The national team has entered the field of AI virtual cells, and 'Baiyao Technology' has completed a new round of financing worth tens of millions of yuan | 36Kr Exclusive
Text by | Hu Xiangyun
Edited by | Hai Ruojing
36Kr learned that the AI Virtual Cell (AIVC) platform company "Baiyao Technology" has recently completed a new round of financing worth tens of millions of yuan. This round of financing was led by Guoxin Venture Capital Fund under the State-owned Capital Operation Platform of China Guoxin, with Daotong Capital and Yunqi Capital following suit. Old shareholders Fengrui Capital and Baidu Ventures also made additional investments. The raised funds will be mainly used for the iteration of a new generation of virtual cell algorithm models, the construction of an exclusive data collection platform, and the acceleration of industrialization.
Baiyao Technology is an enterprise that 36Kr has been continuously tracking and reporting on (Link attached). The company's founding team relies on the scientific research resources of the Institute of Zoology, Chinese Academy of Sciences, and the Beijing Institute of Stem Cell and Regenerative Medicine. Since 2023, it has started to build a basic large model for AI virtual cells to achieve accurate analysis and prediction of cell behavior and state changes, and continuously promote the upgrade of the AIVC model and the practice of industrialization.
Since 2026, the enthusiasm for financing in the AI pharmaceutical track has been continuously high, and the industry's focus is gradually shifting from the early single-point breakthrough in molecular screening to the deeper cell-level life system modeling and dynamic simulation. Different from AI drug discovery that focuses on molecular configuration optimization, the value of AIVC can cover the entire life cycle of drug development: from disease mechanism analysis, target discovery, virtual screening of candidate drugs, preclinical efficacy and toxicity prediction, to clinical patient stratification. Its core breakthrough lies in upgrading the AI modeling level from "molecule" to "cell" to directly simulate the dynamic response of the life system to external interventions.
AIVC is trained based on massive single-cell multi-omics data. By learning the laws of cell state transformation, gene regulatory networks, and cross-scale causal relationships, it constructs a foundation model that can characterize cell behavior and fate evolution. Researchers only need to input the initial state of cells (such as induced pluripotent stem cells, immune cells, or tumor cells) and specific perturbation conditions (such as drug action, gene editing, or environmental changes), and the model can predict the future state transformation of cells, including key biological processes such as proliferation, differentiation, aging, apoptosis, and disease progression.
This ability enables researchers to conduct large-scale virtual experiments in a digital environment in advance, prioritize research hypotheses with the highest probability of success, and then conduct precise verification through wet experiments. As a result, the R & D process is gradually changing from the traditional "experiment-driven" to "model-driven", which is expected to significantly reduce R & D costs, shorten the development cycle, and improve the overall success rate of drug R & D.
From the perspective of industrial development trends, AIVC is evolving from a simple R & D tool to a new type of infrastructure in the field of life sciences. Just as large language models have reshaped the way of information processing, AIVC is expected to become a core platform for understanding, predicting, and designing life systems, providing underlying support for the next generation of drug R & D, cell therapy, and synthetic biology. Benefiting from this, ARK Invest once stated in the "Big Ideas" report that virtual cells will be "one of the most disruptive fields in the future of AI + life sciences".
Currently, the flow of funds in the primary market is slowly verifying this judgment. The founding team of Baiyao Technology said that the recognition of this round of financing by the state-owned capital operation platform and market-oriented technology and medical capital will help the company quickly complete a new generation of basic models and perturbation models, and implement the construction of a self-produced data set platform and cooperation on exclusive data sets. On the basis of ensuring the core capabilities of AIVC, it will take the lead in exploring industrial application scenarios, deploying advanced cell therapies and "AIDD plus AIVC", and accelerating the closed-loop of the industrial chain of "model - data - scenario".
It is reported that as early as September 2023, the founding team had released the knowledge-enhanced multi-species cell large model GeneCompass, which was published as a cover article in "Cell Research". This model is a general virtual cell base model with over 100 million single-cell data and knowledge embedding. In 2025, the team further upgraded and released the CellGraphCompass model, systematically applying graph structure algorithms and prior knowledge to virtual cell modeling. In the same year, the perturbation model based on the basic model helped the founding team win the first place in the preliminary round and the second place in the overall list of the final round of the Global Virtual Cell Challenge.
In addition, Baiyao Technology has also made a large amount of precipitation in data. In addition to relying on the large-scale global AI-ready public data set accumulated by the founding team over 4 years to train the basic model, the team will also implement self-produced data set technology in the near future, collect high-quality time-continuous perturbation data sets suitable for AIVC model training, and carry out large-scale clinical data cooperation self-produced data set projects.
In terms of application, Baiyao Technology has promoted the implementation of scenarios such as model deployment, joint R & D of cell therapy pipelines, early target discovery, and precise stratification of clinical samples with domestic and foreign partners.
Investors' views:
A relevant person in charge of China Guoxin said that Guoxin Fund has long been concerned about innovative tracks with both hard-core technological barriers and national strategic value. AI virtual cell technology, as the underlying infrastructure of the biomedical industry, is in line with the strategic deployment of the "AI +" initiative in the country's 14th Five-Year Plan. We are optimistic about the profound technological heritage and continuous innovation ability of the Baiyao team in this field, and look forward to the company's continuous drive of the transformation of the life science research paradigm with AI to release long-term value. As the first angel round investment project of Guoxin Fund, we will fully support Baiyao Technology to open up the entire chain of "model R & D - scenario verification - industrial implementation" and provide key support for the construction of an independent and controllable AI pharmaceutical technology system in China.
Hu Mingyi, a partner of Mingde Capital, the financial advisor, said that AI virtual cells are a major breakthrough of the new generation of generative artificial intelligence models in the field of life sciences, which will enable the academic and industrial circles to have more comprehensive, accurate, and efficient means to simulate the interaction between drug molecules and the complex human life system in the virtual world. On the basis of the current widespread acceptance of AI in the global pharmaceutical industry, this will give rise to more fundamental and powerful business opportunities. We are glad to see that Baiyao has received strong support from the national team fund and professional medical and technology funds, and continues to build data infrastructure and lead continuous breakthroughs in models.
Huang Jun, a partner of Dianshi Capital, the financial advisor, said that virtual cells represent the most imaginative frontier direction in the field of AI for Life Science. Baiyao is one of the very few domestic teams that start from the native underlying architecture and independently build a basic cell model, and is ahead of schedule. Relying on the profound scientific research heritage of the Chinese Academy of Sciences and verifying its strength with international top journal achievements and global competitions, Baiyao has both a deep data asset barrier and leading original algorithm capabilities. We firmly believe that it will grow into a leader in this track.