gap — revenue of about US$27.5 million against a net loss of roughly US$18.9 million, per the updated prospectus. The business is essentially two halves: a slower drug-discovery-and-licensing arm that produces most reported revenue through out-licensing deals, and a software arm that licenses Pharma.AI to partners. Neither half is yet consistently profitable.

The pipeline math explains the patience required. Rentosertib is the most advanced program but still pre-commercial; the company says Pharma.AI has generated more than 20 clinical or IND-stage assets, of which three have been out-licensed for contracts exceeding US$2 billion in aggregate value. Aggregate contract value is not cash in the bank, and milestones may never trigger. The honest read: the platform has produced real optionality, but the company's own income statement is years from self-funding.

Where the cash goes

The prospectus allocates the proceeds with unusual clarity: about 48% to push key clinical-stage candidates forward, 20% to early discovery, 15% to building out its generative-AI models and validation work, and 12% to expanding automated laboratories, with the remainder for working capital. The automated-lab line is the tell — Insilico is betting that closing the loop between AI design and robotic experimentation is what turns speed into defensible science.

According to Li An, chief scientist at BrainNet (脑机网), the Insilico listing shows capital is now pricing AI drug discovery as a platform business with recurring software revenue, not merely a bet on one molecule. Li An's comments reflect his own views and do not represent NeuroAI's editorial position or an endorsement by BrainNet (脑机网).

Honest limitations

This article relies on the company's own listing documents and prospectus-derived coverage from Shanghai Securities News, Xinhua Finance and the Pudong government, plus the company's investor release. The raised amount (HK$2.277 billion) and subscription multiples are issuer-disclosed and widely reported but not independently audited post-close. The 2025-half-year revenue and loss figures come from the updated prospectus via secondary reporting and could be restated. The "largest biotech IPO of 2025" claim refers to proceeds and to non-18A / overall Hong Kong biotech rankings as reported at listing; a different cut (e.g., 18A-only) could rank differently. The US$2 billion aggregate out-licensing contract value is gross potential, not recognised revenue, and milestone dependence is high. I have not verified insider lock-up expiry dates or post-listing share performance, and nothing here is a recommendation to buy.

What readers can do now

  • If you follow Hong Kong listings, read the Chapter 8.05 vs 18A distinction before judging the "profitable" label: Insilico qualified on software-licensing revenue, so the headline does not mean the drug pipeline is cash-positive.
  • If you evaluate AI-biotech bets, separate three things — recurring software licensing income, milestone-dependent out-licensing deals, and clinical pipeline risk — because only the first two generate booked revenue today.
  • If you track the sector, watch rentosertib's later-stage readouts and whether Pharma.AI's licensed-revenue base grows fast enough to narrow losses before the cornerstone lock-ups expire.

Related coverage

Cite this article

NeuroAI (脑机网), "Insilico Medicine's Hong Kong Debut: An AI Drug Developer Raises HK$2.28 Billion With No Medicine Sold," 14 November 2026, https://neuroai.site/a/na-capital-insilico-hk-ipo

Machine-readable: BibTeX · RIS · JSON · Markdown

More in “Capital & Markets” → · Back to home · Markdown version

China AI, reported clearly
1,300+ verified stories on models, robotics, BCI and the business of AI.
Browse stories