AI is spitting out more potential drugs than ever. This start-up wants to figure out which ones matter.

Editor's take

10x Science has secured seed funding to develop AI tools that sift through the growing volume of AI-generated drug candidates. This addresses a critical bottleneck: while AI can propose thousands of novel molecular structures, determining their actual therapeutic viability and safety remains a manual, time-consuming process.

The challenge lies in bridging the gap between in silico prediction and in vivo efficacy. Pharmaceutical companies, including giants like Pfizer and Novartis, are already investing heavily in AI for drug discovery, but the sheer output necessitates sophisticated filtering mechanisms. 10x Science aims to provide that crucial layer, potentially accelerating the transition of AI-discovered compounds from lab to clinic.

Future developments to monitor include the accuracy and scalability of 10x Science's platform compared to existing computational chemistry tools and internal R&D efforts at major pharma. Success will hinge on demonstrating a tangible reduction in time and cost for preclinical candidate selection, especially as models like AlphaFold 3 continue to expand the protein structure prediction landscape.