Crystal Pharma's HK$20 Billion Market Debut: How Is China's AI Drug Discovery Sector Competing Globally?

On June 13, XtalPi (2228.HK), the first AI pharmaceutical company to go public, officially listed on the Hong Kong Stock Exchange at an issue price of HK$5.28 per share, raising approximately HK$896 million in net proceeds. On its first day of trading, XtalPi opened at HK$5.39 per share and closed at HK$5.80 per share, up 9.85%, with a total market capitalization of HK$19.759 billion.

This is also the first initial public offering (IPO) under the Hong Kong Stock Exchange's Specialized Technology Companies Listing Rules (hereinafter referred to as "Chapter 18C") since its launch over a year ago. According to Chapter 18C, specialized technology companies must secure a substantial amount of investment from seasoned independent investors before going public.

This time, XtalPi received support from five leading senior independent investors and eight globally renowned institutional cornerstone investors, and also attracted nearly 80 global investment institutions to participate in anchor investments. The international placement orders were oversubscribed by 2.13 times. It is also the only listed company with over 100 times HKPO subscription among issuances of more than 50 million US dollars in the past two years.

"Looking at the entire AI pharmaceuticals sector, 'In 2023, the segment that received the most funding among U.S. biopharmaceutical companies was AI pharmaceutical companies, and this trend of high-value and large-scale financing has continued into 2024. The attention and popularity that AI pharmaceuticals will receive in the future will not diminish. However, with the development of the industry, the bubble in this field will gradually be eliminated,' a securities analyst told the 21st Century Business Herald.

The role of AI in the fields of capability and pipeline development is becoming increasingly prominent, and it has been continuously sought after by capital and markets in recent years. "Technology is only an auxiliary tool; the essence of AI pharmaceuticals is still drug research and development," an investor told the 21st Century Business Herald. "Whether the technology is reliable or not must be proven by results, and ultimately, the competition comes down to who can first deliver a mature product."

From Small Molecule Drugs to Beyond

XtalPi was founded in 2015 as a research and development platform enterprise driven by quantum physics, AI, and robotics. According to the prospectus, XtalPi's revenue reached RMB 62.8 million, RMB 133 million, and RMB 174 million in 2021, 2022, and 2023 respectively, showing a year-on-year increase with an annual compound growth rate of 66.7%.

While revenue is on the rise, losses cannot be overlooked. From 2021 to 2023, XtalPi's operating losses were RMB 299 million, RMB 525 million, and RMB 722 million, respectively. Due to anticipated costs and expenses related to the implementation of commercialization plans (especially in the U.S. and Europe) and increased share-based payment expenditures, XtalPi stated in its prospectus that it expects to continue incurring losses in 2024.

But the high customer retention rate also, to a certain extent, verifies XtalPi's capabilities and prospects. From 2021 to 2023, the company's customer retention rates were approximately 67.5%, 51.4%, and 64.9%, respectively. Returning customers include Pfizer, Johnson & Johnson, Chia Tai Tianqing Pharmaceutical, Daewoong Pharmaceutical Co., Ltd. of South Korea, and Merck Group of Darmstadt, Germany.

For AI-driven pharmaceutical enterprises, a high customer retention rate requires strong underlying technical capabilities. XtalPi's main businesses include drug discovery solutions and intelligent automation solutions. Specifically, the drug discovery solutions cover the entire process of drug discovery and research, providing modular independent solutions or collaborating with various biotechnology and pharmaceutical companies and academic institutions for novel drug discovery efforts; the intelligent automation solutions mainly include solid-state research and development services and automated chemical synthesis services.

Among them, the drug discovery solutions achieved revenues of RMB 39.346 million, RMB 87.666 million, and RMB 87.728 million during the reporting period, accounting for 62.7%, 65.7%, and 50.3% of total revenue, respectively. XtalPi stated in its prospectus that most of its revenue comes from small molecule drug discovery.

According to the prospectus, XtalPi has achieved a 100% success rate in all crystal structure prediction projects for small molecules, which can be considered as the ability to correctly predict the thermodynamically stable experimental form of crystal structures using typical small molecule drug calculation methods. For comparison, according to Frost & Sullivan, the average industry success rate for crystal structure prediction ranges from 86% to 93%.

At this stage, XtalPi plans to leverage its technology, experience, and expertise in small molecule drug discovery to explore more therapeutic modalities, such as PROTAC, ADC, peptides, and RNA. From 2021 to 2023, XtalPi's R&D expenditures were 213 million yuan, 359 million yuan, and 480 million yuan, respectively, accounting for approximately 52.4%, 53.5%, and 49.8% of operating expenses, and representing about 338.5%, 269.2%, and 275.6% of annual revenue, respectively.

"The entire AI pharmaceuticals industry is expanding from single small molecules to emerging drug types such as ADCs, antibody drugs, cell therapy, and gene therapy. This is a trend," the aforementioned securities analyst explained. "Because the discovery of small molecule drugs originated earlier and is more mature, and there is a larger accumulation of small molecule drug data, it represents a more classical approach to drug development, while also facing more typical and prominent bottlenecks."

"Now, with the increasing amount of data in antibody drugs, peptide drugs, cell and gene therapies, etc., the industry also hopes that AI-driven drug discovery based on data and algorithms can empower these fields. Overall, making some attempts is the right approach, but many times these attempts cannot be solved by AI alone," the analyst added.

