The field of AI-powered drug discovery has seen a flurry of recent activity. Anthropic has not only established its own wet lab to test the ability of models to direct biological experiments but has also collaborated with pharmaceutical companies such as Novo Nordisk and Bristol-Myers Squibb, and acquired Coefficient Bio. Meanwhile, ByteDance's spun-off lab, Xinsheng Laboratory, has successfully completed its first round of funding. Currently, AI technology is primarily applied in the early stages of new drug development, effectively enhancing R&D efficiency and shortening the time required for search and experimental iterations. However, the time required for clinical trials is difficult to shorten in tandem, and improvements in early hit rates do not necessarily equate to higher clinical success rates. In terms of development paths, AI-driven pharmaceutical companies, large model companies, and traditional pharmaceutical firms each have their own focuses: AI-driven pharmaceutical companies concentrate on solving individual problems in drug R&D, with their models serving specific tasks; large model companies pursue general-purpose capabilities, aiming to build scientific workbenches through a single foundational model plus vertical toolchains, with the goal of shortening single-cycle times and emphasizing the construction of data infrastructure across targets and tasks; traditional pharmaceutical firms, while possessing rich data resources, lack unified data infrastructure. Regarding R&D budget allocation, large model companies are more inclined to invest funds in model training and acquire key data, especially failure data, through self-built labs and external collaborations. In terms of commercialization, AI-driven pharmaceutical companies mostly adopt models combining AI-CRO, software tools, and self-developed pipelines, with self-developed pipelines holding greater potential for value realization. Large model companies currently focus primarily on preclinical research and R&D infrastructure construction, with some also actively advancing pipeline development. Pipelines remain the core basis for evaluating the value of AI-driven pharmaceutical companies, with the market awaiting the Phase III results of Insilico Medicine's drug candidate to validate the value of AI-driven drug discovery. Clinical data remains the ultimate criterion for pricing.