Dialogue with Yao Su Technology's Xie Xin: When organ chips move toward commercialization, pharmaceutical companies will pay to "reduce R&D uncertainty and failure risk."

Animal experiments have long been the “safety foundation” for new drug R&D.

But with the release of relevant roadmaps by regulatory authorities, all of this is quietly beginning to change.

On April 10, 2025, the FDA issued the Roadmap for Reducing Animal Use in Preclinical Safety Studies (hereinafter referred to as the “Roadmap”), stating that it will gradually eliminate animal experiments in the preclinical safety studies of monoclonal antibodies (mAbs) and other drugs, and instead adopt more new alternative technologies such as organoids.

Now that the policy has been in place for nearly a year, the substantive progress brought by this “shockwave” is clearly visible.

This shift is mainly reflected in two dimensions:

First, regulators have taken concrete actions to accept new technologies.

In January this year, the FDA received a letter of intent for the ISTAND organ-on-a-chip initiative led by the 3Rs Collaborative (3R Alliance). Participants include Axiom/LifeNet Health, BioIVT, CN Bio, Xellar, and many other companies.

Not only that, just two months later, the FDA also released a draft industry guidance titled General Considerations for Industry on Using New Methodologies in Drug Development. It aims to encourage stakeholders to submit experimental data related to new methods such as organ-on-a-chips and organoids to regulators, in order to improve the safety of clinical trials and reduce reliance on animal experiments.

This means that non-animal experiments such as organ-on-a-chip studies have already taken the first breakthrough step within the official standards system as alternative solutions. The application of new technologies is no longer confined to the conceptual stage alone.

Second, competition on the industry side is becoming fiercer, and Chinese companies are starting to come into view.

The ISTAND organ-on-a-chip letter of intent that the FDA received this time includes a range of companies, such as Axiom/LifeNet Health, BioIVT, CN Bio, DefiniGEN, InSphero, Lena Biosciences, Inc., PredictCan, TissUse, and Xellar.

Xellar (Yao Su Technology) is the only Chinese organ-on-a-chip company on this list.

Now, capital is putting its full weight behind this.

Recently, Yao Su Technology completed more than RMB 200 million in Series A financing. The round was exclusively led by the State Life Equity Fund, with existing shareholders JingTai Holding, YaYi Capital, and Junlian Capital continuing to increase their stakes. The funds are mainly used to build next-generation bio-intelligent infrastructure centered on humanized models and mechanism research, including expanding the organ-on-a-chip disease model system, establishing a scalable mechanism-research platform, and continuously conducting high-throughput, standardized real biological data collection and accumulation.

But on the other side of the coin is that, despite how close we may be to the era of truly “AI-driven drug discovery,” the industry still faces a long validation period.

The most direct reality is that, to date, no drug has been developed and successfully launched globally that was fully dominated by AI.

The “early mover” that has been heavily expected is also accompanied by high risk. Previously, Exscientia’s globally first AI-designed molecule, DSP-1181, ultimately failed to live up to expectations in Phase 1 clinical trials and was left by the wayside.

The gap between this ideal and reality fundamentally stems from the inherent complexity of new drug R&D itself and from regulatory policies.

Although the finish line for “AI-native new drugs” has not yet been breached, the domestic companies’ push to break through has already begun.

In February 2025, Haisum Biosciences announced a collaboration with JingTai Technology to complete the first innovative drug pipeline project, SIGX1094, for drug discovery. It officially obtained FDA Fast Track designation in the United States.

Now, progress is starting to show.

On April 2, Haisum Biosciences CEO Zhang Haisheng told Wall Street Insights · All-Day Tech that the first indication for SIGX1094 is gastric cancer. Gastric cancer is a rare disease in the United States. With the double boost of receiving both FDA rare disease designation and Fast Track designation, the drug is expected to complete Phase 2 clinical trials and then apply for approval.

