BioGeometry Secures Hundreds of Millions in Strategic Financing to Pioneer Bio AI World Models

Recently, BioGeometry announced the completion of a new round of strategic financing, raising hundreds of millions of RMB. This round was co-led by the Shanghai Biomedical Innovation and Translation Fund, CAS Investment, Fortune Capital, and Link-X Capital, with participation from Gaorong Capital and the Index AI Industry Innovation Fund. Index Capital served as the exclusive financial advisor. The newly raised funds will be utilized to continuously iterate GeoFlow, BioGeometry’s proprietary all-atom biomolecular world model, and to fully accelerate the construction of the company’s high-throughput automated laboratories as well as the advancement of its diversified innovative drug pipelines.

An “AI-Native” Pioneer: Building a Universal Operating System for All Biomolecules

In today’s convergence of artificial intelligence and life sciences, BioGeometry deeply understands the critical importance of developing foundational large models in-house. Therefore, the company has consistently pursued full-stack self-development across both dry and wet labs to provide a more robust algorithmic foundation for the industry. Amidst the current wave of accelerated AI evolution, the tech industry is exploring three clear trajectories:

• Digital AI, represented by large language models and multimodal models, revolutionizing how humanity interacts with information and knowledge;
• Physical AI, represented by autonomous driving and humanoid robots, revolutionizing how humanity navigates the physical world;
• Bio AI, the ultimate proposition BioGeometry is solving—revolutionizing how humanity decodes the fundamental language of life to engineer biomolecules with atomic-level precision.

Instead of building incremental efficiency tools at the application layer, BioGeometry has chosen a more challenging, long-cycle path from its inception: to reshape the R&D paradigm of life sciences from the ground up and build a next-generation AI-native biotechnology company. This confidence to pursue a more foundational path stems from its world-class scientific team: founded by renowned AI4S (AI for Science) scientist Professor Jian Tang, with Turing Award laureate and one of the three godfathers of deep learning, Yoshua Bengio, serving as the Chief Scientific Advisor. Core team members also hold multiple international milestone achievements in fields such as graph machine learning and diffusion generative models. Recently, the team participated as core contributors in the development of NVIDIA’s open-source protein foundation model, La-Proteina, and the AI virtual cell model, PerturbDiff, garnering widespread attention in the research community and demonstrating the scientific prowess to define next-generation life science infrastructure.

Proprietary “All-Atom Biomolecular World Model” GeoFlow Sets the Global Standard

The fundamental challenge of life sciences is whether we can truly understand and precisely design molecular interactions at the biomolecular level. BioGeometry’s solution is to build the GeoFlow series—an “all-atom biomolecular world model” that understands the laws of life and matter. The generational evolution of GeoFlow clearly demonstrates BioGeometry’s sustained technological leadership at the forefront globally:

All-Atom Modeling | GeoFlow V1

By then, mainstream AI4Protein models often treat protein chains as linear sequences, performing approximate modeling at the amino acid residue level. In contrast, GeoFlow directly and precisely characterizes 3D space at the atomic level, incorporating the spatial coordinates and chemical bond angles of every single atom into comprehensive modeling. In protein-protein complex structure prediction tasks, GeoFlow V1 matched the performance level of AlphaFold 3.

Unification of Prediction and Design | GeoFlow V2

Before the advent of BindCraft, even models as powerful as AlphaFold could only perform structure prediction – designing new proteins require specialized tools like RFDiffusion. GeoFlow V2 is among the first in the world to unify protein structure prediction and de novo design within a single model. It also became the first AI foundation model in China to achieve breakthroughs in de novo antibody design.

Simulating Evolution through Multi-Step Reasoning | GeoFlow V3

Just as large language models (such as OpenAI o1 and DeepSeek R1) evolve toward deep thinking and reasoning, BioGeometry is the first to bring this frontier trend into the biomolecular space. GeoFlow V3 breaks through the traditional linear “generation-screening” process by constructing a “generation-evaluation-optimization” multi-step reasoning loop, capable of simulating the natural affinity maturation process of antibodies for in silico protein evolution. In de novo design tasks across over 20 targets, GeoFlow V3 achieved an average hit rate of nearly 20% for hit molecules—an improvement of nearly 100 times over previous-generation AI methods, directly condensing the lead molecule discovery cycle to under three weeks. This evolutionary speed is leading the direction of global Bio AI development.

Currently, BioGeometry is developing the next-generation biomolecular world model, GeoFlow V4, which will expand the modeling scale from independent molecular interactions to more complex levels—moving from “designing single molecules” to “designing molecular systems”—laying the foundational AI infrastructure for next-generation drug discovery and biomolecular engineering.

Moving Beyond the Hype: Validating AI with Proven Successes in Drug Discovery and Biomanufacturing

Driven by a dual-engine strategy of “collaborative development + in-house pipeline,” BioGeometry has delivered multiple industry-leading commercial cases in highly challenging areas such as biomedicine and synthetic biology.

