Shares of AI-driven pharmaceutical companies listed in Hong Kong saw significant gains on Tuesday, with GenScript Biotech Corporation, XtalPi Holdings, and InSilico Medicine each climbing more than 6% by the market close on September 17. This surge reflects growing market enthusiasm for the application of artificial intelligence in drug development.
The sector is also receiving a wave of positive developments. On September 16, market reports indicated that Anew Labs, an AI drug discovery venture spun off from ByteDance, successfully completed its initial external financing round, securing $290 million at a post-investment valuation of approximately $1.5 billion. On the same day, GenScript Biotech Corporation announced a new collaboration with Lilly TuneLab, an AI and machine learning drug discovery platform operated by pharmaceutical giant Eli Lilly and Company.
In an exclusive interview on September 17, Alex Zhavoronkov, co-CEO and Chief Scientific Officer of InSilico Medicine, shared his perspective on the industry's trajectory. He stated that AI drug discovery has now entered its "second half," a phase where the technology's ability to improve drug development efficiency is widely accepted, but the more critical test is whether it can meaningfully enhance the success rate of bringing new drugs to market.
Shifting Focus from Speed to Success
The evolution of AI in pharmaceuticals has already experienced several cycles. Currently, the application scope has broadened from early-stage drug discovery to encompass the entire research and development workflow. Zhavoronkov predicts that within the coming years, virtually all pharmaceutical companies will operate as AI-driven entities in a broad sense. He categorizes the current use of AI in the field into two main types: "AI for Science," which directly addresses scientific problems such as target discovery and molecular design, and "AI for Scientist," which employs large language models to assist researchers by streamlining literature review, data analysis, and workflow optimization to boost overall scientific productivity.
For investors seeking to evaluate the technical capabilities of AI pharmaceutical companies, Zhavoronkov suggests that a key indicator is the commercial value generated by the AI technology. This includes whether established drugmakers are willing to enter strategic partnerships, license pipelines, or pay for access to the AI platform itself. This value is most readily apparent in business development transactions. Data shows that InSilico Medicine has signed contracts totaling approximately $11 billion since 2021, with new agreements announced in the first half of this year alone reaching roughly $7.3 billion. These partnerships involve both international and domestic companies, including Eli Lilly, Servier, Takeda, Qilu Pharmaceutical, and China Medical System Holdings.
The surge in business development activity has correspondingly driven revenue growth. InSilico Medicine's interim results for 2026 indicate that first-half revenue reached $106.3 million, a 287.2% increase year-over-year. Gross margins improved to 90.3% from 84% in the prior year period, and the company reported a net profit of $35.54 million, a significant turnaround from a loss of $19.22 million in the first half of the previous year. This marks the company's first half-year profitability since its listing. Zhavoronkov noted that current revenue is still heavily weighted toward upfront payments from research collaborations and pipeline licensing. However, as collaborations progress, the company anticipates that milestone payments will account for a growing share of future revenue, creating a more balanced income structure over time.
From Usable to Optimized Application
According to data from the U.S. Food and Drug Administration (FDA), between 2016 and 2023, its drug review divisions accumulated over 500 submissions for drugs and biologics that included AI components. This suggests that AI has become a foundational tool in the drug development process. The challenge for companies now lies in optimizing how this tool is utilized.
Zhavoronkov noted a shift in the nature of partnerships between multinational pharmaceutical companies and AI firms. Earlier collaborations often focused on designing differentiated molecules for known targets, addressing issues like toxicity, selectivity, and activity. Now, some multinationals are asking AI companies to help uncover entirely novel targets with fresh biological mechanisms. He indicated that some drugmakers are seeking AI assistance in identifying more reliable and innovative targets for specific disease areas.
Data from InSilico Medicine shows that 13 of its drug candidates have received approval for clinical trial applications. Among them, Rentosertib (ISM001-055), a candidate for idiopathic pulmonary fibrosis where AI played a role in both target discovery and molecular design, has entered Phase III clinical trials with a planned enrollment of 320 patients. Zhavoronkov expressed optimism that if all goes according to plan, top-line data from the Phase III study could be available in 2029. A successful outcome would, in his view, demonstrate a more complete validation loop for AI technology in drug development, spanning from target discovery and molecular design to clinical verification. Furthermore, InSilico Medicine has nominated 33 preclinical candidates, with six of the targets involved being what the company classifies as first-in-class.
As AI becomes a standard tool in the pharmaceutical industry, the future role of AI drug discovery companies is also evolving. Based on InSilico Medicine's current business model, the company continues to advance its proprietary pipeline, holding drug assets and moving them into the clinic to generate clinical evidence that further validates its AI platform. Simultaneously, it engages in long-term strategic collaborations with major pharmaceutical companies, converting its AI-driven research capabilities into drug assets and sharing in the value of innovation through upfront payments, milestone fees, and sales royalties.
Zhavoronkov clarified that InSilico Medicine does not position itself as a traditional contract research organization that simply executes specific research tasks according to client specifications. Instead, it aims to play a more integral role in key R&D processes like target discovery and molecular design. Looking ahead, he emphasized that AI is not destined to replace scientists. "AI cannot yet independently complete all aspects of drug development," he said. "Combining AI with the professional expertise of researchers remains essential." He concluded that the real question the industry needs to answer is not whether a drug can be developed "100% by AI," but whether safe and effective medicines can be created through effective collaboration between humans and AI.
Comments