"Xunming Biotech" Secures Tens of Millions of US Dollars in Series A Funding with AI Agents "Scaling Up" the Design of Novel Antibody Molecules | Exclusive Report by 36Kr

Text by | Hu Xiangyun

Edited by | Hai Ruojing

36Kr learned that Aureka Biotechnologies recently completed a Series A financing of tens of millions of US dollars. This round of financing was jointly led by Matrix Partners China and Qiming Venture Partners. Existing shareholder Newlight Capital continued to increase its investment, and institutions such as Agentic Ventures participated in the follow - on investment. The raised funds will be used to advance the core R & D pipelines into the critical clinical stage, accelerate the application of the Agentic system technology in antibody autonomous design and molecular innovation, and expand cooperation with global biopharmaceutical companies.

Aureka Biotechnologies focuses on the development of generative antibody drugs. When it completed the seed - round financing in 2023, with the help of its three core technology platforms, namely the self - developed generative intelligence platform, the yeast intracellular automatic directed evolution platform, and the microfluidic single - cell functional screening platform Three Core Technology Platforms, the company had already achieved the goal of "allowing autonomous agents to be deeply involved in drug molecule design and development, rather than simply providing assistance."

"During that period, we proved that we had the ability to screen out molecules that traditional technologies couldn't. How to further scale up this ability became the development direction for the company's next stage," explained Zhao Weian, the founder and CEO of Aureka Biotechnologies.

To achieve this goal, Aureka Biotechnologies built an antibody design intelligent agent with autonomous evolution ability, and combined it with the existing technology platforms to create a data - driven wet - dry closed loop, which changed the data generation paradigm at the underlying level and formed a new model for more efficient antibody drug pipeline development.

First of all, in the most intuitive process, this model forms a "data flywheel" through the combination of "autonomous agents + scientists + high - throughput experimental platforms."

"In the past, in the project initiation stage, human experts had to read literature, collect information, and put forward reasonable hypotheses, which involved a large amount of interaction between human resources and experiments. As a result, the number of breakthrough pipelines that could be advanced in a year was very limited. However, our model has transformed the single 'human - calling - model' mode into the 'autonomous - agent - calling - model' mode. The latter can participate in or even lead the proposal of antibody design hypotheses, while human experts are responsible for supervision and guidance, retaining high - quality ideas and eliminating ineffective ones. Then, through high - throughput experiments for large - scale verification, the pipeline screening efficiency has been increased to 10 - 20 per year. This R & D logic is more coherent and more practical," introduced Zhao Weian.

In this process, Aureka Biotechnologies is also continuously optimizing the functional dimensions, phenotypic accuracy, and throughput levels of data collection, and autonomously generating higher - quality data assets. This enables the company to break through the barriers of traditional scarce data and continuously screen out more differentiated molecules. In Zhao Weian's view, as a data - driven biotech company, this is the core advantage of Aureka Biotechnologies.

Traditional antibody drug design mainly involves screening specific antibodies through animal immunization and then optimizing their pharmacokinetics, pharmacodynamics, and immunogenicity through genetic engineering. It has defects such as low efficiency, poor molecular diversity, and insufficient differentiation. Especially for difficult - to - drug targets, it is very difficult to screen out high - value antibodies.

In contrast, Aureka Biotechnologies has "changed the data production model" relying on its own technology platforms. "Currently, we have established two high - throughput digital biology methods. One is to collect affinity - related data, that is, data related to the binding of antibodies to targets, through yeast automatic evolution. Secondly, considering that some antibodies can bind to targets but cannot be translated into actual efficacy, we have built a microfluidic - based single - cell functional screening platform. By constructing large - scale combinatorial libraries and corresponding mechanism cells, functional data of millions of molecules can be screened in a single experiment."

If we make a simple analogy, the traditional method is like "injecting" experimental mice, while Aureka Biotechnologies uses an "electronic mouse." It realizes molecular iteration and screening optimization through autonomous agents and designs molecules that traditional living organisms cannot produce.

As a result, this not only improves the efficiency by at least three orders of magnitude, but more importantly, it may "achieve R & D goals that traditional methods cannot achieve." Zhao Weian gave an example. For instance, most of the antibody drugs previously approved by the US FDA are inhibitors, which work by blocking target binding. Since affinity is directly related to function, they are relatively easy to develop. However, in the future, the R & D requirements for high - value application scenarios will be more complex (such as the development of agonists). Currently, there are no general design rules in the industry for these, and the design is very difficult.

"We generate data by constructing large - scale combinatorial libraries and combining high - throughput digital biology technologies, and then use agents to learn and decode design principles. After that, we can design molecular characteristics that meet the requirements through model regulation. Therefore, we have certain development advantages in such complex scenarios."

It is reported that currently, this is also "the ability that large pharmaceutical companies value the most." There are actually many ways in the industry to accelerate or batch - provide antibody molecules for pharmaceutical company customers to choose from. However, the problem is that many methods only provide benchmark molecules, which are difficult to meet the customized needs of pharmaceutical companies. "But we can make a difference for them and translate the difference into clinical needs."

Benefiting from this, Aureka Biotechnologies has cooperated with several European and American pharmaceutical companies, established joint projects around complex targets and functional mechanisms, and achieved commercial revenues in the tens of millions of US dollars in the past two years. In addition, focusing on fields such as autoimmune diseases and metabolic diseases, the company has several self - developed pipelines entering the PCC or Pre - IND stage.

Zhao Weian mentioned that in the current market environment, the implementation ability to "create differentiated pipelines and advance them to the clinical stage" is a key factor for biotech companies to gain the recognition of investment institutions, because it means whether the enterprise can bring clear commercial return expectations. At the same time, digital biotechnology is developing rapidly. Whether the team has the ability of continuous R & D and self - iteration to better realize the combination of "IT + BT" is also very important, as it "relates to the company's long - term growth curve."

"We believe that science and technology are in a mutually beneficial relationship. Scientific breakthroughs will guide the focus of new technologies. Just like GLP - 1RA is not only a series of blockbuster drugs but also promotes the industry to rethink the potential of the'metabolism' field, making people realize that improving metabolism can bring such great market value. Then, new technologies represented by autonomous agents will focus on the unsolved problems in the metabolism field, such as whether two pathways can be simultaneously regulated through target design to achieve the effects of 'weight loss without rebound' or 'weight loss without muscle loss'. This is what we want to do, which is to use new technologies to expand the scientific field more widely," Zhao Weian said.