“Also, please expect whether we can become the first drug designed and developed by AI and successfully approved,” Zhang Haisheng said.

From building underlying organ-on-a-chip model infrastructure to pushing innovative pipelines toward later clinical stages, AI is sweeping the industry at an unprecedented pace.

In this wave of paradigm shift, the resonance between regulatory standards and industrialization pace is indeed a key variable in shaping the pharmaceutical R&D landscape.

As the only organ-on-a-chip company in China that has made it onto the FDA letter-of-intent list, Yao Su Technology is not only an eyewitness to this wave, but also a keen sensor of changes in the industry’s “water level.”

Around the progress made by China’s regulatory authorities in the organ-on-a-chip field, and the real competitive moat of Chinese innovative companies in underlying infrastructure, All-Day Tech recently held a conversation with Yao Su Technology’s CEO Xie Xin.

Industrial Resonance Between China and the U.S.

“If regulatory policy had not changed, I believe that even 100 years from now, new drug R&D would still follow the experimental sequence from animals to humans. But fortunately, regulators have begun to allow new technologies to be tried,” said an innovation drug investment professional in Shanghai.

Behind this sentiment is the fact that regulation is accepting the changes brought by AI to new drug R&D.

On April 10, 2025, the U.S. FDA officially released the Roadmap, making clear that animal studies for drugs such as mAbs will be phased out, and that it will fully shift to alternative technologies such as organoids.

This is not an isolated signal across the ocean. Just more than ten days later, China’s NMPA and multiple other departments jointly issued the Implementation Plan for the Digital and Intelligent Transformation of the Pharmaceutical Industry (2025–2030), extracting 41 typical scenarios of technological innovation.

Among them, it specifically mentions “data mining from animal models and virtual animal experiments,” meaning that for issues such as high demand for animal replacement and deviations in consistency with human outcomes, data mining and simulation technologies can be used to establish computer simulation models for animal disease modeling.

Against this backdrop, companies such as Yao Su Technology—along with other organ-on-a-chip firms—have entered this real-world industrial test.

In June 2025, the FDA took the lead in launching a joint validation of drug-induced liver injury (DILI) across a global network of nine organ-on-a-chip platforms, including Yao Su Technology.

As the only Chinese company on this list, Yao Su Technology’s core business relies on the underlying platform of “organ-on-a-chip + AI.” It focuses on specific disease models or R&D pain points, and conducts platform-driven joint R&D with global pharmaceutical companies.

In Xie Xin’s view, the performance of Yao Su Technology’s platform in predicting drug responses shows stability and repeatability. It can produce consistent results across different batches and conditions, and it can provide complete, standardized, and traceable raw data and technical documentation—these are key reasons why it can be selected for FDA organ-on-a-chip validation projects.

All-Day Tech: What differences exist between China and the U.S. in the pace and focus of advancing organoid standards? What are the biggest concerns and core risk assessment points for regulators in each place at this stage?

**Xie Xin: **Overall, China and the U.S. show characteristics of “similar pathways but different pace and emphasis.”

The U.S. FDA’s push is more systematic and has a clearer roadmap. Using mechanisms such as iSTAND as the core, it gradually builds a complete framework from technical validation to qualification recognition (DDT). Its emphasis is on promoting the application of new methodologies in specific drug development scenarios through multi-center validation, standardized data requirements, and a clearly defined “Context of Use.”

In other words, the U.S. places more emphasis on whether the technology is sufficiently reliable in a clearly defined use scenario so that it can be incorporated into decision-making systems.

By contrast, China’s CDE is currently in a stage that is more focused on “consensus building,” emphasizing broad absorption of opinions from academia and industry. It refines standards step by step through guidance discussions and special seminars.

In terms of emphasis, China is more concerned with the applicability of technology pathways and industrial feasibility. Under the premise of ensuring scientific rigor, it seeks implementation approaches that align with domestic R&D systems.

