The Application and Investment Outlook of Artificial Intelligence in the Pharmaceutical and Materials Science Fields

1. Technical Overview and Application Scenarios

Pharmaceutical Industry:

Artificial intelligence is being applied throughout the drug development process, significantly accelerating and improving outcomes. Traditional drug discovery can take over a decade and cost billions of dollars, while AI-driven approaches can analyze vast chemical and biological datasets to identify promising drug candidates at a faster pace. Machine learning models assist in designing new molecules with specific properties (generative chemistry), predicting biological targets for diseases, and conducting large-scale virtual screenings to pinpoint the most likely effective compounds. For example, AI algorithms can predict how molecules bind to protein targets (which is key to efficacy) or identify toxicity risks early. Deep learning models like DeepMind's AlphaFold can now predict protein 3D structures in minutes—something that previously took scientists years. This breakthrough has made structure-driven drug design possible, guiding researchers in designing drugs that more precisely target biological objectives. AI is also used for optimizing clinical trials (such as selecting the best trial sites and subject populations). A notable case is Pfizer's COVID-19 drug Paxlovid: the development team used real-time AI prediction models to forecast outbreak areas, helping them choose trial sites and enroll 46,000 patients across 150 sites in just four months.

In addition to discovery and trials, AI also supports drug synthesis planning (proposing synthetic routes) and drug repurposing (mining existing drugs for new uses). Natural language processing (NLP) models, such as GPT-4, are being used to parse vast amounts of biomedical literature and propose hypotheses or experimental designs, thereby enhancing researchers' capabilities. Overall, AI in the pharmaceutical field encompasses new drug design, target identification (e.g., new cancer targets), virtual screening, preclinical modeling, clinical data analysis, and even manufacturing optimization, all aimed at reducing development time and costs to bring new therapies to patients faster.

Materials Science:

AI technology is similarly transforming the discovery of new materials and their performance improvements. Traditionally, materials science relied on repetitive experimentation and physics-based simulations, which are time-consuming. AI offers data-driven predictions: machine learning models can predict the performance (strength, conductivity, reactivity, etc.) of materials based on their composition and structure, enabling rapid screening of candidate materials. This aids in discovering alloys, polymers, or new compounds suitable for batteries, semiconductors, and other demands without complex experiments. AI-driven research has led to breakthroughs such as more efficient battery materials, lighter and stronger composite materials, and new catalysts for clean energy.

A notable case is Google DeepMind's development of the graph neural network model GNoME, which predicted over 2 million new inorganic crystal structures; approximately 380,000 of these were assessed as stable and synthesizable. This single AI model achieved an order-of-magnitude leap in the inventory of known stable materials, which would have taken hundreds of years to achieve through manual experimentation The potential applications of these AI-discovered materials include next-generation batteries, solar cells, and even superconducting materials. AI is also used in the optimization of synthetic pathways in material research: similar to drug synthesis, algorithms can propose ways to manufacture materials (such as optimal temperature/pressure or mixing order) to achieve desired performance. The integration of AI with robotics has given rise to automated laboratories—in these closed-loop systems, AI designs experiments, robots execute them, and feedback results. For example, the A-Lab at Berkeley Lab uses AI-controlled robots to continuously synthesize and test inorganic materials, reportedly processing samples daily at a rate 50-100 times higher than human researchers.

This significantly accelerates the validation of AI predictions, creating a virtuous cycle. Another increasingly focused area is the use of AI to assess the environmental impact of materials—such as predicting their recyclability or toxicity to achieve sustainability goals. Overall, the role of AI in the field of materials science includes material discovery, performance prediction, process optimization, and laboratory automation, enabling researchers to develop higher-quality materials for energy, electronics, infrastructure, and more at a faster pace and lower cost.

2. Representative Companies and Cases (Focusing on U.S. Stock Companies)

Pharmaceutical Industry:

There are several innovative companies in the U.S. leading AI-driven drug discovery. For example, the UK/U.S. company Exscientia (NASDAQ: EXAI) has multiple AI-designed drug candidates currently in clinical trials. In 2021, the company collaborated with Bristol Myers Squibb, using its AI platform to identify a promising immunotherapy molecule in just 11 months, receiving a milestone payment of $20 million, with a total collaboration value of up to $1.2 billion. AI-designed molecules by Exscientia (such as CDK7 inhibitors and MALT1 inhibitors) are entering Phase I clinical trials in oncology.

