Recently, industry events in AI drug discovery have been密集, with multiple overseas AI pharma collaborations and capital deals landing, and several domestic AI pharma companies disclosing strong results and large out-licensing deals. AI drug discovery is becoming a hotspot in pharmaceutical R&D. In September this year, China's first original drug approved with AI-assisted R&D — Islatravir tablets (brand name: Aipusiwei, developed by Westlake University and others) — was officially approved for marketing. Compared with traditional R&D cycles that often take a decade, this drug took only three and a half years from source discovery to completion of clinical trials, with AI playing a key role. Is AI drug discovery likely to be the first direction in healthcare to deliver? Which parts of the value chain may benefit first? To address these questions, let's look at the latest analysis from Liang Furui, fund manager of Great Wall Pharmaceutical Industry Select.
1. Do you agree that "AI drug discovery is the first direction to deliver"? If so, at what stage is "delivery"? Is the first sign clinical data readout, receipt of licensing upfront payments, or improvement in R&D expense ratio? If not, which scenario in healthcare do you think will deliver first?
Liang Furui, fund manager of Great Wall Pharmaceutical Industry Select: I tend to believe that AI drug discovery is one of the high-value application directions in pharma worth prioritizing. Currently, pharma companies are willing to pay for R&D achievements and technical services with practical value, and related industry collaborations have already landed. But the industry as a whole is still in the early stage of commercial delivery — customer payment, company revenue recognition, sustained profitability, and sales returns after new drug approval correspond to different levels of delivery, and it is not yet sufficient to prove that the entire industry has completed full commercial closed-loop validation. In my research framework, the first core sign of commercial delivery is when customers make actual payments for clearly defined deliverables; thereafter, I continue to track project acceptance, customer repurchases, and cash collection. For software and lab service companies, the validation path for delivery acceptance and renewals is relatively direct; for AI drug R&D companies centered on pipeline licensing, upfront payments reflect partners' commercial recognition, while subsequent value depends heavily on R&D progress and clinical data results. In the long run, the commercial value of AI drug discovery still ultimately needs clinical data validation. At present, companies differ in their AI-enabled segments and technical capabilities, making it difficult to form a unified comprehensive evaluation standard externally, which also means there is still a lack of a fair and unified evaluation system for the technical capabilities of different companies in the industry. In addition, AI application scenarios with shorter process chains, such as scientific literature search and document processing, may be more likely to form stable payment models earlier, but there is currently a lack of unified horizontal data, so it is not yet possible to clearly rank the delivery pace of all medical AI scenarios. At this stage, I will focus on whether the R&D value of AI drug discovery can be transformed into sustained customer procurement, and the final distribution pattern of industry value.
2. How do you divide the delivery stages of AI drug discovery? Where are we in 2026–2027? What is the more certain trend in the next 12–24 months, and what is most easily falsified?
Liang Furui, fund manager of Great Wall Pharmaceutical Industry Select: If AI drug discovery companies are narrowly defined as service-oriented companies based on large models, or AI R&D departments within pharma companies, I would start with the iteration of large models and vertical professional models, and break down the industry along three lines: funding sources, procurement spending, and performance delivery. Applied to drug R&D, technical value needs to be reflected in core segments such as prediction, molecule design, and clinical validation; the industry delivery pace can be divided into three observation stages, which may interweave and may also exist simultaneously within the same company. In the first stage, the core focus is on the model's problem-solving ability and customers' willingness to pay; in the second stage, track procurement spending brought by technology penetration — professional model companies can obtain funds through financing, software sales, R&D collaborations, and pipeline licensing, while pharma companies can also allocate procurement from their own R&D budgets, with funds ultimately flowing to computing resources, professional data, and real experiments; in the third stage, verify whether large customers incorporate AI into formal business processes to form sustained procurement, focusing on three signals: expansion of pilot projects, procurement entering annual budgets, and substantial improvement in R&D output. Based on this framework, by 2026 some companies are already showing simultaneous characteristics of investment expansion and deeper application; in 2027 the key observation is whether early collaborations can continue to advance and form a more stable revenue and cash collection rhythm. But this does not mean the industry will enter a broad profitability stage in 2027. Model iteration still depends on experimental validation, and application deployment will continue to generate new data needs, so the industry as a whole remains in an early stage of development. Investment research can be carried out around three dimensions: formal collaboration, actual delivery, and customer repurchases. The next year is a key window to verify expansion of existing customer projects, broadening of procurement scope, and conversion of incremental revenue into profit; if there are situations such as delivery falling short of expectations, weak repurchases, or lengthening collection cycles, we need to be alert to falsification risk.
