AI Pharma's Decade of Rise and Reckoning: From Hype to Pragmatism

A Decade-Long Dream: Breaking Free from Illusions, Demystifying Myths, and Embracing Pragmatism.

On a winter night in Boston in 2016, inside the conference room of Pfizer's headquarters, scientists gazed at the predictive data from the Chinese team, their faces filled with disbelief.

More than a month ago, Pfizer organized a global blind test competition, which mainly focused on crystal form prediction for three unreleased drug molecules developed by Pfizer. This was an extreme challenge to the core technology of drug development, and the list of participants included many well-known institutions from Europe and the United States.

The test results were announced, surprising everyone — XtalPi, a Chinese startup founded just two years ago with a team of only 30-plus people, achieved 100% accuracy in predicting the crystal forms of three drug molecules. Not only that, but they also compressed the traditional R&D process, which usually takes several months, into just a few days.

When the answer was announced, it was past 3 a.m. in China. The three founders of XtalPi, Shuhao Wen, Jian Ma, and Lipeng Lai, gathered around the computer and were moved to tears upon hearing the results.

This technological victory enabled XtalPi to officially become one of Pfizer's suppliers in early 2017, also marking the beginning of China's AI-driven pharmaceuticals stepping onto the global stage.

From 2014, when XtalPi was founded, to the present day, over a hundred startups have emerged in China’s AI pharmaceuticals sector, such as Accutar Biotechnology, BioMap, Deep Intelligent Pharma, and Egret Therapeutics. With their narratives of algorithms and computing power, these companies hold what they call "dragon-slaying skills," painting a disruptive blueprint for traditional pharmaceutical enterprises and investors.

They aim to overturn the "double ten dilemma" faced in new drug development — it takes over a decade to develop a new drug, with R&D investment reaching up to one billion US dollars, while the success rate is only 1%.

Time is money, and the long R&D cycle of innovative drugs has been a bitter pill.

AI drug hunters like XtalPi are entering the pharmaceutical industry with great enthusiasm, with the vision of ensuring that the scenes depicted in *Dying to Survive* will never happen again.And AI is the key, which significantly accelerates the R&D process and exponentially reduces R&D costs.

How short can the time be? Ma Jian, founder of XtalPi, once envisioned a scenario: design a molecule in the morning, and synthesize the compound by the afternoon.

AI has ushered in a new era of faster, cheaper, and more efficient drug discovery.

In the biopharmaceutical field, money seems to have found a new breakthrough, pouring in like crazy. Money is being thrown around, and capital is directly piling up the growth space of the entire track. AI pharmaceutical companies are having their moment in the spotlight.

Since then, the track has started to become crowded, and problems such as enterprise homogeneity and low-level entry barriers have gradually surfaced. The ability of AI to reduce costs and increase efficiency has been fully demonstrated in the drug discovery stage, but no AI-discovered drug has reached the market. Without impressive clinical data, AI-driven drug development has fallen into a winter season, with companies resorting to layoffs, mergers and acquisitions, and transformations.

A Decade-Long Dream: AI-Driven Pharmaceutical Companies That Survived Return to Their Original Essence – Shedding Illusions, Demystifying Myths, and Embracing Practicality.

The rules of survival have been rewritten; the life-and-death race of AI drug discovery has just entered its most thrilling phase.

Setting Out from the Desert

The year XtalPi secured the Pfizer order, across the ocean, Alex Zhavoronkov, the founder of Insilico Medicine, was selling personal assets to keep the company afloat.

Alex was born in Latvia, Europe. Most of the time, he wears a black Polo shirt printed with the company logo, promoting AI target discovery technology to pharmaceutical companies in a Baltimore laboratory.

Every presentation was a tough battle, and the feedback was basically the same: "The idea is good, but please prove it with experimental data."

Although generative AI is very popular now, back in 2014 when Alex connected AI with biology and chemistry, it was still in its very early stages.

Alex's venture into entrepreneurship was a mid-career transition. Previously, he worked in the IT industry and was employed at ATI Technology, which was later acquired by the chip giant AMD. After the acquisition, Alex dove headfirst into biopharmaceuticals, determined to make it his lifelong career.