Another U.S. leader is Recursion Pharmaceuticals (NASDAQ: RXRX), which uses AI combined with high-throughput biology to map the cellular effects of thousands of molecules. Recursion has established a multi-year collaboration with Genentech, a subsidiary of Roche, applying its "Recursion OS" platform to up to 40 neuroscience and oncology projects. In 2023, Recursion also expanded its collaboration with Bayer, using AI for precision oncology, with a total collaboration value of $1.5 billion.

Schrödinger (NASDAQ: SDGR) provides a physics-based modeling platform that combines machine learning, widely used in drug and material design. NVIDIA claims its platform "has led the forefront of computational drug and material discovery over the past 20 years," and the company is now integrating generative AI models to further enhance predictive capabilities Pharmaceutical giants are also heavily investing in AI. Pfizer, Novartis, Roche, Johnson & Johnson, AstraZeneca, and Sanofi have all established AI departments or collaborated with technology companies. In fact, Sanofi, Eurofins, AstraZeneca, Novartis, and Johnson & Johnson are the pharmaceutical companies with the highest number of AI-related job postings. These companies have signed multiple cooperation agreements with AI startups, such as Pfizer collaborating with IBM to develop a drug repurposing project, Merck and GSK investing in AI biotech startups, and Roche acquiring an AI company for clinical trial data analysis.

In the startup sector, Insilico Medicine operates in both Hong Kong and the United States, advancing an AI-designed anti-fibrotic drug (INS018_055) into Phase I trials in 2021 and progressing to Phase II in 2023. Others like BenevolentAI and BioXcel Therapeutics are also utilizing AI for target identification and drug repurposing.

Big tech companies are also actively investing. Alphabet's DeepMind not only developed AlphaFold but also established Isomorphic Labs focused on AI drug discovery; IBM and Microsoft provide AI cloud services tailored for pharmaceutical research and development. This collaborative ecosystem has achieved tangible results— for example, the first AI-discovered drug has entered human trials, and Japan approved an AI-optimized drug (for treating obsessive-compulsive disorder) for market in 2021.

Materials Industry:

The application of AI in the materials industry encompasses chemical giants and specialized startups. Dow Inc. (NYSE: DOW), one of the largest chemical companies in the world, has developed a "predictive intelligence" platform that combines materials science expertise with machine learning, significantly shortening formulation development time. In collaboration with the CAS division of the American Chemical Society (ACS), Dow deployed a customized AI search platform that reduces material candidate screening from weeks to hours, enabling the selection of the best combinations from millions of chemical solutions.

BASF (headquartered in Germany, with operations in the U.S. and China) is also embracing AI. Its catalysis division collaborates with materials informatics startup Citrine Informatics to rapidly screen carbon capture catalyst formulations using AI, with models optimizing material screening through continuous learning. This collaboration has helped identify several promising new materials, and BASF executives stated, "AI-driven material development is the future... early investors will reap substantial rewards."

In the startup sector, Citrine Informatics provides an AI platform to assist in developing new alloys and polymer materials, collaborating with BASF, Boeing, and others. DeepMind is a rising star in materials science, with research findings revealing 2.2 million crystal structures indicating that big tech companies can also have a significant impact on the industry This achievement was jointly completed with the Berkeley Lab under the U.S. Department of Energy, demonstrating the potential of combining public research with AI.

In terms of hardware, companies like NVIDIA (NASDAQ: NVDA) provide GPU and AI toolkits (such as the Materials Discovery Toolkit) for supporting material simulations. Company executives pointed out that "accelerated computing + generative AI" will significantly enhance material research and development capabilities.

In China, XtalPi, a company originating from MIT and now headquartered in Shenzhen, combines quantum physics, cloud computing, and AI on its platform to serve drug and material design. The laboratory of the Chinese Academy of Sciences has built a "self-driven experimental platform" similar to Berkeley's A-Lab for high-throughput synthesis of inorganic materials.

