The AI drug development sector has recently seen a significant surge, with two major players previously outside this space entering the field. Anthropic is building its own wet lab to test its own model's ability to direct biological experiments, while also partnering with Novo Nordisk and Bristol-Myers Squibb, and acquiring Coefficient Bio; the newly established laboratory spun off from ByteDance has also completed its first round of financing. The highly anticipated AI is currently mainly used to speed up the early stages of drug development - compressing the search space for candidate molecules and the time for experimental iteration. However, the costly and time-consuming clinical trial phase is difficult to shorten simultaneously. A higher hit rate in the early stages does not necessarily mean an increased clinical success rate.

Three types of players are actually taking completely different approaches. Companies focused on AI-driven drug development target specific problems within drug development, allowing models to serve clear tasks. Large model companies, on the other hand, pursue general capabilities, using a single base model combined with a vertical toolchain to build a scientific workbench, aiming to shorten the time for each research cycle, and particularly value data infrastructure that can be transferred across targets and tasks. Traditional pharmaceutical companies hold vast amounts of data but often lack a unified data foundation. Large model companies invest more R&D budgets into model training, compensating for the most valuable key data, especially those from failed experiments - they are as valuable as successful samples for training models.

The commercialization paths are also clearly divided. AI drug development companies usually adopt a combination of AI-CRO, software tools, and their own pipeline. Among these, their own pipeline is the easiest way to realize value. Large model companies currently focus more on preclinical research and R&D infrastructure, with some quietly advancing their own pipelines. Ultimately, self-developed pipelines remain the core anchor point for the valuation of AI drug development companies. The entire market is waiting for the results of Insilico Medicine's candidate drug in Phase III trials, using real clinical data to verify the credibility of this approach - after all, no matter how well the story is told, clinical data is the ultimate pricing standard.