Northeast: Demand for R&D efficiency gains drives accelerated industrial adoption of AI drug discovery

Northeast released a research report stating that it focuses on translation of results and grasps the two main lines of commercial validation and clinical validation. Currently, AI drug discovery is in a stage of deepening from technology application to commercial realization. In the short term, focus on pharmaceutical company collaborations and platform delivery; in the medium to long term, focus on clinical evidence and the release of drug asset value. More noteworthy are companies with cross-project replication capability, clear rights to results, and reasonable capital investment. Key areas to track include expansion of collaborations with large pharmaceutical companies, key clinical data, actual milestone payments, and cash flow improvement, while guarding against over-extrapolation from single licensing revenue and individual successful cases. Northeast's main views are as follows: AI drug discovery combines artificial intelligence with computational simulation and experimental validation to assist in target discovery, molecular design, property prediction, and delivery formulation optimization, and is an important direction for the digital and intelligent transformation of pharmaceutical R&D. At present, the industry has formed major technology entry paths such as small molecule design, biologics R&D, and delivery and formulation optimization, gradually building a R&D closed loop of "computational designexperimental validationdata feedback," and is moving from single-point tool application to platform-based delivery and clinical validation. Market space: Demand for R&D efficiency improvement is clear, and AI-related investment is accelerating. According to a Frost & Sullivan report, global pharmaceutical R&D expenditure is expected to increase from US$260.5 billion in 2023 to US$394.3 billion in 2030, with a compound annual growth rate of 6.1%; global AI-enabled pharmaceutical R&D expenditure is expected to increase from US$11.9 billion in 2023 to US$74.6 billion in 2032, with a compound annual growth rate of approximately 22.6%. The expansion of R&D investment and the demand for efficiency improvement jointly support industry development, while the actual revenue space for platforms depends on external procurement by pharmaceutical companies, service scope, and allocation of asset rights. Technology competition: Foundation models are accelerating in adoption, and differentiation is shifting toward data and experimental closed loops. Small molecule R&D focuses on multi-parameter optimization, biologics R&D focuses on function and developability design, and delivery formulations focus on materials, in vivo distribution, and process coordination. As the threshold for foundational tools declines, the focus of competition is gradually shifting from single-point model performance to high-quality experimental data, interdisciplinary collaboration, and cross-project delivery. Whether proprietary data can continuously improve candidate quality and reduce experimental rounds has become an important basis for judging technological barriers. R&D validation: Early efficiency gains have been demonstrated in some cases, and clinical translation remains the key test. Candidate drugs developed with AI participation are gradually entering human studies, providing opportunities to validate the transition from computational capability to clinical value; the termination or failure to meet efficacy expectations of projects such as DSP-1181 and BEN-2293 also shows that AI cannot eliminate drug development risk. Existing research shows some early positive signals, but it is not yet sufficient to prove that AI has improved clinical success rates across the full process. Future attention should focus on multi-project, long-cycle, and confirmatory study results. Commercial realization: Multiple models coexist, and sustained charging and return on investment determine differentiation. The software model relies on workflow embedding and renewals, the service model relies on delivery reuse and cost control, and the asset model relies on pipeline quality and licensing rights. XtalPi, INSILICO, Jitai Technology, and Jiangsu Hualan New Pharmaceutical Material respectively reflect paths such as integrated R&D and automation, generative AI and drug assets, delivery formulations, and integration of industrial resources. Risk disclosure: Technology and clinical validation may fall short of expectations, and commercialization progress may fall short of expectations.