Chinese industrial giants are also adopting AI: for example, Sinopec uses machine learning to optimize catalyst design, and Baowu Steel Group utilizes AI to enhance steel performance while reducing weight, making it more competitive in automotive and aerospace manufacturing. Although the number of material AI startups is not as high as in pharmaceuticals, AI companies like Fourth Paradigm and SenseTime have ventured into scientific applications, and Baidu's PaddleHelix platform is also used by material and biotech companies for molecular design.

3. Comparison of Practices between China and the U.S.

Both the U.S. and China recognize the transformative potential of AI in the pharmaceutical and materials fields, but their paths and progress focus on different aspects.

In the pharmaceutical field, the U.S. currently maintains a lead in cutting-edge AI models and their applications, while China is rapidly catching up, expanding quickly in numbers due to national support and vast resources. U.S. companies and academia have driven key technological breakthroughs like AlphaFold and have led the earliest clinical trials of AI-designed drugs. Many leading AI drug discovery startups (such as Recursion and Schrodinger) come from the U.S. or Europe and closely collaborate with large U.S. pharmaceutical companies. The U.S. also has a sound venture capital system and regulatory interaction mechanisms (for example, the FDA is advancing an AI regulatory framework), which help facilitate the implementation of AI in drug development.

China, on the other hand, is accelerating its layout through national strategy. Since 2017, the Chinese government has planned to include "AI + pharmaceuticals" as a key direction, prompting the establishment of numerous startups. Currently, over 100 AI-driven pharmaceutical companies have been established in China, pushing AI to solve pharmaceutical bottlenecks and leveraging successful overseas experiences to establish business models. As a result, China achieved the world's first AI-designed drugs entering human trials in 2023 (such as the anti-fibrotic drug from Insilico Medicine), almost in sync with Western progress. Chinese AI laboratories are integrating AI capabilities into hospitals and clinical systems by leveraging rich patient data and compound libraries However, experts point out that China still lags slightly behind in certain aspects—most of the most advanced AI models and innovative drugs still originate from the West. But the situation is changing: large tech companies like Huawei, Tencent, and Baidu are increasing their investments in life sciences AI (such as Baidu's PaddleHelix and the ActFound model developed in collaboration with Peking University) and collaborating with international teams.

From the perspective of patent indicators, China has now become the country with the highest number of AI-related patent applications, including patents for generative AI in the medical field. Between 2014 and 2023, China applied for over 38,000 generative AI patents, while the United States had about 6,300. This reflects the breadth of research activity in China, although the number of patents does not equate to quality.

In the field of materials science, the United States has also gained a first-mover advantage due to the launch of the Materials Genome Initiative in 2011, establishing several national laboratories integrated with AI (such as the A-Lab at Berkeley Lab). DeepMind's new materials research is also primarily led by American institutions (such as Berkeley and Oak Ridge). American universities and companies have a high output share in materials informatics research.

China, on the other hand, has strong overall capabilities in materials science and a fast implementation speed in manufacturing. Chinese steel, chemical, and battery manufacturing companies (such as CATL) have deployed AI quality control systems and structural optimization platforms on a large scale. The Chinese Academy of Sciences and Tsinghua University have established multiple AI materials discovery projects and are promoting the "Materials Genome Initiative" and supercomputing platforms (such as the AI-enhanced supercomputing system in Wuxi) with the help of AI.

Frequent cooperation and exchanges occur: many Chinese materials scientists have been trained in the United States and participated in multinational research projects. For example, the ActFound model developed in collaboration between Peking University and the University of Washington is a joint achievement of China and the U.S.

Policy is also one of the key differences: the U.S. primarily relies on research funding for scientific research, driven by industry; China, on the other hand, is government-led and promotes corporate cooperation. For instance, the Chinese government promotes the enhancement of power battery performance through AI as a national goal, forming a cooperation model between upstream and downstream enterprises.

Overall, both China and the U.S. are advancing on dual tracks: China has a clear advantage in execution speed and government mobilization, while the U.S. leads in core algorithms and international ecology. Both countries rank among the top two globally in AI pharmaceuticals and materials patents, papers, and investments. Despite potential policy frictions between China and the U.S. regarding technology exports and cross-border data, cooperation among scientists and the sharing of AI tools (such as AlphaFold data) provide a realistic foundation for global collaboration.