On the intersecting track of biomedicine and artificial intelligence, Alex originally intended to apply AI to understanding and processing fundamental biological data (such as predicting physiological age). After communicating with pharmaceutical companies, he realized that the value of applying AI to drug target discovery was much greater.

The key to successful innovative drug development lies in the target, which refers to the binding site between the drug and biomacromolecules in the human body. The target can be a receptor, enzyme, immune system, gene, etc. Most failures in innovative drug development are due to incorrect target selection.

In order to improve the efficiency of drug research and development, modern medicine proposed the target-based drug discovery approach very early on. If it is possible to make combinations of diseases and targetable sites within the human body, there may be a chance to find treatments for certain diseases.

However, the manual experimental approach is not only time-consuming and labor-intensive but also has a very low success rate.

Therefore, Alex boldly envisioned that if AI could be used to review hundreds of thousands of medical articles, research projects, and clinical trials, it might potentially uncover underlying disease mechanisms and therapeutic targets.

Alex pitched his ideas to pharmaceutical companies. At first, they wanted experimental data. After obtaining the data for validation, they claimed that AI was not powerful enough. When Alex said, "Let's collaborate and create miracles," the pharmaceutical companies still did not believe him.

This is the eve of AI's emergence in the pharmaceutical circle —— it stems from the foresight of a group of people who believe that AI can identify molecules related to drug properties and predict biomolecular structures. Besides significantly increasing success rates, it can also shorten R&D time and save on R&D costs. As for whether this business model will work, no one has the answer.

In 2014, when XtalPi was in its preparation phase, Shuhao Wen led the team to seek seed-round financing, facing a market full of uncertainty. At that time, the awareness of AI-driven drug discovery in China was nearly nonexistent.

At that time, China's pharmaceutical industry was deeply trapped in the "generic drug dilemma." Data shows that among the 37 Class I new drugs launched between 2017 and 2020, only three had original mechanisms of action. Of the 401 research and development targets globally, China covered only 80, making innovative drug development akin to scaling a cliff.

When Ma Huateng saw the project, he once told Wen Shuhao directly, "At that time, no one had reviewed the project after it was submitted, and Tencent had not invested in this type of project before. But what you are doing has social benefits."

After Ma Huateng's decision, XtalPi secured a 24 million yuan Series A financing from Tencent. Tencent has since invested in multiple additional rounds, becoming one of its key investors.

Investment from Ma Huateng and securing an order from Pfizer are two events frequently mentioned within XtalPi. Wen Shuhao said that it’s not necessarily because they are the biggest milestones, but because they happened at the moment when XtalPi needed them the most.

Alex also met his "benefactor."

In 2018, Li Ge, the founder of WuXi AppTec, led a new round of strategic financing for Insilico Medicine. Both parties expressed their intention to closely integrate Insilico Medicine's AI technology with WuXi AppTec's experimental capabilities.

After this round of financing, the gears of fate began to pull this company towards China. On the advice of the investment director of WuXi AppTec, Alex decided to relocate the company's headquarters to Hong Kong.

Alex once explained to the media, "Over the past 20 years, just like in the IT industry, China's biotechnology industry has undergone tremendous changes. China's investment in basic research continues to increase, and there are well-established contract research organizations. In China, the results discovered by artificial intelligence can be validated in biology and chemistry more quickly."

AI pharmaceutical companies have received early investments, but the market remains immature, so they decided to focus on "paving the way" first.

Ma Jian and the team have been continuously engaged in external publicity and market education. At the same time, they started to expand XtalPi's infrastructure and capabilities, such as establishing a crystal form laboratory. With this experimental capability, XtalPi can not only directly meet the entire crystal form research needs of enterprises but also visually demonstrate whether the predictions match the experimental results.

Alex has developed an AI software system for automated machine learning in drug design. After identifying the target, this system can automatically design corresponding protein drug structures based on deep generative algorithms.

At this time, a reform is quietly brewing within Chinese pharmaceutical companies.