3. The AI drug discovery value chain can be roughly divided into three types of roles: "tool sellers" — platforms and automated laboratories; "pipeline sellers" — AI-driven biotechs; and "drug sellers" — traditional pharma companies introducing AI to improve efficiency. Which type do you favor most? Which segments are likely to benefit first? Is the main basis order visibility, cash flow quality, or sustainability of barriers?
Liang Furui, fund manager of Great Wall Pharmaceutical Industry Select: For companies providing R&D tools and automated laboratory services, focus on AI-related business increments in segments such as gene synthesis and protein expression, with core attention to indicators including absolute business scale, value per project, customer repurchases, and profitability. But note that corporate performance growth may also be driven by traditional R&D demand, business structure adjustments, and improvements in their own operating efficiency; it is not appropriate to attribute all improvement to AI enablement, and some listed companies may also tend to position incremental revenue more closely to the AI concept. For AI-driven biotech companies with self-developed pipelines, the research focus is on revenue sustainability, fulfillment of delivery obligations, and pipeline equity value; the tracking logic is basically consistent with the current innovative drug BD research framework, and it is necessary to focus on the R&D progress rhythm of specific pipelines. For traditional pharma companies introducing AI technology to improve efficiency, research should still focus on the R&D progress and cost optimization results of specific products. It is worth noting that if AI drug discovery is sufficiently disruptive to existing drug R&D processes in the future, the industry competitive landscape may concentrate toward leading large pharma companies.
4. What is the current valuation level of AI drug discovery targets? Which segments do you think have already been overextended, and which are still undervalued?
Liang Furui, fund manager of Great Wall Pharmaceutical Industry Select: At present, the valuations of most AI drug discovery targets already largely incorporate the market's premium expectations for their reshaping of the drug R&D paradigm. At this stage, there is still some uncertainty regarding the revenue sustainability of AI drug discovery tools and automated laboratory businesses, as well as the boundaries of AI drug discovery platforms' technical capabilities. In investment research, I will focus on two types of verifiable marginal changes: first, existing customers achieving continuous repurchases and procurement budgets continuing to expand; second, leading service companies leveraging data accumulation and delivery capabilities to obtain higher value-added orders. Only when the above substantive progress exceeds the market expectations already implied in share prices can investment opportunities with cost-effectiveness emerge. At the same time, clinical value validation is the core touchstone of the AI drug discovery industry. If technology can significantly improve R&D efficiency and substantially revolutionize the drug R&D paradigm, the industry will have long-term growth space; conversely, if technological progress falls short of expectations and cannot effectively demonstrate R&D efficiency advantages, then the industry transformation expectations carried by most current targets will face valuation regression risk.
5. Public fund positions in the pharmaceutical sector are at a five-year low, and the AI main line has clearly diverted capital. What conditions do you judge are needed for capital to return to pharma? In your framework, is AI drug discovery a "new story" to absorb returning capital, or rotation within the innovative drug main line?
Liang Furui, fund manager of Great Wall Pharmaceutical Industry Select: The core reason capital is underweight innovative drugs is that industry financial statement improvement has not yet been verified on a broad scale, and the professional cognition threshold for innovative drug clinical R&D is relatively high. To promote long-term capital returning to the pharmaceutical sector, first, industry financial reports need to show significant improvement, and second, a new industry incremental logic is needed. Currently, innovative drug overseas BD is an incremental logic that is being delivered, but its delivery cycle is long and will be intermittently disturbed by clinical trial data. If domestic companies seize first-mover advantages in the AI drug discovery wave, both innovative drug companies and AI drug discovery service companies are expected to share in industry incremental returns. Disclaimer: The information contained in this communication comes from sources the company believes to be reliable and the personal judgment of researchers, but the company does not provide direct or implied statements or guarantees regarding its accuracy or completeness. This communication is not a complete statement or summary of the relevant securities or markets, and any expressed opinions may change without further notice. This communication should not be accepted as a substitute for independent judgment or as a basis for investment decisions. The company or its related institutions, employees, or agents shall not be liable for any person's use of all or part of this content or any losses arising therefrom. Without prior written permission from Great Wall Fund Management Co., Ltd., no one may distribute, copy, reprint, or publish this report or any part of it in any form, and no abridgment or modification contrary to the original meaning of this communication may be made. The fund manager reminds that every citizen has the obligation and right to report money laundering crimes. Every citizen should strictly comply with relevant laws and regulations on anti-money laundering. Investment requires caution. MACD golden cross signals have formed, and these stocks are rising well!
Comments