JPMorgan initiates coverage on China’s AI-driven drug discovery sector, assigning Overweight ratings to Insilico Medicine and XtalPi

Overview

This report is JPMorgan’s initiation of coverage on China’s AI-driven drug discovery sector and on two Hong Kong-listed companies, Insilico Medicine and XtalPi. The report argues that China’s AIDD industry is moving from algorithmic promise toward industrial execution, with business models also shifting from exploration to early commercialization; BD deals, service revenue, and clinical milestones are providing both financial and scientific validation for the sector.

Core views

The core views include three points. First, AIDD significantly improves the unit economics of drug R&D through a dry-lab/wet-lab closed loop and may increase the probability of R&D success. Second, China’s dense pharmaceutical R&D infrastructure, clinical resources, talent, and regulatory environment give domestic AIDD companies faster and lower-cost execution advantages over some overseas peers. Third, partnerships with large pharmaceutical companies, out-licensing of assets, clinical milestones, and revenue growth are validating the commercial viability of AIDD companies. The report is positive on Insilico’s high-risk, high-reward deep drug pipeline, and also on XtalPi’s platform-based growth engine composed of AIDD services, robotic wet labs, and diversified businesses.

Analysis framework

The report analyzes AIDD by combining industry frameworks, technology stacks, commercial validation, company segment valuation, and peer comparison. At the industry level, it focuses on market opportunity, the technology closed loop, China’s infrastructure advantages, and clinical/BD catalysts. At the company level, it separately evaluates Insilico’s pipeline and licensing potential, as well as XtalPi’s services, robotic laboratories, AI4S, and incubation businesses. For valuation, it uses SOTP with DCF as a cross-check.

Methodology notes

• Industry analysisThree-pillar framework

  The AIDD investment thesis consists of three pillars: R&D efficiency, China’s infrastructure advantages, and commercial and clinical validation.

  The report believes the core value of AIDD is not just the algorithm itself, but whether AI predictions can be converted into real experimental and clinical assets; China’s ecosystem amplifies this closed loop in terms of cost, speed, data, and regulation.

• Technical analysisDry-lab/wet-lab closed loop

  AI generation and prediction, automated experimentation, and experimental data feedback together form an iterative closed loop.

  The dry lab is responsible for molecular design and prediction of biological properties, while the wet lab uses automated robotics for synthesis and testing, and the experimental results are then fed back into the models to form a proprietary data flywheel.

• Valuation methodSOTP and DCF cross-check

  Different businesses or assets are valued using AI-enhanced EV/S, rNPV, or discounted cash flow.

  XtalPi’s chemical synthesis services are benchmarked against China CRO EV/S with an added AI premium; Insilico’s drug discovery collaborations and pipeline valuation are more affected by probability-of-success assumptions.

Asset mapping & comparison

Structured mapping from thesis to named assets (strengths, weaknesses, peers, risks).

• Insilico Medicine 3696.HK

  A China AIDD company with a deep drug pipeline and platform model; the report initiates coverage with an Overweight rating.

  Strengths

  It has end-to-end capabilities including the Pharma.AI platform, PandaOmics, Chemistry42, and Science42, with a deep pipeline and out-licensing potential; the report highlights its 12- to 18-month discovery cycle, low PCC cost, and strong clinical catalyst upside.

  Weaknesses

  Its investment upside depends more on clinical success probabilities and pipeline progress, making it a high-risk, high-reward asset.

  Comparison

  Compared with XtalPi, Insilico is more focused on drug pipelines and clinical/BD catalysts; compared with overseas AIDD companies, the report believes it benefits from China’s R&D infrastructure advantages in cost and speed.

  Risks

  Clinical trial failure, downward revisions to PoS assumptions, BD revenue below expectations, licensing frictions, and model generalization failure.

• XtalPi 2228.HK

  A China AIDD platform company focused on services, robotic wet labs, and AI4S; the report initiates coverage with an Overweight rating.

  Strengths

  Its business is diversified, including drug R&D services, robotic laboratories, pharma collaboration, AI4S infrastructure, and an incubation model; it serves 17 of the global top 20 pharmaceutical companies and has a customer and data flywheel.

  Weaknesses

  Valuation is sensitive to changes in AI-enabled EV/S multiples, and revenue may be affected by project-based services and BD timing.

  Comparison

  Compared with Insilico, XtalPi is more like AIDD infrastructure and a services platform, potentially with lower volatility but still dependent on commercialization delivery; the report recommends holding it together with Insilico.

  Risks

  Service and BD revenue below expectations, compression of the AI premium, robotic wet lab execution below expectations, and intensifying industry competition.

Key data

• Global AI-enabled drug R&D spendingFrom US$11.9bn in 2023 to US$74.6bn in 2032, CAGR 22.6%From the F&S forecast cited in the report.
• Global generative AI marketExpected to reach US$37.4tn by 2032, 2023-2032E CAGR 25.8%The report views this as the broader technological infrastructure and adjacent commercialization backdrop for AIDD platforms.
• Insilico target priceHK$71Initiation rating is Overweight.
• XtalPi target priceHK$10Initiation rating is Overweight.
• Implied 2027E P/SInsilico 19x; XtalPi 25xThe report says this is below the roughly 63x to 66x average level of overseas AIDD companies.
• Insilico discovery cycle and costAverage 12 to 18 months from target discovery to PCC; cost per PCC about US$3mn to US$5mnThe report compares this with the traditional industry average of about 4.5 years.
• China R&D and clinical cost advantageTime from early discovery to IND shortened by about 50% to 70%; direct clinical execution costs about 30% to 60% lower than in the United States and EuropeThe report believes this advantage comes from China’s clinical, preclinical, and industrial-chain infrastructure.
• XtalPi customer and data assetsServes 17 of the global top 20 pharmaceutical companies and works with more than 300 biotech and pharma customersThe report also mentions that METiS has more than 30 global partnerships, over 100,000 wet-lab data points, and a proprietary lipid structure library of over 10 million compounds.

Impact & implications

If the report’s view proves correct, Chinese AIDD companies could see a multi-year re-rating: Insilico’s clinical readouts, BD licensing, and pipeline advancement would drive share-price upside, while XtalPi’s service revenue, robotic laboratory capabilities, and AI4S expansion would provide a more infrastructure-like growth path. At the industry level, AIDD may reshape drug R&D cost, timelines, and success rates, but valuation still depends heavily on clinical success probabilities, BD sustainability, and regulatory/geopolitical frictions.

Risks

• Clinical trial failure or key pipeline data missing expectations.
• BD revenue below expectations, or project-based revenue volatility causing the business model to be questioned.
• China-US drug licensing frictions, US sanctions, or geopolitical regulatory risks.
• AIDD model generalization failure, making it unable to continuously improve R&D success rates.
• Downward revisions to AI-enabled valuation multiples or probability-of-success assumptions.
• Changes in regulatory pathways, medical data circulation, or AI drug development rules falling short of expectations.

What to watch

• Key clinical readouts from Insilico’s pipeline, especially assets such as Rentosertib and ISM5411.
• Insilico’s subsequent out-licensing, upfront payments, milestone payments, and progress in partnerships with major pharmaceutical companies.
• XtalPi’s service revenue growth, robotic wet lab throughput capacity, and progress in AI4S commercialization.
• How China’s NMPA implements policies for innovative drugs, AI medical products, and IND review pathways.
• Whether China’s medical data circulation policies are truly implemented and provide high-quality real-world data for AIDD models.
• Changes in overseas AIDD company valuations and whether the relative valuation discount of Chinese AIDD narrows.