In 2016, Janssen Pharmaceuticals, a subsidiary of Johnson & Johnson, reached an agreement with BenevolentAI, an artificial intelligence company, to use AI technology to evaluate the clinical potential of small molecule compounds, particularly in the field of Parkinson's disease.

Li Xing is a member of the original new drug development leadership team at Johnson & Johnson China, where he worked on new drug research and development for seven years and witnessed the company’s evolving recognition of AI-driven drug discovery.

In the past, she often discussed with her team how to bring new drugs to market earlier, but they rarely considered it from the perspective of IT technology. This cross-disciplinary collaboration allowed Li Xing to see the great potential of AI and informatization in accelerating new drug development.

She began self-studying AI, devouring dozens of professional books related to AI and attending numerous AI-related lectures and conferences. In 2017, she gave several presentations on AI-driven pharmaceutical R&D operations within Johnson & Johnson, earning the opportunity to present at the headquarters. This journey made her realize, "The informatization and intelligent transformation of pharmaceutical R&D must be done."

In the second half of 2017, Li Xing resigned from Johnson & Johnson and founded Deep Intelligent Pharma, a company specializing in information technology and intelligence for pharmaceutical R&D. The slogan and vision she set for this company is, "Make it easy to develop new drugs."

"The car is coming."

Although pharmaceutical companies have not eliminated their doubts about the true value of AI in drug discovery, the pain points in new drug development persist, prompting them to adopt a "try and see" attitude and collaborate with these emerging enterprises.

In May 2017, Sanofi, a French multinational pharmaceutical company, announced a collaboration with Exscientia, a British AI-driven drug discovery company, to develop bispecific small-molecule diabetes drugs, with an investment of 250 million euros.

This was a huge collaboration between pharmaceutical companies and AI drug discovery companies at the time, also meaning that pharmaceutical companies acknowledged and valued the role of AI drug discovery companies in drug discovery.

Two months after gaining recognition from pharmaceutical companies, Exscientia secured its first $15 million in Series A funding, five years after its establishment.

Starting from this point, large pharmaceutical companies are increasingly willing to collaborate with AI companies in drug development, and venture capital firms are also showing growing interest in investing.

However,The AI pharmaceuticals market really heated up in 2020.

This year, the disruption caused by the "black swan" of the pandemic has pushed the medical field to the forefront, making the demand for the application of new technologies in healthcare more urgent. On the other hand, in the secondary market, there are two "AI pharmaceutical" concept stocks: Schrödinger and Relay Therapeutics, which went public on Nasdaq in February and July 2020, respectively.

Among them, Schrödinger, the first computational drug discovery company to go public in the industry, reached a market value of 40 billion US dollars at that time.

Schrödinger's outstanding performance has completely ignited the secondary market, with this wave of enthusiasm spreading from the US stock market all the way to the domestic market. Although no companies have gone public in China, the primary market has welcomed a significant financing boom.

When he learned that XtalPi was about to undergo Series C financing, Jing Xutian, Managing Director of 5Y Capital, was thrilled. The company had previously discussed XtalPi multiple times internally, and he realized: "If I don't invest in it, this will become a major regret of my career."

He issued a TS (Term Sheet) as the lead investor to this company and invested tens of millions of US dollars.

In September 2020, XtalPi officially announced the completion of a $318.8 million Series C financing round, setting a new record for the highest financing amount in the global AI drug discovery field at that time.

The reason why capital began to favor AI pharmaceuticals is that AI + pharmaceuticals seemed to have made substantial breakthroughs in new drug research and development.

In February 2020, DSP-1181, a new drug for treating obsessive-compulsive disorder co-developed by the British AI pharmaceutical company Exscientia and Sumitomo Dainippon Pharma of Japan, entered Phase I clinical trials. This is the world's first AI-designed drug to enter the human trial stage.And it took less than a year from the concept proposal to entering clinical trials, while the industry average is 4.5 years.

In November of the same year, AlphaFold2, developed by Google's DeepMind, successfully predicted the three-dimensional structure of proteins. It solved a 50-year-old难题 in the biology界, driving significant advancements in large-molecule drugs through protein spatial prediction.

An Outsider to the Pharmaceutical Industry Delivers a Heavy Blow to a Pharmaceutical Giant.

