□ Wang Kaixuan, Han Shitong, Chen Zhu

近年來，AI製藥市場規模迅速擴張，預計2026年將達29.94億美元。AI藥物在早期臨床試驗中展現出較高成功率，尤其是Ⅰ期臨床試驗成功率遠超傳統製藥。隨著AI製藥技術的不斷革新，包括深度學習和強化學習等技術的應用，AI有望重塑傳統製藥範式，為新葯研發帶來更多可能性。

The market size and the number of pipelines are increasing rapidly

In recent years, the scale of the AI pharmaceutical market has expanded rapidly. According to the data of the China Commercial Industry Research Institute, in 3 years, the global AI pharmaceutical market size was 00 million US dollars, and it is expected to reach 00 million US dollars in 0 years, with a compound annual growth rate of 0.0% in 0-0 years. According to the data of the qubit think tank, in 0 years, the market size of China's AI pharmaceutical foreign cooperation pipeline was 00 million yuan, and the market size of self-developed pipelines was 00 million yuan, and it is expected that the two will reach 0 billion yuan and 00 million yuan respectively this year, and 0 years will reach 0 billion yuan and 00 million yuan respectively, both showing a rapid growth trend.

Since 10 years, the number of AI pharmaceutical pipelines has increased rapidly. According to Nature Reviews Drug Discovery, the number of AI-driven new drug discovery/preclinical pipelines worldwide was 0 in 0, doubling from 0 years. In 0, this number climbed significantly to 0. According to the statistics of the "qubit" WeChat official account, at present, AI-driven clinical trial pipelines have accounted for 0% of the existing clinical trial pipelines. In recent years, there have also been cases of the use of AI technology in the research and development process of new drugs marketed, such as Pfizer's new coronavirus treatment drug nirmatrelvir tablets/ritonavir tablets combination packaging (i.e., Paxlovid), which has in-depth cooperation with XtalPi in the screening of candidate small molecule compounds and crystal form screening, and has made use of AI technology. As the pipeline products produced by AI pharmaceuticals become more and more mature, the feasibility of using AI technology for in-depth drug research and development will be further enhanced, and the number of AI pharmaceutical pipelines is expected to continue to increase. At the same time, AI technology will also be more involved in the pipeline of traditional pharmaceuticals.

According to the data of the WeChat official account of "Zhi Pharmaceutical Bureau", there are currently 5 AI-driven drug pipelines approved for clinical trials around the world, of which 0 are in Phase I., 0 are in Phase II, and 0 are in Phase III. According to the schedule of routine clinical trials, the first AI-driven drugs are expected to be approved for marketing next year, and AI drugs will shift from technical verification to commercialization considerations, and the economic value of the AI pharmaceutical industry will further grow. At the same time, it is expected that a large number of drugs in the AI pharmaceutical industry will enter the clinical trial stage next year, and with the continuous advancement of the pipeline process, the pattern of the AI pharmaceutical industry will gradually become clear.

The therapeutic field of anti-tumor and neurological diseases is an area where AI-driven small molecule new drugs are concentrated. In terms of diseases, according to the statistical results of the report of BCG (Boston Consulting Group), the pipeline of AI small molecule new drugs under development is widely distributed in anti-tumor, anti-coronavirus, neurological disorders, infectious diseases, cardiovascular diseases, metabolic diseases and other therapeutic fields, of which anti-tumor drugs account for 16% and neurological disease drugs account for 0%, which is the field where AI new drugs are concentrated. Different from the multi-disease treatment pattern of AI small molecule drugs, AI-driven vaccines are concentrated in the fields of anti-tumor, anti-COVID-19 and infectious diseases, while macromolecular antibody drugs are concentrated in the field of anti-tumor and anti-COVID-19.

Dramatically increase the success rate of clinical trials

Clinical trials are the most time-consuming, labor-intensive, and costly stage of the drug development process, and whether they can improve the success rate of clinical trials is crucial for AI pharmaceuticals. At present, the success rate of phase I clinical trials in the AI pharmaceutical pipeline is significantly higher than that of traditional pharmaceuticals, but there are still some problems, the most important of which are safety and efficacy in clinical trials.

