Anthropic's acquisition of Mithril Medicine, a one-person drug company, breaks the precedent of large model vendors acquiring only computational biology tool companies. Currently, the buyer pool in the AI pharmaceutical M&A track continues to expand, with transactions centered on competing for controllable, repatriable, and clearly owned data assets. Four parallel main lines have emerged: AI pharmaceutical companies acquiring wet-lab assets downstream to complete the validation chain; industry players acquiring AI tools and datasets upstream to upgrade capabilities; AI pharmaceutical peers merging for horizontal integration; and general large model and computing power vendors entering crossover to complete vertical data and execution layers. At present, targets meeting M&A requirements are very limited. Most early-stage AI pharmaceutical companies have not created substantive value for pharmaceutical R&D, and coupled with rising data ownership thresholds, many companies have lost acquisition eligibility, even heading toward liquidation or bankruptcy. The industry's core criteria for evaluating high-value M&A targets are no longer simply algorithmic models or data scale, but increasingly emphasize data scarcity, continuity and clear ownership, human validation results of models, and quantifiable commercialization capability. The AI pharmaceutical track is accelerating stratification and reshuffling: pure-algorithm AI pharmaceutical companies' moats continue to thin, and small and medium-sized CROs lacking differentiated capabilities are being squeezed from both sides and gradually cleared out; dry-wet closed-loop assets with automated wet-lab platforms, clearly owned deep data, human validation records, and customer repeat purchases have become scarce targets. The M&A market has shifted from buying stories to buying validation, and entities mastering full-chain data closed-loop production rights will possess long-term industry discourse power.