4. Market Trends and Policy Support

Market Trends—Pharmaceuticals:

The integration of AI in the pharmaceutical field is transitioning from experimental to mainstream, as evidenced by the growth in the "AI in Drug Discovery Market." Analysts estimate that the global AI pharmaceutical market size will be approximately $2-3 billion in 2024, with a compound annual growth rate of about 25-30%. Some forecasts that include generative AI even suggest that the market size could reach hundreds of billions of dollars by the early 2030s Growth momentum comes from successful cases and rising adoption rates: Almost all large pharmaceutical companies now have AI strategies, including internal development and external collaboration. The number of AI/pharmaceutical-related collaborative deals continues to grow—by the second quarter of 2024, the number of AI/pharmaceutical collaborations increased by 14% compared to the same period last year. We have seen AI-driven IPOs (multiple AI biotech startups went public in 2020-2021) and large financing rounds, indicating strong investor enthusiasm.

In terms of R&D output, there are currently more than ten AI-designed drug molecules in clinical trials globally (covering areas such as cancer, fibrosis, and immunology), and this is expected to grow significantly each year. Related trends also include the rise of "Pharma 4.0"—the comprehensive digital transformation of drug manufacturing and supply chains, integrating AI and the Internet of Things.

For example, companies are using AI for predictive maintenance on production lines, real-time quality control, and supply chain optimization. This improves efficiency and compliance (the FDA also encourages such digital technologies for quality management). On the commercial side, AI is also applied in marketing and drug vigilance (such as patient AI Q&A and algorithms for scanning safety reports). As the value of technology is validated, industry confidence grows—executives from large pharmaceutical companies frequently mention the role of AI in R&D during conference calls, and some companies (like GSK and Merck) have established Chief Digital Officer (CDO) positions to oversee AI integration.

Of course, there is also a "period of adjustment between hype and reality": not all AI-discovered drugs can achieve clinical success (Exscientia has a compound that did not meet its endpoints), reminding us that the complexity of biology still requires cautious handling.

Market Trends—Materials Science:

The application of AI in materials science (often referred to as materials informatics) started slightly later but is now growing rapidly. The market is estimated to be worth $135 million in 2023, with an expected average annual growth rate of over 16% by 2030.

Although smaller in scale, the customer base is concentrated, and the growth momentum is strong. Traditional materials companies face faster innovation pressures (such as developing better batteries, superconductors, and sustainable materials), and AI provides a competitive advantage.

One trend is the increase in cross-industry collaboration: technology companies and research institutions are partnering with manufacturing companies. For example, DeepMind collaborates with national laboratories to discover materials; startups are also co-developing materials with aerospace and automotive companies.

Sustainability demands are an important driving force—companies need to reduce their carbon footprint and are looking for new types of materials (such as recyclable plastics and rare-earth-free battery chemistries). AI can quickly assess options: for example, algorithms can screen for biodegradable, performance-compliant bio-based polymer combinations.

Policy measures are also driving this trend: the U.S. Department of Energy and the National Science Foundation provide funding for AI+ materials projects, and the 2022 CHIPS and Science Act explicitly supports semiconductor materials research and mentions the use of AI tools.

China has listed advanced materials as a priority in its "14th Five-Year Plan" and explicitly identified AI as an enabling technology, leading local governments to establish AI materials innovation centers Therefore, we see Chinese companies like CATL applying AI to enhance battery material performance, while Huawei uses AI to discover new semiconductor packaging materials.

Another trend is data democratization—data provided by initiatives like the Materials Project in the U.S. and material databases in China is gradually moving to the cloud for broader access.

On the enterprise side, acquisitions and collaborations are ongoing: for example, large engineering software companies are acquiring materials simulation companies to integrate AI (such as Dassault Systèmes adding machine learning modules to its BIOVIA software).

"Laboratory automation" and "future laboratories" have become buzzwords: more companies are investing in robotic laboratories (not just A-Labs, with companies like Chemify and Open Discovery selling automated chemistry experiment platforms). These laboratories generate big data, further feeding back into AI models—creating a virtuous cycle.

It is worth noting that compared to the biological field, the research subjects of materials AI are more controllable, progress is often faster, and the barriers to implementation are relatively lower. Therefore, we have already seen patents filed for AI-designed materials, some of which have been commercialized (such as catalysts discovered by AI being applied in chemical plants, or machine learning-optimized alloys used in jet engines).

