Top 10 Most Promising AI for Science Companies in 2026: Global Leaders Driving the Scientific Revolution

2026: AI for Science Explodes!

According to incomplete statistics from Zhiyaoju, global AI for Science financing has exceeded in the past six months$4.43 billion。

Capital clustering forms a microcosm,AI for Science has officially been elevated to a national-level strategic direction.In the 15th Five-Year Plan, “artificial intelligence leading the transformation of scientific research paradigms” is explicitly included in the national medium- and long-term plan.

US Launch“Genesis Project,” a national-level scientific research initiative tackling critical challenges, directly targetsAIfields of scientific discovery driven by; the European Union has also successively introduced the "ApplicationAIStrategy》and《ScienceAIStrategy” two major documents; Japan and Singapore have also successively introduced relevant policies.

Nowadays,AI for ScienceCapital, policy, and industry forces are now in resonance; the industry’s inflection point is imminent.

This article reviews materials related to the AI for Science (AI4S) field and selects 10 noteworthy companies each from China and overseas to help readers understand capital flow trends. It does not constitute any investment advice, guidance, or commitment.

Overseas AI for Science Companies (10)

Google DeepMind: Leader in AI4S Underlying Algorithms and Foundational Models

The absolute technological leader in the global AI4S field, leveraging the vast resources of its parent company Alphabet, it holds a dominant position in fundamental scientific discovery and multimodal large models.

Its developed AlphaFold2 basically solved the "protein folding" problem that had plagued the biological community for 50 years in 2020.. With this breakthrough, scientists Demis Hassabis and John Jumper were awarded2024 Nobel Prize in Chemistry.

Google DeepMind has achieved a series of landmark accomplishments in the field of AI for Science, with itsIts influence has expanded from fundamental biology to multiple disciplines, including physics, materials science, and mathematics,GNoME、AlphaEvolve and a series of other models.

Isomorphic Labs—GeneralAI Drug DiscoveryEngine

A Global Benchmark in AI-Driven Drug Discovery, Isomorphic Labs Directly BuildsBuilt on DeepMind’s Nobel Prize-winning achievement, AlphaFold, with the mission to “reimagine the entire drug discovery process from scratch using AI.”

The company's milestone-based technologyAlphaFold, independently developedIsoDDE UnificationAI Drug Design Engine,IsoDDE Protein Prediction- Ligand structure accuracy rateAlphaFold3more than twice as much, with binding affinity prediction accuracy on par with traditional physics-based simulations“Gold Standard”, with speed increased by several orders of magnitude。

With IsoDDEEngine,Isomorphic Labs Built an independent pipeline covering core areas such as oncology and immunology, with the first batchAI Design of Candidate Drugs Expected2026 Entering at Year-EndIPhase Clinical TrialIsomorphic Labs. Meanwhile, the company has established deep strategic collaborations with top global pharmaceutical enterprises。Cumulative potential value of collaborations with Eli Lilly and Novartis exceeds30100 million USD。

At the capital level,Isomorphic LabsContinuously Breaking Industry Records:2025Year3Monthly Completion6$100 millionASeries round financing,2026Year5Monthly Completion21 hundred million USDB -round financing, setting a global recordAIPharmaceutical Sector Financing Reaches New High。

SandboxAQ – AI + Quantum Platform

Originating from Alphabet’s X lab, it was spun off as an independent entity in 2022, focusing on the integration of quantum technology and AI.Pioneered the LQM (Large Quantitative Model) technical paradigm,Specializes in solving real-world physical problems that are difficult for traditional AI to handle.

To date, SandboxAQ has built a comprehensive product portfolio comprising AQBioSim (a drug discovery platform), AQChemSim (a chemistry and materials simulation platform), AQtive Guard (quantum-safe encryption), and AQNav (a quantum navigation system), covering multiple fields including biopharmaceuticals, chemistry, and materials science.

In June 2026, the company secured $500 million in dedicated R&D funding under the U.S. CHIPS Act to develop novel chemicals and materials required for semiconductor manufacturing. To date, the company has completed multiple rounds of financing, with total cumulative funding exceeding $950 million.

Periodic Labs—AI-Driven Autonomous Materials R&D Platform

Co-founded by former OpenAI researcher Liam Fedus and Ekin Dogus Cubuk, lead of the DeepMind GNoME materials project, it is dedicated to bridging the data loop between theoretical simulations and physical experiments, creating an “AI Scientist” system capable of autonomously conducting scientific discovery.