At present, the regulators in both places share two core concerns: first is repeatability and consistency—whether stable and comparable results can be obtained across different batches, experimental conditions, and even different laboratories; second is relevance to human outcomes—whether the model truly improves the ability to predict clinical outcomes, not just “looks more complex.”

On this basis, the U.S. FDA is more focused on whether it can be trusted in specific use scenarios. Therefore, it will重点评估 data integrity, standardization, and performance under multi-center conditions. Meanwhile, Chinese regulators are relatively more focused on “how to implement it within the existing system,” including practical operability of technology, costs, and industrial supporting capabilities as real-world factors.

All-Day Tech: What stimulative effects did the FDA’s Roadmap bring to your performance and financing?

**Xie Xin: **Since 2023, we have been working with the FDA and related 3Rs organizations, and we have gradually entered multi-center validation and the iSTAND review framework. This has driven our platform capabilities to evolve continuously into “tools that regulators can accept.” When the policy clearly encourages alternative technologies, and we are also within the regulatory validation pathway ourselves, customers’ understanding of our technology shifts from “frontier exploration” to “a solution with forward-looking regulatory value.”

From a commercialization standpoint, the depth of collaboration and the fee structure have both changed to some extent.

In the early stage, pharmaceutical companies more often start with exploratory, small-scale projects to test whether the technology is feasible. But over the past year, as regulatory signals have become clearer and continuous validation of the data we deliver has progressed, the collaboration model has begun to evolve to a deeper level.

For example, after some global pharmaceutical companies complete initial projects, they no longer treat organ-on-a-chip as merely a “supplementary experiment.” Instead, they incorporate it into standard evaluation systems within specific R&D workflows, and conduct ongoing collaborations centered on specific models or application scenarios.

These collaborations are typically no longer one-time validation projects. Rather, they focus on data production in stages and across multiple batches, along with mechanism research, reflecting a stronger willingness to invest long term.

In other words, customers are no longer paying solely for “technology trials,” but for “reducing R&D uncertainty and failure risk.”

All-Day Tech: Why was the company selected as a partner by the FDA? Does this, to some extent, also help the company build competitive barriers?

**Xie Xin: **The core reasons are twofold: first, our platform’s performance in predicting drug responses is stable and repeatable, and it can output consistent results across different batches and conditions; second, we are able to provide complete, standardized, traceable raw data and technical documentation—this is especially critical for regulators.

Regulatory pathways themselves do form a barrier built from time and trust accumulation. Even if later entrants have similar technical capabilities, they still need a longer time to complete the same validation and alignment process.

All-Day Tech: Compared with other organ-on-a-chip companies at home and abroad, what is your most irreplaceable differentiating competitive advantage?

**Xie Xin: **Our main advantage still lies in building a bio-intelligence system with 3D Bio Intelligence at its core, featuring a dry-wet closed-loop approach, and demonstrating competitiveness across three dimensions: regulatory recognition, industrial validation, and engineering capability.

In the regulatory pathway, from the start we designed the platform around the new drug R&D and submission system, and we deeply participated in advancing relevant FDA standards and the new methodologies (NAMs) framework. Regulatory alignment capability, in essence, determines whether a technology can move from “usable” to “adopted.” This is one of the most critical—and hardest—thresholds for the entire industry.

On the industrial validation side, we have already established in-depth collaborations with multiple global top pharmaceutical companies and have continued to deliver high-quality data and research results in real R&D scenarios. By continuously accumulating high-quality human-derived data and application scenarios, we are forming a long-term competitive advantage driven by real-world data.

In terms of engineering and system capabilities, we have advanced our organ-on-a-chip platform from research-grade capability to industrial-grade stable operation. For example, in high-throughput experimental systems, we achieve a CV below 10% and a Z’ factor above 0.5. This means the platform has reached standards for scalable applications in consistency, repeatability, and signal discrimination—providing the foundation for stable large-scale data production and model training.

