At the 2026 World Artificial Intelligence Conference in Shanghai, the Shanghai AI Laboratory unveiled ten joint scientific innovation achievements under its Co-Creating with Shusheng initiative. Presented at the Science Frontier Forum, the releases anchor on the laboratory's Shusheng Duan Yan platform, designed to transition artificial intelligence from passive analysis to active research. The portfolio demonstrates how foundation models are closing the loop between computational simulation and experimental validation across multiple disciplines. In life sciences, Shusheng AMix unifies protein structure prediction, natural language interaction, and de novo sequence design, enabling researchers to specify functional requirements directly for accelerated drug discovery. ViraHInter integrates protein sequence and structural data to identify shared host factors across influenza subtypes, outperforming prior benchmarks and advancing broad-spectrum antiviral development. The laboratory also introduced the first virtual cardiac cell model, which simulates ischemia-reperfusion injury to predict therapeutic gene targets. For neuroscience, BrainOmni standardizes heterogeneous EEG and MEG signals into a unified representation, while SigmaBrains-Lab launches a digital zebrafish system that links neural dynamics with behavioral output. Additionally, CENO, a genomic world model, decodes three-dimensional genome folding and designs functional DNA elements, boosting antibacterial gene screening efficiency nearly ninety-fold in practical applications. Across earth sciences and engineering, Earth-o1 processes over ten petabytes of observational data to deliver atmospheric forecasts that exceed traditional numerical models by fifteen percent within thirty-six-hour windows. The Shusheng Fuyao platform automates aerospace aerodynamic design, compressing concept iteration from months to hours through AI-driven simulation and wind-tunnel validation. An electroplating research agent reduces molecular screening from hours to milliseconds, shifting material development from empirical trial-and-error to high-throughput AI optimization. In semiconductor manufacturing, the ChipDesign Agent and AgenticDTCO framework autonomously orchestrate electronic design automation workflows, accelerating layout planning by over fifty percent and increasing computational energy efficiency by ninety percent. These deployments mark a structural shift toward AGI for Science. By embedding continuous feedback mechanisms between AI reasoning and wet-lab verification, Shanghai AI Laboratory is establishing reproducible pipelines for hypothesis generation and cross-disciplinary breakthroughs. The laboratory will continue expanding the Shusheng model ecosystem alongside the Duan Yan platform, scaling collaborative research infrastructure to tackle foundational scientific challenges and global technical bottlenecks.