For a long time to come, the value of the AI pharmaceuticals industry will be widely discussed. People's attitudes towards AI and drug development are either "completely unfeasible" or seen as disruptive innovation. But at this moment, these entrepreneurs in AI pharmaceuticals have opened the door to the capital market.——According to statistics from the Arterial Network, from January 2010 to October 2020, more than 50 AI + new drug companies at home and abroad received financing, with a total financing amount exceeding 4.5 billion US dollars.

Investors' valuation of AI drug discovery has undergone a dramatic shift.

In the summer of 2020, Alex realized that the market in China was becoming increasingly important, and he reached out to Ren Feng, who would later become his partner in China.

Ren Feng has an impressive resume, a Harvard alumnus, head of two departments at the world's second-largest pharmaceutical company, and the behind-the-scenes driving force in the IPO journey of MedChemExpress (MCE), China's leading CRO company.

There is a saying that before Ren Feng joined Insilico Medicine, the company did not have anyone who truly made drugs; it was basically a group of AI engineers. Ren Feng's intention was clear: to transform the company into a real AI pharmaceutical company.

He single-handedly built a team, recruited talents, and in just over two years, formed a drug research and development team of nearly 150 people; with the other hand, he took bold actions, halted low-value pipelines, and concentrated resources on advancing high-value projects.

At the end of February 2021, Ren Feng and his team quickly nominated Insilico's first PCC (Preclinical Candidate Compound). According to media reports, this is also "the world's first AI drug preclinical candidate compound."

AI Drug Development Continues to Heat Up, Many New Players Emerge in the Market. According to media reports, 38 AI drug development companies were established in China in 2020, such as Accutar Biotechnology, Zhixing Technology, SuperDimension Pharma, Deep Potential Technology, and StoneWise. Unlike the early founders, most of the founders of these new companies come from technical backgrounds without experience working in pharmaceutical enterprises; they choose to assist pharmaceutical companies in completing AI-driven drug development from various stages.

Besides, China's BATs are also unable to sit still.

Robin Li personally took the lead in establishing the AI life health platform "BioMap".

Tencent COO Ren Yuxin also announced in this year that Tencent officially launched its first AI-driven drug discovery platform "Yunshen Zhidrug".

Alibaba Collaborates with Zhejiang University and Multiple Academic Institutions to Establish the Alibaba Health Pharmaceutical Laboratory, Focusing on Key Areas Such as Drug Screening and New Drug Development with AI Technology at Its Core.

Huawei Offers Over a Million Yuan Annual Salary for Drug Research Algorithm Engineer Positions.

At this point, these AI entrepreneurs are facing a group of formidable opponents – internet giants, who have almost monopolized the top AI technical talents in China.

The three major forces in AI-driven drug discovery are gradually forming: traditional pharmaceutical companies, new AI-driven drug discovery players, and internet giants.

In 2020, the number of financing rounds in China's AI pharmaceuticals track reached 12, with a total financing amount of 2.723 billion yuan. This year was called the first year of AI pharmaceuticals.

Hard Bones to Crack in the Midst of Celebration

It is worth mentioning that the frenzy of AI pharmaceuticals has been driven by outsiders.

A large portion of the incoming hot money originates from TMT funds.

An investor who had previously sought to invest in such funds told the media that, in his view, investing in AI technology and the Internet has always been their forte. AI drug discovery presents a rare opportunity to tap into the trillion-yuan pharmaceutical market.

With the distinct style of the Internet, TMT funds are aggressively expanding in the AI pharmaceuticals field. They hope to burn through cash within months to create leading players, similar to what happened with shared bicycles.

AI Drug Development Embodies Both Speed and Slowness. On one hand, it follows the internet's mantra of "speed above all," while on the other hand, it reflects the pharmaceutical industry's "ten years to hone a sword."The collision between the two modes of thinking was very intense.

As for the so-called AI pharmaceuticals, the core is "AI" or "pharmaceuticals". For a long time, the industry itself has no answer.

This, in turn, has laid down some contradictory foreshadowing for the future development of AI-driven drug discovery.

