2228 JingTai Technology Visit Record and Thoughts 01

$XTALPI(02228.HK)

Regarding last week's interview, I plan to write 3 to 4 paragraphs, and today I will first publish the first paragraph. The community's concerns and responses are still waiting for JingTai IR to organize and provide the final information (as a listed company, we cannot affect or violate relevant compliance regulations, please understand). Today, I will write down my thoughts and some observations from the process:

To state upfront, the sole purpose of organizing this thought is to help me decide whether this company is worth holding for 3 years, or even longer? It can be said that holding a stock for a long time is, from a certain perspective, a painful choice. I have experienced this several times. Should I take the leap again this time?

Design of the Business Model:

A company's valuation is greatly related to its cash flow, and the source of cash flow largely depends on the design of the business model. Before going, I wasn't very clear on how JingTai makes money, but through communication with the company, I can now see some interesting phenomena:

• Large long-term strategic cooperation contracts, such as those signed with Pfizer/Eli Lilly, include some long-term service contracts. The service content mainly involves AI-based research and development services, including polymer design, preliminary compound validation, etc. This involves specific professional vocabulary, and I find myself at a loss for words. Here, large contracts are broken down into different smaller projects, and each small project will have some one-time payments associated with different milestones. However, I suspect that the amount of these one-time payments is not too large at the moment. For example, the major strategic contract with Pfizer has an intended amount of USD 250 million. However, based on the one-time fees already collected, the amount is still not too significant.
• At the same time, in the research and development cooperation with the aforementioned large overseas pharmaceutical companies, about 40% of the contracts have a long-term revenue-sharing model after the drug is launched. If the drug developed through the services provided by 2228 is successfully launched, then within a certain period, say 5 years? 10 years? This depends on the specific contract, a portion of the drug sales revenue will be shared with 2228. This model is certainly good. However, the problem is that the drug launch cycle is too long; in the short term, I estimate it will be difficult to quickly provide scalable revenue within 2 years.

2228's starting point is from drug development, but the biggest problem with drug development is that the time cycle for generating large cash flow is too long, often on the order of 10 years. Therefore, in addition to the aforementioned revenue models, 2228 is currently entering the following business revenue schemes:

• Research and development of new materials, especially a polymer material, based on the required polymer characteristics such as solubility, water absorption, etc. They aim to quickly realize the design of related materials and even enter production and sales based on their algorithms for constructing polymer materials. During this visit, I learned about some product lines that I believe have significant potential for scalable revenue, although this still needs some time for validation. If possible, conducting on-site investigations for future production/validation might be a better choice
• The research and development of beauty and cosmetic materials has a much shorter market launch cycle compared to pharmaceuticals, but the essence of the technology required is similar. During this communication, we also saw some products that are about to be launched in the short term (within 6 months).

At the same time, they are also selling their integrated workstation (robots, a type of high-precision chemical experimental verification robot). Based on the advantages of the supply chain in the Greater Bay Area and practical chemical experiments, 2228 has developed a complete set of scalable chemical experimental robots (workstations). This part of the sales is also underway.

Comparing the latest quarter's revenue sources, we can see that the aforementioned business model is reflected in the revenue:

• Strategic Cooperation: Jingtai Technology signed a 5-year strategic cooperation agreement with Xiexin Group, with a total amount of approximately USD 135 million (about RMB 1 billion). Jingtai will provide Xiexin with research and development services for new energy materials such as perovskite, supramolecular, lithium-ion batteries, cathode materials, and carbon-silicon materials, and jointly create an AI+ automated intelligent creation system driven by large models in the materials field.

• Drug Discovery Solutions: Revenue of approximately RMB 60.85 million, a year-on-year increase of about 68%, mainly due to an increase in the number of clients and revenue-generating projects.

• Intelligent Automation Solutions: Revenue of approximately RMB 41.78 million, a year-on-year decrease of 4.8%, mainly due to a decrease in solid-state research and development service revenue, partially offset by an increase in revenue from automated chemical synthesis services.

The Evolution Path of This Company's Technical Capabilities and Competitive Barriers

Regarding whether to invest in this company long-term, "technical capabilities" and the corresponding barriers built are what I care about the most. The focus of this communication was also to understand this process, and more side insights are needed to verify authenticity and sustainability. To briefly explain how this company develops and builds competitiveness:

• The background of the founders of 2228 is impressive. Three physicists from MIT (at the current microscopic research level, physics and mathematics are the basis for determining chemical properties, while the macroscopic verification of chemical substances is more of an experimental issue) stand out, which can be seen from the numerous excellent institutional investors in the company. The founders' professional research background, along with the opportunities for scaling at the microscopic level through AI algorithms and models, has built the core competitiveness of this company in its first stage (founding stage). Simply put, if you have a need to construct a new type of polymer (whether it is a drug or a material), you can state your requirements, and 2228's AI model can efficiently help you calculate and predict, providing molecular material models that meet your requirements among tens of thousands of possibilities, and importantly, significantly shortening the time.

• With the support of the aforementioned AI algorithm models, exploring new types of polymer materials has become a low-cost endeavor. However, once the mathematical models of polymers are available, verification is needed, which involves the synthesis of molecules and the verification of related chemical and physical properties, such as stability, etc., while also considering related costs. I understand that this is quite cumbersome, and initially involves a lot of manual processes. During my visit, I saw many experimental personnel using beakers and measuring cups to conduct practical experiments In this process, humans are indispensable, but human operation poses a significant problem; the efficiency, stability, and measurability of human operations are quite low. During the conversation, a simple example was mentioned: for instance, in a certain synthesis process, whether the mixture in the measuring cup is stirred two times or three times can affect the synthesis itself. However, when humans operate, it’s hard to remember whether it was two times or three times. 2228 has leveraged the supply chain advantages in the Greater Bay Area to build a complete set of workstations that cover everything from algorithms to operations and assembly line requirements (which they refer to as robots; I was initially puzzled about what kind of robots this company produces). This workstation can replace humans to conduct large-scale, continuous synthesis experiments, verifying as many possibilities as possible and accurately collecting experimental process data, with as much automation as possible. I believe this process is crucial because it can generate a vast amount of data needed for high polymer synthesis processes. The scale advantage of this data will strengthen 2228's advantage in algorithm models in the future, creating a positive cycle. The role of data scale advantage in AI models has been widely recognized in the industry. I also discussed this with a professional in bioengineering (a good friend of mine, a Harvard PhD and Tsinghua University professor), who indicated that this data scale advantage could be crucial for 2228. Moreover, it may have its leading and unique aspects globally (I will also communicate with friends at Pfizer in the U.S. to seek more validation; understanding the core competitive advantage is essential).

• From a certain perspective, 2228 is a data company for polymer materials, following a process of model/data/model iteration—engineering practice—drug/material production validation to realize its value. The starting point is its core competitiveness: the cycle of model—data—model iteration.

  • Frankly speaking, I have high expectations for this. If one wants to confirm a long-term investment in this company (looking forward to the realization of long-term alpha), this must be the foundation of all stories.

• The core commercial value of 2228 lies in the large-scale time savings, with a clear case: Signet Pharmaceuticals, in which 2228 invested, utilized its model data advantages to help the founder of Signet transform his research paper into drug practice in a very short time: