Exclusive | ByteDance's AI Drug Discovery Unit Initiates Spinoff and Funding, AI4S Enters Industrialization Phase
Exclusive | ByteDance’s AI Drug Discovery Unit Initiates Spin-off for Financing, Signaling Industrialization Phase for AI4S.
ByteDance’s AI drug discovery business line has begun the process of spinning off as an independent entity to raise outside funding, marking a key step toward industrializing its AI for Science (AI4S) efforts. Post-spin-off, ByteDance will remain the controlling shareholder. The new company will inherit the core team, algorithms, technology platform, and existing pipeline assets. It will also continue to receive computing power support from ByteDance's Volcano Engine.
The AI drug discovery team, established in 2021 and led by Liu Kai with a core team of about 50 AI and pharmaceutical experts, has been responsible for foundational model research and industrialization. It has consolidated ByteDance's protein structure prediction model team and launched several key technologies. These include the molecular structure prediction models Protenix and Seedfold, the protein design tool PXDesign, and the AI drug discovery platform "Anew Labs." Anew Labs has produced research covering protein-ligand dynamics, molecular generation, and free energy calculation, and has developed early-stage drug pipelines like IL-17 and IL4R inhibitors. Notably, its IL-17 small molecule program, presented in April 2026, demonstrated the first small-molecule blockade of three IL-17 dimers, a significant step in autoimmune disease research.
This progress demonstrates ByteDance’s AI capabilities have advanced from model development to validating specific drug targets and molecules. The company believes the opportunity to transition from research to industry is now ripe. The spin-off aims to establish an organizational structure better suited to the business's unique needs—long R&D cycles and complex validation processes involving wet labs and clinical trials—to attract top talent and drive deeper integration of AI with the pharmaceutical industry.
The move responds to the pharmaceutical industry's pressing need to improve efficiency amid high R&D costs, long timelines, and high failure rates. The field of AI4S is rapidly advancing, as seen with tools like AlphaFold evolving from protein prediction to modeling complex biological interactions and the emergence of multimodal molecular generation models for drug design.
ByteDance has been building its AI4S capabilities for years, exploring areas like computational biology, molecular simulation, and materials science. This spin-off represents its first major attempt to industrialize AI4S. An insider stated the company attaches great importance to this move, hoping that an independent entity with greater decision-making flexibility can pioneer a viable industrial path for AI4S in China.
marsbit06/10 06:26