Company Structure“AI Scientist + Autonomous Robotics Laboratory”Closed-Loop R&D System: AI proposes hypotheses for materials exploration, while an automated experimental platform conducts thousands of physicochemical experiments in batches, continuously generating proprietary empirical data to feed back into model iteration and optimization. According to insiders, the company has secured its first batch of customers in the semiconductor industry and achieved partial revenue generation.

In September 2025, the company completed a $300 million seed funding round, with investors including prominent firms such as a16z, DST Global, NVIDIA NVentures, and Accel, as well as individual investors like Jeff Bezos, Eric Schmidt, and Jeff Dean. By May 2026, the company was advancing a $500 million financing round at a $7.5 billion valuation, marking a nearly six-fold increase in valuation in less than eight months.

Xaira Therapeutics – The Frontier of Virtual Cell Technology

The company is composed ofNobel Laureate in Chemistry, Pioneer in Protein DesignDavid Baker Professor and Former Chief Scientific Officer of GenentechMarc Tessier-Lavigne Co-founded by PhDs, with a core team comprisingAI Top Expert in Protein Design, Antibody Engineering, and Clinical Development。

Xaira TherapeuticsUpon its debut, it immediately10 $100 Million Seed Funding Round Sets New Global RecordAIPharmaceuticalsAngel RoundRecord，Highlighting the Capital Market's AI Extremely Bullish on the Large-Molecule Drug Sector.

Technologically, the company has consistently bet on the cutting-edge field of virtual cells. Previously releasedX-Cell，It is the largest virtual cell model to date, with globally leading performance across all metrics.In addition, the company,Possesses a technology platform for the de novo design of novel antibodies, enabling the development of therapeutics that differ from traditional antibodies and can bind to new targets.。

Pipeline-wise,The company has made clear strategic choices in therapeutic areas and drug modalities.，The company currentlyFocusing on Inflammation and Immunology ScienceDue to the complexity of the disease, there are significant unmet medical needs.。

Lila Sciences—Scientific Superintelligence Platform

A leading enterprise in the global autonomous laboratory sector, dedicated to building the world’s first scientific superintelligence platform and fully autonomous laboratory for life sciences, chemistry, and materials science.

Its AI Science Factory integrates advanced automation equipment and robotics technology, enabling 7×24 hour uninterrupted experimentation to continuously provide real experimental feedback for AI models, forming“Hypothesis Generation – Automated Experimentation – Data Feedback Iteration”Closed Loop. Leveraging this system, the company identified a novel green hydrogen catalyst in just four months and completed the discovery and validation of multiple new antibodies and novel carbon capture materials, continuously accumulating proprietary empirical datasets.

Bloomberg reported that Lila is negotiating a Series B funding round of approximately $2 billion at a pre-money valuation of $8.5 billion. Prior to this, the company had raised a total of $550 million in cumulative funding.

PhysicsX—Leader in Industrial AI

A representative enterprise in the global industrial AI simulation sector, deeply engaged in the R&D of high-end equipment, providing AI-native engineering simulation solutions for the aerospace, automotive, semiconductor, new energy, and advanced materials industries.

its self-developedLarge Physical Model (LPM) Platform,High-fidelity simulations that previously took days can be compressed to second-level inference, enabling engineers to batch-validate thousands of design solutions. The platform seamlessly integrates into existing enterprise R&D workflows, balancing simulation accuracy with iteration efficiency, and continuously supports the forward design and optimization of complex hardware products.

To date, the company has established deep strategic partnerships with industry giants such as Siemens and NVIDIA, with its products widely applied in aerospace, automotive, and semiconductor sectors. In June 2026, the company completed a $300 million Series C financing round, achieving a post-money valuation of approximately $2.4 billion. The round was led by Temasek, with continued increased investment from strategic industrial shareholders including NVIDIA, Siemens, and Applied Materials.

Chai Discovery—Macromolecular Foundation Models

Chai Discovery Established in2024Year3Month, headquartered in San Francisco, California, USA, is a company dedicated to leveraging AI to empower drug development.

The success of Chai Discovery is rooted in its disruptive breakthroughs in the core technologies of AI-driven drug design. The Chai series of models published by the company have overcome the limitations of previous AI drug discovery models,Capable ofEnhancement of Drug Target Binding Affinity to Original Levels100 times.The latest Chai-3 success rate is double that of the previous generation, exceeding approximately 30%.

