Zheshang Securities: AI Pharmaceuticals Reshape the Scientific Research Paradigm, Long-term Focus Can Continue to Iterate R&D Platforms

The Zhitong Finance App learned that Zhishang Securities released a research report saying that the pharmaceutical research and development industry is undergoing similar industrial changes: the number of large biochemical models has exploded, and bio-based models have been iterated many times, and capabilities have been greatly improved. The development logic of the AI pharmaceutical industry is in a new stage from scientists' decisions to intelligent autonomous decision verification. The paradigm of scientific research may be reshaped, and the value of the platform is increasing. In the long run, it is recommended to focus on R&D platforms that can be continuously iterated.

The main views of Zheshang Securities are as follows:

Industrial stage

AI pharmaceuticals have experienced developments from computational chemistry to intelligence. From the 1960s to the present, the impact of computer science on the pharmaceutical sector has gone through the era of computational chemistry, the era of flux automation, and finally the age of machine learning, and finally the age of intelligent devices. The industry is also leaping from early computational chemistry, simulation, structure prediction, and molecular generation to automated technology.

Landing rhythm

The development of AI pharmaceutical technology is a reflection of the development of general AI technology capabilities in the pharmaceutical sector. The bank noted that general AI technology has experienced a rapid improvement in large language model capabilities, followed by intelligent products such as Claude Code and Workbuddy, which actually changed the workflow and results of some industries. The bank discovered that the drug research and development industry is undergoing similar industrial changes: the number of biochemical models has exploded, and the biobasic model represented by AlphaFold has been iterated many times to achieve significant improvements in capabilities. From predictive evolution for specific tasks to simultaneously predicting complex structures of all types of biomolecules such as proteins, nucleic acids, small molecules, ions, and modification residues, the accuracy has improved markedly, and most basic models are open source.

Paradigm shift

The early drug development automation platform has entered the implementation verification stage. The closed loop of multiple Agent+ dry and wet tests forms an AI pharmaceutical platform. Essentially, it automates pre-clinical drug development and verification. Through full-process task orchestration and independent experimental design, scientists lead and decide to transform the scientific research process into the entire process of early drug discovery by intelligent agents, shifting early drug discovery from a paradigm of low frequency verification to a new paradigm of high frequency, high throughput, and multiple rounds of rapid iteration. The pipeline for the development of a new automation platform has not yet entered the clinical verification stage, but through MNC's large cooperative orders and the explosion of the performance of US companies such as Twist and Kingsley, it can be seen that this round of paradigm change has moved from theoretical research to implementation verification.

Industry outlook

The platform iteration process benefits the industrial chain. The result is an explosion of capabilities and the establishment of barriers for AI pharmaceutical platforms. The bank believes that from the perspective of industry development, high-frequency iteration will bring about a blowout in demand at the verification stage in the short term, but in the long run, high-speed iteration may bring about an improvement and generalization of platform capabilities. Discovery platform capabilities that can be continuously iterated will improve or be the core of the industry, and Claude in the pharmaceutical industry may be coming.

In the short term, it is recommended to pay attention

1) Gene synthesis. Follow: Kingsley Biotech, Yiqiao Shenzhou, Bepsis, Heyuan Biology. 2) Protein Expression and Purification: Focus: Optomax, Nanowei Technology, Saifen Technology; 3) Recombinant Target Protein/Cytokine Link: Focus: Bepsis, nearshore protein, Yiqiao Shenzhou, Haoyuan Pharmaceuticals, Novizan, Uningwei. 4) Small molecule synthesis: Focus: Bide Pharmaceutical, Haoyuan Pharmaceutical, Haofan Biology, Aladdin, Titan Technology. 5) Mouse models: Focus on: Baiosetu, Yakang Biology, Nanmo Biology. 6) Pre-clinical CRO: Focus on Pharmacovigilance, Kanglong Chemical, Pharmacovigilance, Zhaoyan New Pharmaceutical, Innox, and Medici. In the long run, it is recommended to focus on R&D platforms that can be continuously iterated: Jingtai Holdings, Insili Intelligence, Jietai Technology, Hualan Co., Ltd., etc.

Risk Alerts

The progress of AI pharmaceutical research and development may fall short of expectations; industry competition may intensify; clinical trial results may fall short of expectations; and tense international relations may cause multinational cooperation to fall short of expectations.