Currently, through deep integration of AI with human-derived biological models, we have formed a positive feedback loop of “data generation—mechanism parsing—model training—experimental validation,” enabling the platform to continuously self-evolve.

In the long term, our goal is not only to provide tools or services, but to build a data and computational infrastructure centered on human biology, driving new drug R&D to move from experience-driven approaches to mechanism-driven approaches.

All-Day Tech: AI algorithms and data quality are inseparable. Where does your company’s current data mainly come from, and does it have exclusivity?

**Xie Xin: **It mainly comes from the 3D wet-lab experimental data generated by our own organ-on-a-chip platforms—this is the company’s most important and most exclusive asset base.

Unlike relying on publicly available databases or data sources from a single in vitro model, our data is generated continuously through standardized, repeatable, and quantifiable experimental systems under conditions that are closer to real human physiological environments. It has a clear source, complete process records, and traceability.

These data include multiple cell types, dynamic fluidic environments, and multi-dimensional functional readouts. They not only cover traditional indicators, but also high-dimensional dynamic physiological information such as metabolic activity, barrier function, and mechanical signals.

Actually, the value is not only “result data,” but also “process data” closely tied to biological mechanisms, which can support deeper mechanism parsing and model training.

In addition, we also selectively integrate and clean public databases as supplemental data sources to enhance the models’ generalization capabilities.

But overall, the core competitive strength still lies in the human-derived high-quality data continuously generated by our proprietary platforms. This type of data has inherent non-replicability: on the one hand, it comes from complex engineering systems and experimental capabilities; on the other hand, as projects progress, an advantage accumulates by compounding time and scale.

Capital operations are a means, not an end

With regulators becoming clear about an open attitude toward organ-on-a-chip technologies, domestic pharmaceutical companies are gradually increasing their acceptance of new technologies.

As Xie Xin explained to All-Day Tech, domestic pharmaceutical companies typically start with single-project collaborations. After validating the value of the technology, they move toward long-term framework collaborations. Currently, Yao Su Technology has already signed long-term cooperation framework agreements with multiple leading domestic and international biotech companies.

On the “B side” of the acceleration of commercialization, Yao Su Technology’s planning for an IPO has also drawn attention.

On this, Xie Xin believes that capital operations are a means, not an end. The specific timeline still needs to be judged comprehensively based on the company’s development stage and the market environment.

All-Day Tech: Focusing on the domestic market, how receptive are Chinese pharmaceutical companies to organoid technologies? What pain points and concerns do they generally face when introducing the technology? In terms of commercial implementation, what characteristics do you see in local pharma companies’ true willingness to pay and the procurement decision cycle for this cutting-edge technology?

**Xie Xin: **Acceptance is indeed increasing, especially among innovative drug companies, whose demand for IND filing support is urgent.

The pain points they commonly face are doubts about the credibility of data from new technologies, as well as the complexity of internal validation processes.

Based on what we observe regarding commercial implementation, local pharmaceutical companies’ willingness to pay shows a pattern of “start with pilots, then build platforms.” In other words, they typically begin with single-project collaboration, and after validating the technology’s value, they shift to long-term framework collaborations.

At present, we have signed multi-year cooperation agreements with multiple leading domestic biotech companies, and procurement decision cycles have also shortened significantly.

All-Day Tech: What are the company’s main business models at this stage? Is it more of a CRO model providing technology services? Are your main customers pharmaceutical companies, or CXO companies? Could you break down the company’s current revenue structure in detail?

**Xie Xin: **In essence, Yao Su Technology is not a traditional CRO company. It is a bio-intelligence platform company centered on 3D Bio Intelligence.

Organ-on-a-chip and AI are our technological foundation. Our goal is not to provide one-time services, but to build platform capabilities that continuously generate data, parse mechanisms, and drive discovery.

At this stage, our business model is mainly reflected in platform-driven collaboration models.

