Citrini Research, in collaboration with Anthropic, utilized the Claude Science model to identify potential molecules for neutralizing scorpion toxins. By processing a 35-page prompt, the AI generated approximately 180 candidate molecules in a few days, a task that typically requires a year of traditional research.
Twist Bioscience synthesized and tested 211 designs, resulting in 17 molecules that successfully bound to the target toxins. Citrini Research noted that this outcome exceeded their initial expectations for early-stage drug discovery.
Following the experiment, Citrini published a report on October 20, 2026, suggesting that the acceleration of computational design shifts the bottleneck in drug development toward physical validation. The firm anticipates increased demand for laboratory infrastructure services.
The report identifies Charles River Laboratories International Inc. and Revvity Inc. as key beneficiaries of this trend. Citrini argues that investors should focus on companies providing downstream experimental validation rather than solely on AI model developers.
The analysis highlights that while AI lowers barriers to early-stage design, it increases complexity in the supply chain due to the higher volume of synthesis and testing required. The firm advises a cautious outlook, noting that binding to a toxin is only an initial milestone.
