Exclusive Debut｜Xellar Biosystems Completes Nearly RMB 400 Million Series A Financing Rounds, with AI4S and Humanized Data Super-Factory Poised for a Promising Future

Artificial intelligence-driven drug research and development is undergoing a fundamental shift in its underlying driving force: moving from "computation-first" virtual simulation toward infrastructure development centered on "high-quality, high-fidelity human-derived data."

Xellar Biosystems has completed a Series A+ financing round of nearly RMB 200 million. This round saw participation from Guofang Venture Capital, Tsing Song Capital, GF Xinde Investment, Oriental Fortune Capital, and Cedarlake Capital, with existing shareholder Yael Capital continuing to increase its support.

This marks another significant financing round for Xellar Biosystems, following the over RMB 200 million Series A financing round led by China Life earlier this year. To date, Xellar Biosystems has raised a cumulative total of nearly RMB 400 million across its Series A financing rounds, further consolidating its leading position in the artificial intelligence for biology field, which is underpinned by artificial intelligence and organ-on-a-chip technologies.

The proceeds from this round will be primarily used to continuously advance the company's "3D Bio Intelligence" platform, including further enhancements to vertical-specific artificial intelligence models and biological mechanism analysis capabilities, the implementation of high-throughput automated experimental systems, the continued accumulation of multimodal biological data, and the acceleration of Virtual Cell and Biological World Model construction, comprehensively promoting the formation of a new-generation humanized data infrastructure for drug research and development.

Over the past decade, artificial intelligence-driven drug development was once captivated by the "computation-first" paradigm, attempting to overcome the challenges of drug research and development through exponentially growing computing power. However, numerous cases of "high computing power with low output" have demonstrated that the mere stacking of algorithms and computing power cannot bridge the gap between virtual simulation and real biological systems.

The underlying reason is that the key bottleneck constraining artificial intelligence-driven drug development lies in the scarcity of high-quality, high-fidelity human-derived data. An increasing number of industry professionals are coming to recognize that:

Based on this fundamental shift in underlying driving forces, Xellar Biosystems was founded in Boston, United States, at the end of 2021. The company is dedicated to building a next-generation biomedical research and development platform that integrates organ-on-a-chip technology with artificial intelligence and automation. It is the world's first biotechnology startup to leverage organ-on-a-chip combined with high-content three-dimensional cell imaging, computer vision, and artificial intelligence for drug discovery and precision medicine research under the "3D Bio Intelligence" framework.

For a long time, drug discovery has relied on static two-dimensional cell lines and animal models. These data often differ significantly from the complex physiological environment of the human body, resulting in persistently low clinical translation rates and frequent failures in late-stage human clinical trials. When faced with traditional biological data that are noisy, limited in sample size, and highly heterogeneous, artificial intelligence models often find themselves in a predicament of being unable to perform effectively due to insufficient quality data.

Over the past two decades, advances in single-cell sequencing, organ-on-a-chip, and spatial omics technologies have provided us with unprecedented microscopic visibility and data accumulation. Foundation models and world models are now extending beyond language and vision domains into the biological field.

To advance artificial intelligence models from "digital simulation" to "digital twin," the core of competition lies in who can build a "living" data super-factory—that is, a data production capability characterized by three key attributes: authenticity, standardization, and scalability.

By deeply integrating engineering systems, biological models, imaging and multi-omics data, and artificial intelligence computing capabilities, Xellar Biosystems has established a closed-loop system of "data generation, mechanism analysis, model training, and experimental validation."

Through the continuous construction of real-world human biological data systems, Xellar Biosystems provides "ground truth" for biological artificial intelligence models, advancing artificial intelligence from "pattern recognition" toward "mechanistic understanding." In this process, humanized biological systems based on organ-on-a-chip technology, combined with artificial intelligence capabilities, are enabling complex biological processes to move from being "observed" to being "simulable and predictable," thereby significantly improving research and development efficiency.

Within this flywheel, the organ-on-a-chip and automated experimentation platforms continuously generate high-quality human biological data; the multimodal analysis system is responsible for deciphering complex biological mechanisms at the cellular, tissue, and disease-state levels; the biological foundation models continuously learn and understand the operational principles of living systems; and experimental validation feeds new discoveries back into the platform, forming a continuously iterating and self-reinforcing data engine.

Unlike traditional preclinical research pathways that rely on animal models or single-endpoint measurements, Xellar Biosystems' platform is capable of capturing multidimensional signals from cells and tissues during dynamic changes under conditions that more closely resemble the human physiological environment, thereby enabling in-depth analysis of mechanisms of action, toxicity sources, and disease progression processes. This high-dimensional, multimodal data capability allows its 3D Bio Intelligence platform to understand biological systems with greater accuracy and continuously iterate to optimize predictive performance.

At the same time, the company is actively building humanized drug research and development artificial intelligence agents. Unlike traditional analytical tools, artificial intelligence agents are capable of autonomously performing tasks such as experimental design, hypothesis generation, result analysis, and decision optimization based on experimental data, scientific literature, knowledge graphs, and historical research experience. In the future, as the organ-on-a-chip platform continues to generate high-quality human-derived data, artificial intelligence agents will serve as an important bridge connecting experimental systems and computational systems, advancing drug research and development from a "human-driven experimentation" paradigm toward a new paradigm of "intelligent system-assisted research."

Furthermore, Xellar Biosystems will accelerate the construction of Virtual Cells and Biological World Models, comprehensively promoting the formation of a new-generation humanized data infrastructure for drug research and development.

Virtual Cells aim to reconstruct a digital twin in silico capable of simulating physical, chemical, and biological interactions by integrating multi-omics data through artificial intelligence. In simpler terms, they are no longer limited to describing "what is inside a cell," but are able to predict "how the entire signaling pathway, cellular phenotype, and microenvironment will dynamically respond if a drug effect occurs." This transition from descriptive biology to predictive biology represents a critical step in moving beyond the limitations of animal models.

The Biological World Model is an extension of the World Model into the biological domain, seeking to establish a "universal biological engine" within digital space that integrates multiple coupled systems including metabolism, immunity, and neurology. By abstracting and reproducing the complex macro-systems of the human body in vitro, it addresses a key pain point in drug research and development—"systemic effects."

Xellar Biosystems believes that the core of future drug research and development will no longer be a single technology, but rather a new type of research and development infrastructure composed of organ-on-a-chip, automated experimentation, biological foundation models, world models, and artificial intelligence agents working in concert. Within this framework, the organ-on-a-chip is responsible for generating real-world human biological data, the biological foundation models are responsible for understanding living systems, the artificial intelligence agents are responsible for planning and executing scientific decisions, and the world models are responsible for simulating and predicting complex biological processes. The synergistic interaction of these multiple layers of technology will jointly drive order-of-magnitude improvements in the efficiency and success rate of drug research and development.