After a period of frenzied progress, it's time to test the results. For the entire pharmaceutical industry, theoretically, no matter how excellent a molecule is, it must undergo validation during the clinical stage.

However, the anticipated new drug did not arrive as expected.

In July 2022, the world's first AI-designed drug, DSP-1181, was discontinued by Sumitomo Pharma due to its Phase I clinical study not meeting the expected standards.

A clinical pipeline, from its sudden emergence to its quiet disappearance, lasts only about two years.

At its root, it is because of the severe homogenization of data and algorithms.

High-quality, large-scale, and standardized biopharmaceutical data are the cornerstone of AI model training. However, in reality, data is scattered, siloed, and varies in quality, limiting the performance and generalization ability of AI models.

Ren Feng from Insilico Medicine mentioned to Yiou that the majority of data for AI pharmaceutical companies comes from publicly available data provided by pharmaceutical companies, research institutions, or universities. However, most drug development data is held by pharmaceutical companies and considered a core asset, which they are unlikely to share easily.

Early AI pharmaceutical entrepreneurs were keen on telling stories about AI-driven drug discovery.Zhou Jielong, founder of StoneWise, told Yiou that before 2020, many AI pharmaceutical companies were doing a lot of everything, such as molecular screening, protein structure prediction, etc. It sounded very "technological" to introduce AI, but in fact, no real breakthroughs were seen.

Since 2020, AI pharmaceutical companies have discovered an increasing number of drugs with the help of AI technology. However, compounds identified by artificial intelligence cannot guarantee success in clinical trials. On the other hand, if the computational results are inaccurate, pharmaceutical companies no longer want to pay, as they consider it not a worthwhile investment.

At the same time,The commercialization path of AI pharmaceuticals has also been criticized.

Currently, the main commercial directions for AI pharmaceutical companies in China are: Biotech (innovative pharmaceutical enterprises), CRO (contract research organizations), and SaaS (software tool-based companies).

Some media once pointed out incisively that XtalPi, which boasts itself as the "first AI drug discovery stock," is essentially a CRO company providing services to pharmaceutical enterprises once the AI label is removed. Indeed, pharmaceutical companies are its target clients, and XtalPi mainly offers drug discovery solutions at the preclinical stage along with some intelligent automation solutions.

This model is highly dependent on large pharmaceutical companies and is closely related to the overall pharmaceutical financing environment. If pharmaceutical companies face increased difficulties in financing, their willingness to develop new drugs will also significantly decrease.

Insilico Medicine, which follows its own drug development model, is internally researching multiple self-developed pipelines. However, the long pharmaceutical cycle means its drugs have yet to be commercialized. As a result, R&D investment has caused its net debt to continuously rise, leading to enormous financial losses.

Starting from the second half of 2022, the capital frenzy that had lasted for more than two years entered a severe winter.

In April 2023, BEN-2293, a topical pan-Trk inhibitor for the treatment of atopic dermatitis developed by AI drug discovery company BenevolentAI, was terminated due to not meeting secondary efficacy endpoints in Phase IIa clinical trials, which triggered layoffs within the company.

In June of the same year, Insilico Medicine's failed IPO attempt further deepened the industry's chill.

Internationally, the largest M&A deal in the AI pharmaceuticals industry has been born: Recursion and Exscientia have announced that they have reached a definitive agreement to merge. As a star company in the AI pharmaceuticals field, Exscientia will cease to exist after the merger.

A Decade-Long Dream: Exscientia Becomes a Footnote in the Boom and Bust of the AI Drug Discovery Field.

Reviewing the whole year of 2023, only one AI pharmaceutical company in China received B+ round financing, while the stock prices of Insilico Medicine, XtalPi, and others plummeted, with a reduction of over 70% in primary market financing.

The Story of AI Drug Development on the Brink of Bankruptcy: Industry Truth Revealed – Among 300 Global AI Pipelines, Less Than 5 Have Entered Phase III, and No New Drugs Have Successfully Reached the Market.

"A PPT with storytelling cannot match a clinical report." A senior executive of an investment bank lamented at a Lujiazui drinking party.

Amid the burst of the bubble, 80% of funding flows to leading companies, while small and medium players struggle to survive in the financing winter.

