Sumitomo Dainippon Pharma

Traditional drug discovery from project initiation to a clinical candidate typically requires about 5 years and the synthesis and testing of roughly 2,500 compounds. Sumitomo needed a faster, more efficient path to optimized candidates for OCD, where high attrition and costly iterative chemistry are persistent bottlenecks.

Sumitomo partnered with Exscientia, which deployed its AI-driven generative chemistry platform to design and optimize DSP-1181. The system explored chemical space computationally, proposing and ranking candidate molecules to dramatically reduce the number requiring physical synthesis and biological testing.

DSP-1181 advanced from project start to Phase I clinical trial in just 12 months—roughly one-fifth the conventional timeline. The AI platform required synthesis of only ~350 compounds to identify a viable candidate versus ~2,500 historically, an 85% reduction. Though Sumitomo later discontinued DSP-1181, it established a landmark proof point for AI-accelerated small-molecule discovery.

• AI generative chemistry can reduce compound synthesis burden by over 80%, compressing multi-year discovery programs to under 12 months.
• Speed in AI-driven discovery does not guarantee clinical success—downstream biological and safety performance still determines outcomes.
• Pharma companies can access AI discovery capabilities rapidly by partnering with specialist AI biotech vendors rather than building in-house.