Multiple Challenges Remain to Be Overcome

The traditional drug research and development process is costly and time-consuming, typically requiring at least about 10 years and an investment of over 1 billion US dollars. In particular, discovering a drug usually takes about one to two years and requires an investment of approximately 400 to 450 million US dollars. Moreover, during the drug discovery phase, it is generally necessary to select a commercially viable drug from thousands of compounds. The total cost of developing a new drug can reach 2.6 billion US dollars.

However, artificial intelligence has been applied to various stages of the drug research and development process, significantly reducing the time and cost required for drug development while increasing success rates. Feedback from drug development helps improve the functionality of AI-powered drug discovery platforms and enriches AI databases. During the learning and validation processes, algorithms, computing power, and data—the three core elements of artificial intelligence—continuously enhance AI-powered drug discovery platforms.

But at present, the data dilemma is a challenge that AI pharmaceutical companies need to face. "Currently, data regulation in China is still relatively strict, and enterprises face significant challenges in utilizing data," the aforementioned securities analyst explained. "At present, data from hospitals in China are not only inaccessible to enterprises, but also not interconnected between each hospital. Moreover, these data involve many issues, including whether the ownership and usage rights of the data belong to the hospitals or the patients."

The analyst added, "If the overseas data is not open to Chinese companies, then it's possible that the amount of data we can access will decrease, and we may find ourselves in a less advantageous position in terms of training and optimizing models. But for now, it seems unlikely to reach that point. Restrictions on biomedical data are not expected at this time, though there could be some limitations on data related to human genes."

Moreover, the birth of a new drug involves not only solving scientific research problems but also many other processes where AI cannot "flex its muscles," such as quality and cost control in the laboratory. The industry also seems to be working hard to expand the application of AI to later stages, including clinical trial design and data interpretation.

"From the current progress, the strength of AI pharmaceutical companies still lies in the front end, that is, in the early stage of drug discovery, from target discovery to the nomination of clinical candidate compounds," the analyst believes. In the backend, especially after entering the clinical stage, there are actually not many areas where AI pharmaceuticals can empower or significantly accelerate the process. For now, clinical trials still need to be conducted following traditional drug development methods.

"Currently, the fastest-progressing AI pharmaceutical projects globally are in Phase 2 clinical trials. Based on this, it can be inferred that it will take at least three years for AI pharmaceuticals to achieve conceptual validation," the aforementioned analyst pointed out. "On the other hand, AI pharmaceuticals are exploring more diversified commercialization models. For example, some companies focus on certain preclinical segments with AI+CRO, some companies provide preclinical drug discovery services, and others generate revenue by out-licensing preclinical or early clinical pipelines, thereby achieving commercialization of AI pharmaceutical companies rather than the commercialization of drugs developed by these companies."

In the analyst's view, the entire AI pharmaceuticals industry needs a genuine project that can achieve clinical proof-of-concept to boost confidence across the sector. "As an increasing number of AI-driven pharmaceutical projects are entering clinical trials, none have successfully reached the stage of clinical proof-of-concept. Since 2023, there have been several failures. The industry is awaiting a successful clinical proof-of-concept result to elevate AI pharmaceuticals to the next level."

Surpassing the Globe

According to data from灼识咨询, currently, the largest market for global AI drug discovery is in North America, with the Asia-Pacific region ranking third. The market size in 2018 was $51 million, and it is expected to reach $1 billion by 2025, with a compound annual growth rate (CAGR) of 53.8%. The global market size is projected to reach $3.68 billion by 2025.

"AI pharmaceuticals industry: Comparing China and the US, there is still a gap in progress." The aforementioned securities analyst stated, "In terms of establishment time, international AI pharmaceutical companies (small molecules) were mostly founded between 2012-2014, including Exscientia, Recursion, Atomwise, Relay, etc., while in China, XtalPi was one of the earlier established AI pharmaceutical enterprises; most were founded after 2018. However, each company possesses unique technology and is developing rapidly."

"Foreign pharmaceutical companies have a relatively high level of recognition for AI drug discovery companies. They are willing to pay higher upfront fees for strategic cooperation, and they are also willing to spend large sums to purchase AI products and pipelines. In comparison, China is somewhat lagging behind and is still in an exploratory stage," the brokerage analyst explained.

For example, in January 2022, Exscientia plc entered into a strategic research collaboration with Sanofi to develop an AI-driven pipeline of up to 15 novel small-molecule drug candidates in the fields of oncology and immunology, with a total deal value of up to $5.2 billion. At the end of 2021, Fosun Pharma signed a $13 million upfront payment agreement with Insilico Medicine for AI-driven drug discovery collaboration.

"Generally speaking, China's AI pharmaceutical companies are currently about 1-2 years behind the global leaders, but there are still opportunities to catch up." The analyst emphasized, "From 2018 until now, over a hundred AI pharmaceutical companies have been established in China. Each company has its own unique algorithms and addresses different problems. Under the 'swarm of wolves' tactic, eventually, a few companies will emerge as leaders."

The successful listing of XtalPi has undoubtedly injected a strong boost into China's AI pharmaceuticals industry, but its sustainable commercialization and profitability after the IPO still need to be proven. According to XtalPi's estimation in the prospectus, if calculated based on 100% of the estimated net proceeds from the global offering, it can sustain operations for 59.3 months. Within these five years, how the first batch of AI pharmaceutical companies will compete globally remains to be seen.

Source: 21st Century Business Herald 21 Finance APP Author: Han Liming