The Fork in the Road of Fate

At the beginning of 2023, Schrödinger, the first AI pharmaceuticals public company, is attempting to distance itself from AI.

When the analyst referred to Schrödinger as an AI pharmaceutical company, CFO Jeffrey Boggess immediately interrupted, clarifying that Schrödinger is not an artificial intelligence company but a pharmaceutical company with proprietary software.

Schrödinger CEO Ramy Farid once said in an interview with the media: "To me, being described as an AI company is like describing yourself as a company that uses Office software."

In Farid's view, there is no need to "mythologize" AI, and he hopes everyone should downplay the AI label.

Schrödinger's official website now introduces its computational platform as follows: Schrödinger’s computational platform, powered by physics, is transforming the way therapeutics and materials are discovered.

As a bellwether, Schrödinger pointed out the repositioning of AI pharmaceutical companies: AI is an add-on, but ultimately, it must return to drug manufacturing.

In 2024, with the industry reshuffle, AI pharmaceuticals has also moved past the stage of focusing on technology discussions and concept presentations, becoming more pragmatic.

Players in the constantly evolving industry landscape have also moved towards their respective destinies.

In August 2024, significant news emerged in the field of scientific and technological innovation as XtalPi officially signed a five-year strategic cooperation agreement with energy giant GCL Group, involving an amount as high as US$135 million (approximately RMB 1 billion). The collaboration focuses on researching and developing new materials in the fields of perovskite, supramolecular, lithium-ion batteries, cathode materials, and silicon-carbon materials using artificial intelligence and robotic automation.

Wen Shuhao regards entering the new materials field as the company's second venture. He told the media, "XtalPi will definitely become an AI platform company in the future. We define ourselves as an AI innovation and R&D platform for vertical industries."

After making a giant leap from pharmaceuticals to new materials, XtalPi has further expanded its commercial exploration into agriculture, new energy, and consumer goods sectors. Compared to drug development, these new businesses do not require lengthy clinical trial processes and can generate returns more quickly.

As for pharmaceuticals, Wen Shuhao said, "We do not push for the market launch of our own drugs, but rather leave the subsequent clinical development to clients, allowing professionals to handle professional tasks."

XtalPi,找准自身定位, 迎来了高光时刻。On June 13, 2024, the bell echoed in the Hong Kong Stock Exchange.The three founders, Shuhao Wen, Jian Ma, and Lipeng Lai, exchanged smiles as the market value of 19.759 billion Hong Kong dollars appeared on the screen.

In China's AI pharmaceuticals industry, XtalPi is the first AI pharmaceutical company to go public. The name XtalPi combines "Xtal," representing crystal forms, and "Pi," representing the mathematical constant π, symbolizing an entry into pharmaceuticals from a physics perspective. Starting from biopharmaceuticals, XtalPi has stepped into a broader market.

Some people make it to the shore, while others disappear without a trace.

In August 2024, two companies under the AI drug discovery enterprise SuperDimension Pharma have both shown as deregistered. SuperDimension Pharma was founded in 2020, during the peak years of entrepreneurship and financing in the AI drug discovery sector. In July 2021, SuperDimension Pharma announced that it had secured tens of millions of RMB in angel round funding, with the investor being Chunxin Changying, a subsidiary of CITIC Capital.

HyperDimension once had a clear business plan and was benchmarking against Insilico Intelligence in its operations.

However, whether it was self-developed pipelines or collaborations, XtalPi failed to deliver impressive results. Until it vanished from the market.

During a market downturn, how to survive becomes the key question that AI pharmaceutical companies need to answer.Almost all of China's leading AI + drug discovery companies are undergoing cross-sector transformation.

At the 2024 World Power Battery Conference, Deep Potential Technology and the Yibin government jointly launched a new energy materials digital intelligence innovation project. Based on their AI for Science large model, this project provides support for the Yibin government's new energy industry, with a total investment of several hundred million yuan.

Similarly, BioMap announced a strategic partnership with Dabeinong to jointly build a large AI model for biotechnology in agriculture, thereby entering fields such as synthetic biology and bio-breeding.

