Anthropic wants to develop its own drugs
TL;DR
Anthropic introduced Claude Science at The Briefing: AI for Science, pitching it as a workbench that brings fragmented tools, datasets, analysis, figures, and visuals into one environment for scientists. The bigger move: Anthropic says it wants to develop drugs of its own, with life sciences head Eric Kauderer-Abrams pointing to neglected diseases as an initial focus.
Nauti's Take
This sounds big, but it is still closer to a statement of intent than a breakthrough. Claude Science makes sense as a product for research teams because modern science often runs across poorly connected tools, databases, and analysis workflows.
Developing actual drugs is a different league: wet, expensive, slow, and brutally regulated. That is where useful AI separates from pitch-deck magic.
Briefingshow
Anthropic is moving from infrastructure toward owning part of the pharma value chain. That creates a strategic tension: it sells Claude to biotech and pharma customers while potentially pursuing its own candidates. The real bottleneck is less the model than data quality, lab validation, toxicity, clinical trials and capital.
Sources