AI Highlights · The full text is about 2175 words, and it takes 7 minutes to read

1. 2024 is a year of sharp rise in the popularity of artificial intelligence, and the domestic AI pharmaceutical industry has accelerated its evolution, and pharmaceutical companies such as Salubris and Haoyuan Pharmaceutical have entered the game.

2.AI has great advantages in the field of drug discovery, such as virtual screening, small molecule generation models, etc., and the number of AI-driven drug development projects worldwide has surged from 6 in 2010 to 158 in 2021.

3. China is also a hot spot for AI pharmaceuticals, with about 100 related companies in pharmaceutical companies, but most of them are still in the A round or even the angel round and seed round.

4. However, AI pharmaceutical companies face the problem of data silos, with very limited available data and unsatisfactory data quality, which restricts the development of AI pharmaceuticals.

5. Only the leading companies that have established a leading edge can survive after the financing boom, and pure AI pharmaceutical companies are expected to achieve leapfrog fission.

The above content is generated by Tencent's hybrid model and is for reference only

Author / Boiled under the stars

Edit/Spinach's Starry Sky

Typesetting/egg yolk crisp under the stars

2024 is a year of artificial intelligence (AI) popularity, and the Nobel Prize is not only awarded to AI, but also to three scientists who use AI technology to study the structure of proteins. This is both a reward for the existing achievements of AI and an affirmation of the huge potential of this field.

In the past year, it has also been a year of accelerated evolution of the domestic AI pharmaceutical industry, not only Salubris (002294), Haoyuan Pharmaceutical (688131) and other pharmaceutical companies have entered the game, but also the authentic concept stock XtalPi Technology (02228) has successfully landed on the Hong Kong Stock Exchange with the halo of "the first AI pharmaceutical stock", and it is also the first H-share company to be listed under the 18C rule. Another star company, Insilico Medicine, just announced earlier this year that it had raised more than $100 million in Series E financing. With a succession of good news, this still immature track seems to be thriving.

However, in fact, under the cold winter of pharmaceutical financing, a large number of AI pharmaceutical start-ups are actually facing financing difficulties. It can be said that the head company is hot like a flame, and the rest is immersed in the biting seawater.

1. AI solves R&D pain points

As we all know, the development of drugs from R&D to marketing is a very long and costly act. The drug discovery process alone can take about seven years and require an investment of $600 million to $800 million. In the R&D process, there are actually many links that can be assisted by AI.

For example, in the drug discovery stage, AI can be very efficient in conducting virtual screening, greatly improving screening efficiency. Some researchers say that AI models can complete the virtual screening of billions of compounds in a matter of days, which is unimaginable for traditional methods.

In addition, AI-based small molecule generation models have proven to have great advantages in creating complex molecular structures. Generative AI can redesign proteins or small molecule compounds for specific targets, which provides a foundation for the development of precision diagnosis and treatment in the future. Especially in the field of protein design, many AI pharmaceutical companies are lying on this track, and the pioneer Nabla Bio has reached cooperation with AstraZeneca and other giant pharmaceutical companies.

Of course, AI can also intervene in all aspects of the drug development process by screening the subjects with the best response to improve the efficiency of clinical trials, etc., which I will not repeat here. In short, the introduction of artificial intelligence is not hesitating to be a revolution for the biomedical industry.

The cost and cycle of drug development

It is based on these advantages that the practical application of AI in the field of drug discovery has exploded in recent years, and between 2010 and 2021, the number of AI-driven drug development projects worldwide has surged from 6 to 158. And the quality of AI drug discovery is high, with data from the Boston Consulting Group showing that AI can push the overall success rate of drug discovery from 5-10% to 9-18%. By the end of 2023, there are nearly 900 AI pharmaceutical companies in the world.

Second, Chinese companies are frequently present

China is also a hot spot for AI pharmaceuticals.