Specifically, we conduct joint projects with pharmaceutical companies around specific disease models, organ systems, or R&D issues. Through organ-on-a-chip and AI capabilities, we provide high-quality human-derived data and mechanism-research support.

These collaborations may take the form of project revenue, but in essence they are closer to technology collaborations based on platform capabilities, rather than a traditional service model charging for one-time experiments.

In terms of customer structure, our core customers are mainly from the industrial side, including global pharmaceutical companies and biotech companies, and they also extend to fields such as cosmetics, food, and nutrition and health.

A common point among these customers is that they are all seeking evaluation and mechanism-research tools that can more closely reflect real human responses.

In addition, we also collaborate with regulatory authorities and research institutions to jointly advance the development of technology standards and methodologies.

From the perspective of revenue structure, currently it mainly includes three categories: first, platform collaboration revenue tied to specific R&D questions—joint research projects based on organ-on-a-chip and AI; second, continuing data and model service revenue that gradually forms as collaborations deepen, such as multi-batch studies and data accumulation around specific models; third, platform-based licensing and longer-term R&D collaboration models that we are progressively exploring—for example, deeper collaborations around disease models or mechanism research.

In the short term, we focus on platform collaborations, quickly entering real R&D scenarios and accumulating data. In the mid-term, we will gradually extend toward scalable outputs of model platforms and data capabilities. In the long term, based on continuously accumulated human-derived data and AI capabilities, we will evolve toward mechanism research, new target discovery, and the development of our own pipelines.

From the surface, our business model may appear somewhat similar to CROs, but the underlying logic is completely different. We are not delivering one-time experimental results; we are building a bio-intelligence system that can continuously create value and keep evolving.

All-Day Tech: Beyond pharmaceutical R&D, we also see your technology extending into scenarios such as skincare product screening. What is the core logic behind these cross-industry initiatives?

**Xie Xin: **In essence, what organ-on-a-chip and AI solve is the same problem: how to more realistically reproduce human physiological responses in vitro, and convert that into computable and analyzable data.

This capability naturally has cross-industry attributes. Whether it is drug development, skincare evaluation, or food and nutrition and health, fundamentally it is answering the core question of “the mechanism and safety of a certain substance in the human body.”

These cross-industry applications will, in turn, strengthen our core capabilities. Different fields bring different types of biological stimuli and phenotypic data, enabling us to accumulate human-derived data under a wider range of conditions and continuously enrich the AI models’ understanding of human biological systems. This “multi-scenario data input—unified model learning” process continuously improves the platform’s generalization ability, and ultimately feeds back into the core scenario of drug development.

From a business perspective, on the one hand, it can achieve rapid revenue and data accumulation in relatively shorter application cycles; on the other hand, it can reduce dependence on a single industry’s cycle, making the company more stable across different market environments.

Overall, we do not think this is “cross-industry.” It is a natural extension of the same underlying capability into different scenarios.

All-Day Tech: Do you have any timeline plans related to an IPO?

**Xie Xin: **I always believe that capital operations are a means, not an end.

At this stage, our core task remains to continuously strengthen technology barriers, deepen collaboration with global customers, and accumulate high-quality human-derived data assets—so the company has the capability to create long-term, stable value. When the company reaches industry-infrastructure-level maturity across three dimensions: technology, business, and regulation, the capital markets will naturally provide the answer.

Of course, we are also actively making the relevant preparations and arrangements. In the right time window, we hope to help the company move toward the capital markets and become a benchmark listed company in the 3D Bio AI field, injecting more confidence and certainty into the entire industry. But the specific timeline still needs to be judged comprehensively based on the company’s development stage and the market environment.

Risk Warning and Disclaimer Terms

        The market involves risk; investment should be approached with caution. This article does not constitute personal investment advice, and it does not take into account any specific investment objectives, financial situations, or needs of individual users. Users should consider whether any opinions, viewpoints, or conclusions in this article are appropriate for their specific circumstances. Invest accordingly at your own risk.