Biomedicine: Revolutionizing Antibody Drug and Vaccine Development

• Tumor Therapy — De novo design of highly specific antibodies: Faced with highly similar proteins in homologous families that traditional methods struggle to distinguish, GeoFlow directly embedded “specificity” as a prerequisite constraint into the generation phase. By designing no more than 100 sequences, it yielded two antibodies that simultaneously possessed high selectivity and high affinity, pushing the boundaries of existing AI tools.
• International Pharma Collaboration — Simultaneous multi-objective optimization of lead antibodies: For a multi-objective optimization project that an internal team of a renowned international pharma company struggled to resolve over a year, GeoFlow—in a zero-shot scenario (without fine-tuning on target-specific data)—delivered a perfect molecule in just one design round. The resulting molecule featured a more-than-ten-fold increase in affinity, an 8-fold increase in expression levels, and a human germline content optimized to over 90%, shortening the project delivery cycle by more than 80% compared to client expectations.
• Vaccine R&D — From “naturally impossible” to engineered solutions: Addressing a long-standing industry challenge where a natural viral antigen protein could not stably form dimers, GeoFlow increased the proportion of the protein’s dimers from less than 10% to over 90%, while simultaneously boosting expression levels by more than 45 times, fundamentally overcoming the limitations of natural proteins.

Synthetic Biology: Reconstructing Industrial Enzyme Design, Multiple Pipelines Enter Pilot Scale

• Natural Borneol: Achieved biosynthesis for the first time globally, with a chiral purity up to 99.9%, drastically reducing costs by 80% compared to traditional plant extraction.
• α-Ketoglutarate: Directed optimization of key enzyme activities, reducing costs by over 60% compared to existing biosynthesis technologies on the market.

From molecules to cells, from understanding life to designing life—BioGeometry is building the foundational infrastructure that allows life to be engineered and written.

The Inflection Point: Accelerating the Convergence of R&D and Commercialization

The injection of hundreds of millions of RMB in this round officially hits the fast-forward button for BioGeometry’s growth. The development of the next-generation GeoFlow V4 model will extend the modeling scale from “single molecules” to “molecular systems at the cellular level.”

Qiushan Guo, President of the Shanghai Biomedical Innovation and Translation Fund, stated: “The development of macromolecular drugs has long been constrained by the tediousness of traditional screening and the step-by-step error amplification of previous AI toolchains. BioGeometry has achieved atomic-level precision modeling of biomolecular interactions and created an integrated closed loop of structure prediction, sequence generation, druggability evaluation, and wet-lab feedback. This represents an AI-native path closer to underlying scientific logic and real industrial needs. This all-atom de novo design philosophy grants the company an unparalleled generational advantage in highly difficult pipelines—such as traditionally undruggable targets, complex antibodies, and multispecific macromolecules—and has enabled the delivery of Preclinical Candidate (PCC) level molecules. We look forward to BioGeometry rapidly advancing the clinical development of its proprietary pipelines and expanding global collaborations, driven by its rapidly iterating, independently controllable GeoFlow algorithmic foundation.”

Kun Zhang, Head of the Smart Healthcare Team at CAS Investment, remarked: “AI-driven drug R&D is poised to break the industry’s ‘Eroom’s Law’ dilemma, propelling macromolecular drug development into a new paradigm of ‘structural understanding, targeted design, and dry-wet lab closed-loop validation.’ We highly recognize the technical capabilities and global influence of the team led by Professor Jian Tang in the AI4S field. Their proprietary GeoFlow model has already demonstrated differentiated technological advantages in de novo design scenarios for antibody drugs and industrial enzymes. We believe that, against the backdrop of AI reshaping drug discovery, BioGeometry will deeply empower innovative pharma companies and the biomanufacturing industry, accelerating pipeline translation and commercialization, and continuously delivering industrial value.”

Dr. Dakui Wang, Managing Director of Fortune Capital, stated: “The ‘intelligence emergence’ moment for AI in the biomedical field has arrived faster than the industry expected. AI can design and screen candidate molecules almost infinitely, shifting drug discovery innovation from traditional trial-and-error to computation-driven. BioGeometry is a top-tier AI4S team led by Professor Jian Tang, deeply rooted in biological computing for years, possessing both solid academic accumulation and the capability to translate cutting-edge algorithms into engineering solutions. The team’s proprietary all-atom biomolecular world model, GeoFlow, can highly accurately predict the structures and interactions of biological macromolecules like proteins. Its technical capabilities are in the top global tier and serve as a crucial breakthrough to shatter the monopoly of related overseas closed-source models. Coupled with the domestic advantages of low costs and rapid iterations in wet-lab stages, Chinese AI drug discovery companies, represented by BioGeometry, are fully capable of achieving latecomer catch-up.”

Wenjue Li, Partner at Link-X Capital, noted: “Life science is entering a brand-new era: transitioning from relying on experience and serendipitous discovery to precision innovation driven by computation and design. With generative AI as its engine, BioGeometry explores the programmable design of proteins—the underlying language of life. Through the closed loop of dry and wet labs, it continuously accelerates model iteration and experimental validation, enhancing the efficiency and success rate of new molecule discovery and functional design. We are highly optimistic about BioGeometry’s systematic accumulation in AI foundation models, protein design capabilities, and experimental validation systems, as well as the long-term innovative potential demonstrated by its globalized, cross-disciplinary team. We expect BioGeometry to continuously drive the deep integration of AI and life sciences, opening up a more efficient, predictable, and highly engineerable new paradigm for biomedicine and synthetic biology.”