Compared to other major players opening up a second front, another giant in AI drug discovery, Insilico Medicine, has chosen to stick to the path of pharmaceuticals and is attempting to carve out a self-sustaining route.

Ren Feng told Yiwu that over the past decade, Insilico Medicine has been exploring the commercialization of AI-driven drug discovery, going through multiple stages — from initially providing software platform services, to balancing software development with drug discovery, and finally evolving into an AI-driven biotech company.Every step taken brings us closer to the essence of the industry: Setting aside the halo of AI, progress in R&D and collaboration, along with financial metrics, have become the key factors in determining a company's value.

In March 2024, Insilico Medicine filed its prospectus with the Hong Kong Stock Exchange for the second time. The rush for the IPO was driven not only by investor pressure but also signaled the company's ambition to transition into a biotech firm. However, the transition period has been extremely short.

Insilico Medicine's IPO Journey Hits Another Roadblock.

Heading Towards the Next Decade

In 2024, AI pharmaceuticals is at a transformative moment, with NVIDIA founder Jensen Huang stepping forward to advocate for biology: The era when everyone had to learn computers has passed; human biology is the future.

Not only Huang Renxun, but also Elon Musk, Li Yanhong, Zhang Yiming, Huang Zheng and other successful figures have shown a strong interest in life sciences.

In 2021, Huang Zheng stepped down as chairman of Pinduoduo and turned to the fields of food science and life science. Before stepping down as chairman, Huang Zheng had already established the "Stellar Charity Foundation" with his team, which is committed to promoting scientific research in biomedicine, agricultural food, and other fields.

When Zhang Yiming stepped down as CEO of ByteDance, he mentioned in an internal letter to all employees that the dawn of virtual reality, life sciences, and scientific computing's impact on human life has begun to show. To seize this opportunity, innovators need to break through the inertia of their businesses to explore. For this reason, Zhang Yiming mentioned that he would participate in this wave of technological transformation over the next ten years.

Wang Xiaochuan Steps Down as CEO of Sogou, States in Internal Letter: "For the Next Twenty Years, I Hope to Contribute to the Development of Life Sciences and Medicine."

These internet tycoons have experienced entrepreneurship and ups and downs in the internet wave, and when they have achieved success, life sciences have become their common choice for the second half of their careers. Clearly, they are all aware that compared to fields such as computer science, software, and chips, life sciences are much more complex and represent the most difficult science to comprehend.

Has the Hype Around AI-Driven Drug Discovery Faded?

Not at all. Anyone can look up at the stars, but more importantly, we need to keep our feet on the ground.

The ability of AI technology platforms to create and generate innovative molecules has become unquestionable today. A research report by Boston Consulting Group (BCG) shows that AI molecules have demonstrated clinical success rates surpassing the industry average:

AI-Generated Drug Molecules Achieve Success Rates of 80%-90% in Phase I Clinical Trials, Compared to Only 50% Previously.

In the past, it was inevitable that there were bubbles in the industry, but they weren't all bad. Pioneers like XtalPi and Insilico Medicine began their arduous explorations much earlier than many startups over the past decade. They initiated the marketization of technology and R&D, laying the groundwork for the industry to build its foundational infrastructure from scratch.

At the same time, acknowledging the current limitations of artificial intelligence and rationally viewing the relationship between AI and drug development marks the beginning of the industry's "disenchantment" to preserve authenticity.An obvious trend is: AI pharmaceuticals, which initially emerged as a surprising force, have now merged into the mainstream of traditional pharmaceuticals as "new pharmaceutical infrastructure."

Many large pharmaceutical companies have brought in AI drug discovery enterprises as partners to jointly develop marketable drugs.

For example, top global biopharmaceutical companies such as Pfizer, GlaxoSmithKline, and Novartis have established machine learning centers driven by data science, and Sanofi is committed to becoming the first large pharmaceutical enterprise massively driven by AI.

AstraZeneca also claims that artificial intelligence empowers over 50% of its small-molecule pipeline, while Johnson & Johnson has hired more than 6,000 data science and AI experts to optimize R&D and organizational decision-making.