Benefiting from the activity of pharmaceutical venture capital funds in the past few years, not only many large pharmaceutical companies have participated in it, but also a large number of start-ups have sprung up. Up to now, there are about 100 relatively "pure" AI pharmaceutical-related companies in China. At the end of 2024, the National Health Commission and other three departments also issued the "Reference Guidelines for Artificial Intelligence Application Scenarios in the Health Industry", which further clarified the specific application scenarios of AI in the pharmaceutical and other medical and health fields.

2016-2021 Financing amount of AI pharmaceutical start-ups in China Source: Toubao Research Institute

Specific to listed companies, in fact, many of the companies we are familiar with have already laid out.

For example, Yige Hengrui Pharmaceutical (600276) introduced AI drug design software Makya™ in 2021, and also built a research and development platform called "Hengrui-Lingshu", which can support R&D scenarios such as target discovery. Salubris announced last year that it had established an AIDD team and R&D platform and had delivered four lead compounds. In an investor research event at the beginning of this year, Haoyuan Pharmaceutical said that it had established strategic partnerships with AI pharmaceutical companies such as Delis and Insilico Medicine. Of course, companies such as Huadong Medicine (000963) have also said that they have built artificial intelligence-assisted research and development platforms.

It can be seen that many Chinese pharmaceutical companies are unwilling to miss the feast of AI, but they have limited trust/investment in AI at this stage. Everyone said in unison that the layout had been put into construction, but it seemed that it was not much used, after all, there were already a lot of pipelines in the hands of domestic pharmaceutical companies.

Compared with these large-scale pharmaceutical companies that have already been listed, AI pharmaceutical start-ups with pure pedigrees often realize commercial monetization through out-licensing pipelines.

For example, Insilico Medicine granted the global rights to the USP1 inhibitor ISM3091 to Exelixis in September 2023 and received an upfront payment of US$80 million, and in January this year, it disclosed that it had reached a total license agreement with Menarini for a total of US$550 million. Domestic pharmaceutical companies are also cooperating with AI start-ups, such as Jianxingyuan (600380) won the exclusive rights and interests of Fermion Technology's new analgesic drug FZ008-145 in 2023, and BeiGene (688235) introduced Ensem Therapeutics' differentiated CDK2 inhibitor ETX-197. According to statistics from research institutions, there will be as many as 5 pipeline transactions directly involved by Chinese AI pharmaceutical companies in 2024.

Statistics on the business model of domestic AI+ new drug companies Source: Intelligent Pharmaceutical Bureau, Eggshell Research Institute

Third, data silos need to be broken urgently

While the business of these leading companies is rolling out in an orderly manner, in fact, more AI pharmaceutical start-ups are struggling on the brink of life and death. According to industry statistics, except for a few companies that have gone to the C round and beyond, the vast majority of domestic AI pharmaceutical companies are still in the A round or even the angel round and seed round. In the context of the significant contraction of investment and financing in the biopharmaceutical field since 2022, there may not be many companies that can successfully obtain the next round of financing.

This is related to the pain points of this segment, such as data. Data in the pharmaceutical field is highly confidential, especially in the R&D stage, it is impossible to disclose a large number of data, so compared with general AI models, the available data in the pharmaceutical field is very limited, and the data quality is not ideal, and there are often gaps and errors. However, if an AI model wants to achieve more functions, it must be trained and iterated by "feeding" a large amount of data.

It can be said that the special industry background directly restricts the development of AI pharmaceuticals. In fact, this is also reflected in the existing business, and the current domestic AI-driven pipeline tends to focus on mature targets, which are more likely to be supported by data. For targets that are rarely involved, what AI can do at this stage is relatively limited.

At a time when artificial intelligence is surging, AI has long passed the embryonic stage for biomedicine, and has grown into a new driving force for every core link of drug research and development. But in this extremely high threshold industry, the requirements for AI will also increase accordingly. Therefore, the author believes that in this environment where the financing boom has passed, only the leading enterprises that have truly established a leading edge can survive. Moreover, pure AI pharmaceutical companies are forced by the financing environment, and they may be able to achieve leapfrog fission more than traditional pharmaceutical companies that apply AI.

Note: This article does not constitute any investment advice. The stock market is risky, and you need to be cautious when entering the market. There is no harm in buying and selling.

View original image 54K