Shuttle Pharmaceuticals Holdings, Inc.

On November 21, 2025, we acquired substantially all assets and liabilities of Molecule.ai, a pharmaceutical software company. Molecule.ai builds an artificial intelligence (AI) driven platform for molecular discovery and early-stage drug development. This platform combines modern AI techniques with structured scientific workflows to help researchers explore chemical space more efficiently, evaluate molecular ideas with greater clarity, and make more informed decisions in early drug development. The platform is designed to accelerate drug discovery iteration cycles while maintaining scientific reproducibility, traceability, and operational reliability. Molecule.ai adapts state-of-the-art AI algorithms to create a practical, domain-specific AI infrastructure layer for molecular research and development. We intend to use Molecule.ai's molecular modeling and predictive analytics platform to significantly enhance our drug discovery and development business. On November 20, 2025, we also committed to winding down our clinical trials of Ropidoxuridine, our lead product candidate. Molecule.ai is built on three core components: a unified inference engine, an API-first integration layer, and a modular model framework. The unified inference engine manages model execution and multi-step reasoning through a deterministic and traceable sequence of operations. The API-first design allows all platform capabilities to be accessed programmatically. All predictive and reasoning functions are modular, enabling the platform to expand without changing its underlying infrastructure. Molecule.ai currently supports three scientific and computational functions: (1) molecular property prediction, (2) cross-molecule and cross-property evaluation, and (3) prediction reasoning and structured molecular insights. The platform predicts various molecular properties relevant to early-stage discovery and medicinal chemistry and provides inference pipelines for predicting these properties. Using transformer-based models, the platform computes predictive outputs across a wide range of therapeutic tasks.