The trend is clear: materials companies adopting AI are innovating faster, and competitors that do not keep up will face the risk of falling behind. Market analysts expect materials informatics to become a standard component of the R&D process. Once "moonshot-level" results appear (such as room-temperature superconductors or super batteries), market attention and investment could surge.

5. Investment Opportunity Assessment

The integration of artificial intelligence with pharmaceuticals and materials science presents significant market opportunities, but careful evaluation of commercialization timelines and potential risks is necessary. From an investor's perspective, AI is expected to significantly enhance R&D productivity—once successfully deployed, it will lead to more drug candidates, faster development cycles, and the discovery of high-value materials, translating into competitive advantages and financial returns. However, the maturity and risk levels vary across fields: AI commercialization in pharmaceuticals is relatively more mature (with several drugs already entering clinical trials and some companies starting to generate revenue), while materials AI, despite its broad prospects, is often embedded within large corporate systems, leading to longer return cycles.

Commercialization and Market Acceptance:

In the pharmaceutical field, we are beginning to see AI deliver concrete results (such as candidate drugs). Some AI-designed drugs have entered human trials, and large pharmaceutical companies are signing high-value collaboration agreements, validating the commercial viability of this field. It is expected that within the next 5–10 years, we will see the first batch of AI-discovered drugs officially approved—this will trigger a new wave of investment and technology adoption.

Industry executives generally accept AI as a key to enhancing competitiveness: most R&D heads in pharmaceutical companies view it as a "necessary tool." However, the drug development cycle is long and costly—many AI-driven biotech companies will still be unable to launch commercial products (i.e., approved drugs) in the short term, with their current revenue primarily coming from platform licensing and collaboration payments. This means that if AI's promises are not fulfilled or trials fail, their stock prices may experience significant volatility In the materials field, AI is more embedded in processes, with results such as new materials and new processes being launched in improved product forms. For example, aerospace manufacturers are adopting lighter and stronger alloys, and battery companies are applying new electrode materials screened by AI.

The development cycle for materials is relatively short (some catalysts can be commercialized within 2-3 years), but customer acceptance and industry certification processes may extend the time to market. Most materials companies improve efficiency and product performance through continuous improvement and gradual integration of AI optimization tools, allowing them to stand out in competition.

Technical and Regulatory Barriers:

Both industries face different technical thresholds.

In pharmaceuticals, the quality of AI models is limited by data quality—experimental data often contains noise or uneven samples. Companies that can overcome this issue (such as those building automated laboratories to generate uniform data or developing new algorithms) will have a competitive advantage. Another challenge is the difficulty of integrating AI into existing R&D processes: it requires not only system modifications but also team trust in model outputs.

In terms of regulation, the FDA currently does not impose specific limits on AI drugs, but additional safety validations may be required in the future, especially when AI is used for target discovery or molecular design.

In materials, AI predictions require experimental validation—if experimental resources are limited, this will become a bottleneck. However, automated laboratories are gradually alleviating this issue. Another barrier is the ownership of intellectual property: how to define the inventor's identity for innovations generated by AI and how to apply for patents still has legal gaps. Many companies adopt a "dual-track system"—patenting molecular structures while keeping algorithms as trade secrets.

Market and Policy Support:

Multiple "tailwind" factors support the investment logic.

The global pharmaceutical industry invests over $200 billion annually in R&D, and AI tools address its "low output efficiency" pain point. Once they help improve R&D ROI, their value will be quickly realized.

The materials field plays a key role in clean energy and high-tech manufacturing (chips, energy storage). If AI can help discover new materials and enhance experimental efficiency, it means significant gains.

Policy promotion: Whether through funding from the U.S. government via DARPA, FDA, NIH, or through special projects and local incubators in China, the industry is provided with policy and financial support.

Talent support: Companies that can attract top AI and scientific talent are more likely to succeed, and investors can monitor their hiring trends. For example, in Q2 2024, the number of job postings related to pharmaceutical AI increased by 10% year-on-year.

