WAIC 2025: Molecular Heart Launches 10 Major Solutions, AI Protein Design Enters 'Programmable' Era}

At the 2025 WorldArtificial IntelligenceConference (WAIC), Molecular Heart, founded by AI protein folding pioneer Xu Jinbo, showcased a remarkable achievement that injects powerful momentum into the bio-economy.

The industry’s first comprehensive AI protein optimization and design platform, MoleculeOS, was unveiled with upgrades surpassing AlphaFold 3 in complex structure prediction, ultra-high precision dynamic protein design, and dozens of solutions targeting biopharmaceuticals and bio-manufacturing industries. Xu Jinbo stated, "With world-leading AI technology, even biologists without AI backgrounds can precisely design functional proteins on computers."

As an industry-grade AI protein infrastructure platform independently developed by Molecular Heart, MoleculeOS integrates over a dozen top-tier AI protein prediction, optimization, and design technologies, including the global leading multimodal AI large model NewOrigin (Darwin). It combines scientific computing methods like molecular dynamics and quantum chemistry, covering from static observation to dynamic design comprehensively.

MoleculeOS architecture

Using advanced algorithms, Molecular Heart’s platform achieves accuracy beyond AlphaFold 3, with better physical properties in predicted structures. It excels in predicting and designing complex structures like antigen-antibody and enzyme-substrate complexes. By integrating AI with first principles, it breaks through high-precision molecular dynamic structure prediction and protein dynamic design, boosting simulation accuracy by hundreds of thousands of times to reach industrial levels. Achievements include zero-shot AI enzyme design, peptide surface design, and more, published at top conferences like ICML, KDD, and MLHC.

Relying solely on foundational AI protein technology cannot fully meet industry needs. Complex real-world demands such as long-acting drug design, precise drug delivery, and enzyme catalysis require multi-objective AI algorithms. MoleculeOS encapsulates these into automated workflows, creating drug design, enzyme design, and industrial enzyme optimization platforms, including antibody humanization, affinity maturation, protein developability, enzyme activity, stability, and expression optimization. Validated in multiple projects, it enables one-click customization of proteins with specific functions for innovative drug development and synthetic biology.

It also features a conversational AI agent that allows biologists without AI expertise to quickly and accurately design high-value molecules through dialogue.

MoleculeOS interface

As the material basis of life, proteins are in huge demand in the bio-economy. Traditionally, scientists relied on laborious lab screening, which was time-consuming and low-yield, creating a "double ten law"—a ten-year, billion-dollar curse in drug discovery.

Now, with AI, these pain points are gradually being addressed. The new "AI design + minimal experimental validation" approach revolutionizes biology. Instead of months or years of lab screening, biologists can now focus on input and results, greatly improving R&D efficiency. AI-designed high-value molecules have also increased success rates. In a long-acting antibody project for a multinational pharma, Molecular Heart designed molecules far superior to traditional wet lab methods in a shorter time.

Currently, Molecular Heart’s MoleculeOS is open to industry and academia. Xu Jinbo hopes it will amplify biologists’ capabilities, freeing them from tedious lab work to focus on strategic design and judgment, leading to safer, longer-lasting drugs and higher-performance, lower-cost biologics.