"In the future, almost everything in biology will largely start in silico, and largely end in silico." When NVIDIA founder and CEO Jensen Huang repeatedly emphasized this vision under the spotlight at the J.P. Morgan Healthcare Conference and GTC大会, biology was transitioning from an experimental science reliant on trial and error into a predictable, programmable data science. Insilico is precisely the practitioner of this "first principles" approach that Jensen Huang has frequently highlighted. On December 30, 2025, Insilico (3696.HK) officially commenced trading on the main board of The Stock Exchange of Hong Kong, with its opening price at HK$35, marking a 45% surge from its issue price and giving the company a market capitalization of HK$19.5 billion.
This event represents the largest IPO in the Hong Kong stock market's biopharmaceutical sector for 2025, raising a total of HK$2.277 billion; it is also a crucial market test for the "AI+Biotech" business model. Unlike many unprofitable biotech companies that list relying on "Chapter 18A," Insilico is the first AI biopharmaceutical company to list via Rule 8.05 of the Exchange's Main Board Listing Rules—signifying that the company not only possesses promising future pipelines but has also passed stringent profitability or revenue tests, demonstrating proven commercial viability. Insilico's listing marks a watershed moment for the AI drug discovery industry, signaling a shift from "proof-of-concept" to "industrial-scale output." The capital market's "vote with its feet" was evident in the prestigious cornerstone investor lineup, spearheaded by Eli Lilly and Tencent. According to the offering results, the Hong Kong public offering segment recorded approximately 1,427.37 times oversubscription, locking in over HK$328.349 billion in subscription funds; the international placing segment was also oversubscribed by 26.27 times. Both figures set new records for non-Chapter 18A healthcare IPOs in Hong Kong for the year. In Hong Kong IPOs, the list of cornerstone investors serves as a microcosm of institutional confidence in the issuer's fundamentals. For this offering, Insilico attracted 15 global cornerstone investors, with total subscription amounts reaching approximately $115 million. The most impactful names on the list were the appearance of Eli Lilly and Tencent. Eli Lilly's participation as a cornerstone investor marks its first such bet on the AI drug discovery arena. This sends a strong signal: multinational pharmaceutical corporations not only endorse Insilico's technology platform but are also paving the way for future pipeline collaborations. Similarly, this is Tencent's first involvement as a cornerstone investor in a Biotech IPO. It represents a confirmation by tech giants of the cross-disciplinary trend of "AI+Science." AI drug discovery demands immense computing power and cloud infrastructure; Tencent's contribution extends beyond capital, potentially enabling deep collaboration on computational infrastructure. On another front, Oaktree Capital also made its return to the Hong Kong Biotech market this year for the first time. Oaktree is renowned for its expertise in distressed investing and value discovery. Its selection of Insilico as the "first shot" for its market re-entry likely indicates that the company's risk-reward profile has become sufficiently attractive. Additionally, cornerstone investors included Temasek, Schroders, UBS, E Fund Management, and Taikang Life Insurance. This "feverish" subscription demand confirmed its AI premium. Since ChatGPT ignited the generative AI wave, AI-concept stocks have commanded extremely high premiums, and Insilico has inherited these valuation expectations. On its first trading day, Insilico's opening price surged 45% above the issue price. This performance demonstrates that, even amidst a capital winter for biopharma, the market remains willing to grant high valuation tolerance for "hard tech." The market's enthusiasm for Insilico stems fundamentally from its unique "dual-engine" business model: Artificial Intelligence plus Innovative Drug Discovery. Insilico licenses its proprietary generative AI platform, Pharma.AI, to pharmaceutical companies for a subscription fee. This not only generates predictable annual recurring revenue but, more importantly, establishes extremely high customer stickiness. Once pharmaceutical companies become accustomed to using Chemistry42 for molecule generation, the switching costs become prohibitive. This creates a natural customer pool for subsequent pipeline collaborations. Simultaneously, widespread software deployment allows Insilico to collect vast amounts of external user feedback data, which in turn refines its algorithmic models, creating a closed-loop system of "data-algorithm-product." Since 2020, the Pharma.AI platform has been launched commercially in a modular software format, building a global合作 network. As of the last practicable date, it had secured software licensing agreements with 13 of the world's top 20 pharmaceutical companies. Innovative Drug Discovery is the true engine for explosive growth. This follows the typical Biotech model but operates with higher efficiency. Insilico utilizes its own platform to develop innovative drugs, generating revenue through out-licensing agreements or by advancing assets to clinical stages itself. This segment currently contributes over 90% of the company's revenue. The core logic lies in leveraging AI's high success rate to mass-produce preclinical candidate compounds and monetize them at their peak value. This "SaaS + Biotech" hybrid model addresses the pain points of traditional Biotech firms—zero revenue and high risk in their early listing stages—while retaining significant valuation elasticity. This is the key source of valuation premium that distinguishes Insilico from traditional CXOs and pure-play Biotech companies. The AI drug discovery arena is not short on narratives; what has been lacking is validation through clinical data. Insilico's greatest moat lies in having used clinical data to prove AI's efficacy. According to a Frost & Sullivan report, the traditional drug discovery process, from target identification to PCC nomination, takes approximately 4.5 years on average. By utilizing its Pharma.AI platform, Insilico has dramatically compressed this timeline to just 12-18 months, requiring the synthesis and testing of only 60-200 molecules per project. Under the same budget and time constraints, Insilico can explore more targets, granting it more opportunities for trial and error. For the high-risk, high-reward gamble of drug R&D, AI systematically alters the odds by increasing the number of bets placed and improving the probability of success for each individual bet. ISM001-055 (Rentosertib) stands as the definitive proof point for this logic. As the world's first AI-discovered and AI-designed drug candidate to reach Phase II clinical trials, ISM001-055 targets idiopathic pulmonary fibrosis. PandaOmics identified TNIK as a potential target, and Chemistry42 generated a novel molecular structure, forming a complete, closed-loop process. Top-line data from the Phase IIa study released in October 2024 showed the drug exhibited positive efficacy signals in patients, with excellent dose-dependent performance. The improving trend in FVC validated the accuracy of the AI's predictions. The success of ISM001-055 constitutes the "Turing Test" for the AI drug discovery industry—proving that AI can not only generate molecular structures but that the resulting drugs are genuinely safe and effective in humans. Beyond this, the company has built a deep pipeline. ISM3091 (a USP1 inhibitor) has been licensed to Exelixis in a deal potentially worth up to $955 million; ISM5043 (a KAT6 inhibitor) has been licensed to Stemline, a Menarini Group company; and ISM5411 (a PHD1/2 inhibitor), developed in-house for IBD, has also entered Phase I clinical trials. These pipelines not only demonstrate the replicability of the AI platform but also provide the company with sustained revenue-generating capability. Over the past decade, we witnessed the miracle of Moore's Law in the semiconductor industry. In the coming decade, we may witness AI breaking "Eroom's Law" (the inverse of Moore's Law, stating that drug development costs increase exponentially over time). Under the prophecy of Jensen Huang's "In Silico" vision, Insilico has taken a critical step. Whether this step can deliver truly accessible, affordable, and breakthrough treatments for patients worldwide is a question only time will answer.
Comments