At present, most of the molecules discovered by AI are in phase I clinical trials, and a few are progressing to phase II clinical trials and beyond. According to a study, as of the end of 40 years and 0 years, a total of 0 AI-discovered molecules in the world have completed phase I clinical trials, of which 0 have been successful, with a success rate of 0.0%, much higher than the average level of the traditional pharmaceutical industry; 0 AI-discovered molecules completed Phase II clinical trials, of which 0 were successful, with a success rate of 0%, which is comparable to the average of the traditional pharmaceutical industry. AI pharma can greatly improve the success rate of phase I clinical trials, which is related to the AI algorithm's ability to design or select drug-like molecules, including the design and screening of novel molecules with optimized ADMET (drug absorption, distribution, metabolism, excretion, and toxicity) and safety characteristics. Phase II clinical trials typically involve proof of biological or mechanical concepts, and there is still room for improvement in AI algorithms.

There is reason to believe that AI technology can further improve the success rate of clinical trials, especially in Phase II and Phase III clinical trials. AI Pharma can speed up the clinical trial process by analyzing large amounts of data and identify suitable candidates efficiently and accurately; Predictive models and clinical trial designs can be optimized to improve trial accuracy and validity.

Understanding the drivers of disease and identifying and validating drug targets are areas in which many AI native biotechnology companies, pharmaceutical companies, and academic institutions are actively investing, focusing on phenotypic data generation, reverse translation, novel patient-derived models, large language model applications, and more. These technologies can help bridge the gap between molecular design and clinical efficacy, further improving the success rate of clinical trials.

Multiple factors promote the sustainable development of AI pharmaceuticals

2024年，AI製藥領域發生了多起重大事件，包括模型發佈、企業上市、大額併購等，推動行業快速發展和整合。

2024年5月，Alphafold3模型發佈。該模型在結構預測方面取得進展，將預測範圍拓展到DNA、RNA等領域，且準確率大幅提升，有力推動AI製藥行業發展。

2024年6月，晶泰控股上市。AI製藥企業晶泰控股在香港上市，融資11.38億港元，標誌著AI製藥正式進入二級市場，為行業注入新活力。

In 7/0, NVIDIA invested in Recursio, aiming to accelerate the development of breakthrough foundational models in the field of AI drug discovery. This investment further boosts the development of AI pharma technology.

2024年8月，Recursion收購Exscientia。Recursion以6.88億美元收購Exscientia，成為AI製藥領域迄今為止金額最大的一筆併購，標誌著行業整合加速，行業龍頭嶄露頭角。

In 2024 and 0 months, the 0 Nobel Prize in Physics and Chemistry were awarded to researchers related to AI pharmaceuticals, marking the upsurge of the combination of AI technology and life sciences in the scientific research community.

Today, AI Pharma continues to empower traditional drug discovery processes such as target discovery and validation, candidate discovery, and deep learning and reinforcement learning to facilitate virtual screening and de novo molecule generation. In the short term, AI pharmaceuticals are showing a rapid growth trend in terms of both the number of pipelines and the market size, and a number of AI pharmaceutical companies stand out with the AI+Biotech/SaaS/CRO model and accelerate the process of new drug research and development with their own platforms. In the medium to long term, companies such as Recursion are beginning to move away from the traditional drug discovery model through image recognition combined with apparent screening, and AI is expected to reshape the pharmaceutical paradigm. In addition, AI drugs perform well in the early stages of clinical research, and the success rate of phase I and phase II clinical trials is superior to that of traditional drug discovery models. With the acceleration of the preclinical drug candidate process and the paradigm shift, and the coverage of AI technology in the clinical trial stage, the data barriers between in vivo and in vitro are expected to be broken, and the AI pharmaceutical industry will usher in a broader space for development. (Author's Affiliation: CITIC Securities)