At the beginning of 2025, DeepSeek triggered a paradigm shift in the global AI industry with its low-cost, high-performance technological breakthrough, while also injecting new momentum of "inclusive computing power" into drug research and development. This milestone event brought AI-driven pharmaceuticals back into the spotlight.

AI Drug Discovery Enters the Second Evolution.

Ma Rui, a partner at Frees Fund who has been studying biotechnology for many years, divides AI drug discovery into two eras: 1.0 and 2.0. AI 1.0 refers to discriminative AI, while AI 2.0 is generative AI. In his view, during the AI drug discovery 1.0 era, people generally believed that AI was unreliable and that physical methods were more accurate. However, generative AI will now change this situation by moving from merely making predictions to being capable of design and generation, which will be the biggest technological breakthrough.

After a decade of industrialization, the AI pharmaceuticals industry has completed its initial build from 0 to 1; it has now entered the second phase, where artificial intelligence can independently discover targets and synthesize drug molecules, with promising applications in clinical and real-world research.

Former FDA reviewer Du Tao predicted,The next investment hotspot in the AI pharmaceuticals track lies in AI clinical trials.

Globally, the AI clinical track has already taken shape. There are dozens of companies worldwide engaged in AI clinical services, and many multinational pharmaceutical enterprises have directly invested in AI clinical design companies.

Currently, there are more than a hundred AI pharmaceutical companies in China, most of which are concentrated in the early-stage target/site or molecular phase, making this part of the track extremely crowded.

And the "big cake" that AI clinically intervenes in accounts for 80% of the cost. Whether it is to save 80% of the money by making drugs independently or to earn 80% of the money by providing services to large pharmaceutical companies, there is tremendous commercial potential.

At present, AI pharmaceutical companies have also realized the importance of "differentiation," and are gradually broadening the application of AI technology in the new drug development industry. Instead of making general statements, they are now focusing on solving problems and delivering results.

In May 2025, Insilico Medicine launched its third attempt to go public on the Hong Kong Stock Exchange. The twists and turns of its IPO journey serve as a microcosm of the challenges faced in implementing AI pharmaceutical technologies—full of obstacles, yet never ceasing to move forward.

In conclusion

Looking back from 2025, the rise of China-produced innovative biopharmaceuticals and the explosive development of artificial intelligence technology marked a decade of AI in drug discovery — a history where people wielded the "hammer" of artificial intelligence to drive the "nail" of biology.

We can't expect to wield the big hammer of AI and knock around everywhere to disrupt the pharmaceutical industry, but every human industrial revolution has been a revolution of tools.

As in 2015, China's innovative drug industry ushered in a period of vigorous development.

Almost at the same time, China's first batch of AI pharmaceutical companies, led by XtalPi, were also established one after another, becoming an emerging force in China's innovative drug industry.

After a decade of arduous journey, China's AI pharmaceutical industry has transformed from a desert wanderer into an oasis builder. This time, entrepreneurs are no longer talking about disruption but instead focusing on refining every dataset and advancing each pipeline. Their goal remains unchanged: to place the first pill designed by Chinese AI truly into the hands of patients.

The golden age of AI pharmaceuticals has just begun.

References:

"Unveiling AI Drug Discovery" —— China Entrepreneur Magazine

"This Latvian-founded company is striving to become 'China's First AI Pharmaceutical Stock'" — China Entrepreneur Magazine

"From the Outsiders' Celebration to Calm: AI Pharmaceutical, a Gradual Revolution Returning to the Essence of Medicine" — Deep Blue View

"Li Yanhong's Life Science Company Completes New Round of Financing: Why Do Tech Titans All Favor 'Eternal Life'?" — Health News Consulting

"When AI Pharmaceutical Companies Start to Huddle for Warmth" —— Amino Observation

"AI Pharmaceutical R&D Kicks Off the Breakthrough Competition, the Beginning of a New Era" —— Amino Observation

"Hidden Dragon in the Abyss: AI Pharmaceuticals Rushing Towards the Next Decade" — Zhi Medicine Bureau

This article comes from the WeChat Official Account"Yiou Network" (ID: i-yiou), Author: Zhou Jing, 36Kr authorized release.