A research report from Orient Securities Company Limited highlights that as global regulatory shifts accelerate, the large-scale application of AI in pharmaceuticals is rapidly advancing. It has evolved from an early-stage research tool into an end-to-end system covering the entire R&D chain, potentially overturning the "Moore's law of diminishing returns" characterized by long cycles, high costs, and high failure rates in innovative drug development. The report notes that overseas multinational corporations and tech giants are fully embracing AI, while domestic AI-driven pharmaceutical companies are emerging, with related industry chain firms also poised to benefit. Key stocks include 剂泰科技-P (07666), 英硅智能 (03696), 晶泰控股 (02228), 金斯瑞生物科技 (01548), and 百普赛斯 (301080.SZ).

Technical Disruption of Costs: AI's "Speed and Savings" Reshape Pharma

The foundational logic behind China's rise in innovative drugs lies in AI redefining pharmaceutical "speed and savings," rooted in demographic dividends and cost advantages. In drug development, AI has shifted from an early-stage research tool to an end-to-end system covering the entire R&D chain, gradually expanding from small molecules to more emerging modalities, becoming a key battleground for China and the US. If domestic AI-driven pharma first completes the industrial closed loop, China's innovative drug sector will experience its true "DeepSeek" moment, transitioning from human-driven "speed and savings" to AI-driven "speed and savings." Leveraging massive data, high-frequency iteration, and real-world evidence feedback forming a positive flywheel, domestic innovative drugs will evolve from Chinese high-end manufacturing to a global innovation hub.

Drug Discovery First, New Targets Emerge

Currently, AI's large-scale application in medicine is concentrated in the drug discovery phase: AI models multi-omics data to identify disease-driving factors, and uses large language models to rapidly integrate literature data, efficiently discovering new targets. According to data from 英硅智能, its AI platform has identified 29 pipelines, completing the process from project initiation to candidate molecule determination in an average of 12 to 18 months, reducing R&D cycles by 50%-70% and costs from tens of millions of dollars to millions, a decrease of 80%-90%.

Preclinical Replacement: New Methodologies Accelerate Progress

Since April, regulatory authorities in both China and the US have issued policies, marking a global consensus on New Approach Methodologies (NAMs). NAM technologies fall into three categories: stem cells as the foundation, organoids as the core, and AI as the key accelerator. AI computational models extract potential determinants of drug sensitivity through multi-omics data, then validate and optimize via organoid experiments, upgrading organoids from simple functional tests to decision-support platforms linking patient biology with predictive reasoning. In the current preclinical space, AI-driven ADMET prediction for small molecules is mature, and domestic companies have formed applications in frontier areas like crystal form prediction and delivery optimization, potentially accelerating drug-to-clinic conversion.

Clinical Trials Begin Validation, Industry Ushers in Change

Phase I-III clinical trials typically consume 60%-70% of total R&D costs and take 6-7 years, representing the highest failure rate, most expensive, and most deserving of AI restructuring. Currently, AI's role in clinical trials leans more toward data integration and patient management, with its true value limited. Notably, in April 2026, the FDA announced two major initiatives as part of its plan to advance real-time clinical trials (RTCT), and plans to launch broader AI-optimized early clinical trial pilots this summer, signaling a revolution in the 60-year-old clinical trial model. The report suggests AI can transform highly fragmented, complex, heterogeneous global data into evidence for expert and regulatory judgment, improving early-stage decision quality and reducing hidden costs and sunk costs of later failures. Collaboration between experts and AI will truly enhance the success rate of new drug development and shorten timelines.

Risk Warning

Risks include data, technical, clinical, regulatory, and geopolitical factors.