Beneficiary Classification from a Securities Investment Perspective:

1. AI-driven drug development companies (pure targets): Such as Exscientia (EXAI), Recursion (RXRX), Schrödinger (SDGR), Relay Therapeutics (RLAY). These companies represent direct exposure to the AI pharmaceutical theme, with their stock prices significantly fluctuating with clinical progress and collaboration news. If a certain AI candidate drug is approved or the company is acquired by a large pharmaceutical company, it will generate huge returns. However, they also exhibit typical volatility characteristics of biotech stocks 2. Large pharmaceutical companies embracing AI: Such as Johnson & Johnson (JNJ), Novartis (NVS), Bristol-Myers Squibb (BMY), Pfizer (PFE), AstraZeneca (AZN). Although AI does not account for a high proportion of their valuations, if deployed properly, their R&D efficiency and profit margins are expected to gradually improve. The benefits of AI can be verified by analyzing indicators such as the speed of new drug output and the number of collaborations.

3. AI technology providers (computing power + platforms): Such as NVIDIA (NVDA), Alphabet (GOOGL), Microsoft (MSFT), Amazon (AMZN). These companies provide infrastructure to AI pharmaceutical/material companies through GPUs, cloud platforms, and algorithm services. The expansion of the AI industry will drive growth in their cloud services and software revenues, making them relatively low-volatility indirect beneficiaries.

4. Pioneers in the materials/chemical industry: Such as Dow Chemical (DOW), BASF (BASF), Albemarle (ALB), Applied Materials (AMAT), Tesla (TSLA). These companies accelerate the development of new materials through AI, enhancing market share and product premium capabilities. For example, Dow uses AI to develop new polymer formulations, and Huawei uses AI to discover new chip packaging materials.

5. Representative enterprises in China and cross-border: Such as Baidu (BIDU) venturing into AI pharmaceutical platforms through PaddleHelix, and Tencent (700.HK) investing in several AI biotech companies (such as XtalPi, Atomwise). WuXi AppTec (2359.HK), BeiGene (BGNE), and other Chinese innovative drug companies are also actively laying out AI collaborations. Insilico Medicine, although not publicly listed, will be a high-potential target if it goes public in the future.

Market Investment Strategy:

Build an "AI + Science" themed investment portfolio, balancing innovative companies with mainstream giants.

Pay attention to policy directions and regulatory trends: For example, if AI drugs receive special fast-track approval from the FDA, or if the government encourages AI materials to be included in procurement lists, these will act as catalysts.

Focus on ETFs and index products, such as ARKG (which includes some AI pharmaceutical stocks), or future potential material AI-themed ETFs.

Overall, this theme represents the modernization trend of healthcare and manufacturing infrastructure, similar to the wave of internet infrastructure 20 years ago. The pharmaceutical industry can release significant profits by saving 10% on R&D costs; and once revolutionary new materials are discovered in the materials field, it will trigger a revaluation of the industrial chain's value.

Potential beneficiary targets (example categories):

AI-driven drug development companies: Exscientia (EXAI), Recursion (RXRX), Schrödinger (SDGR), Insilico Medicine (potential IPO) Pharmaceutical Giants Embracing AI: Bristol Myers Squibb (BMY), Johnson & Johnson (JNJ), AstraZeneca (AZN), Novartis (NVS), Moderna (MRNA).

AI Infrastructure Companies: NVIDIA (NVDA), Alphabet (GOOGL), Microsoft (MSFT), Amazon (AMZN), Thermo Fisher (TMO), Danaher (DHR).

Materials and Chemical Innovators: Dow Inc. (DOW), BASF (private), Albemarle (ALB), Applied Materials (AMAT), Corning (GLW).

Chinese and Hong Kong Companies: Baidu (BIDU), Tencent (700.HK), WuXi AppTec (2359.HK), BeiGene (BGNE), Contemporary Amperex Technology Co., Limited (CATL).

Summary

Artificial intelligence is gradually becoming the core driving force behind innovations in pharmaceuticals and materials science. The market has transitioned from the validation phase to the commercialization phase, with major global economies continuously increasing investments in policies, capital, and talent. From an investment perspective, the widespread application of AI in science not only enhances industry efficiency but also nurtures a group of high-growth companies. We are currently in a window period for positioning, and it is recommended to closely monitor the transformation of industrial achievements, adjustments in regulatory policies, and trends in multinational cooperation, entering at the right time and holding long-term.

This wave of "AI + Science" integration will be one of the most important themes in technology and capital over the next decade