The company, founded just two years ago, has already completed four rounds of financing, with a valuation exceeding $3.8 billion. Currently,Chai Discovery has partnered with top global biopharmaceutical companies such as Pfizer and Eli Lilly.

Edison Scientific——Autonomous AI Scientific Agent Platform

A representative enterprise in the global AI for science sector, spun off from the non-profit organization FutureHouse, dedicated to deeply integrating AI into the entire workflow from basic discovery to clinical trials to accelerate scientific breakthroughs.

Its core productsKosmosIt is a collaborative AI scientist system composed of specialized AI agents, compressing doctoral-level research that originally took months into just 12 to 48 hours. Kosmos has made seven original scientific discoveries that have been independently verified externally, spanning multiple fields including neuroscience, materials science, metabolomics, and statistical genetics.

To date, the company has completed a $70 million seed financing round, achieving a valuation of $250 million. It has attracted attention from multiple leading global enterprises and established a strategic partnership with the biopharmaceutical company Incyte, continuously driving the intelligent transformation of scientific research paradigms.

CuspAI—An AI-Driven Unicorn in New Materials

CuspAI was founded in 2024 by its two co-founders: Dr. Chad Edwards, a chemist, and Professor Max Welling, an expert in the field of artificial intelligence.

Domestic AI for Science Enterprises (10 Companies)

XtalPi——"Quantum Physics+AI+Machine"Robot" Innovation Platform

China's earliest focus on artificial intelligence+one of the pharmaceutical companies, which first achieved “AIParticipated in Experimental Design+Robot-Executed Experiments+Experimental Data FeedbackAI” has gained recognition from mainstream research institutions both domestically and internationally, empowering multiple sectors including pharmaceuticals, new materials, and agriculture.

Its intelligent autonomous experimental platform enables high-throughput, precise experimentation, operating 24/7, boosting labor efficiency by fivefold and enhancing data collection capacity by fortyfold. Currently, the company has deployed over 300 robotic experimental workstations, forming the world’s largest cluster of AI-driven experimental workstations in commercial operation.

At the customer level, XtalPi has established collaborations with 17 of the global top 20 pharmaceutical companies, serving over 300 international enterprises and research institutions. The company is involved in the R&D of more than 100 innovative drug pipelines, with multiple candidates having entered clinical stages, covering various serious and chronic disease areas such as oncology, autoimmune disorders, and neurodegenerative diseases.

To date, XtalPi has attracted investment from top-tier institutions including Tencent, Sequoia China, and Microsoft, and listed on the Hong Kong Stock Exchange in 2024. In 2025, the company reported revenue of RMB 803 million, becoming the first profitable HK-listed company in the AI for Science sector.

Jitai Tech – The World’s First AI-Driven Drug Delivery IPO

As a global leader in AI-driven drug delivery innovation, Metagenius’s core competitiveness stems from its independently developedThe World's FirstAI Nanodelivery Platform NanoForge，The platformPossessThe World's Largest and Most Diverselipid library, containing over1000A myriad of structurally diverse lipids.

The company has pioneered the industry's first implementation of8Precise Targeted Delivery to Key Organs and Tissues——Liver, Lung, Heart, Muscle, Tumor Tissue, Immune System, Central Nervous System, Gastrointestinal Tract。

The Company ownsExceeding10pipeline products, covering oncology, immunology, central nervous system, metabolic diseases, respiratory system, cardiovascular, and musculoskeletal disorders.

In terms of commercialization, MetagenTech adopts a dual-drive model of “platform collaboration + product collaboration,” establishing partnerships with over 30 pharmaceutical and biotechnology companies globally and laying out more than 10 pipeline products. In June 2026, MetagenTech reached a licensing agreement with Deerfield, a leading global health investment platform, with a potential total transaction value exceeding $1.6 billion.Set a new record for the highest single overseas licensing deal value for a preclinical TCE project among Chinese pharmaceutical companies.

In May 2026, it successfully listed on the Hong Kong Stock Exchange as the “world’s first AI-driven drug delivery stock,” raising over HK$2.1 billion in its IPO, with public subscription oversubscribed by more than 6,900 times, setting a new record for healthcare listings on the Hong Kong Stock Exchange in 2026.

DP Technology — AI4S Intelligent System

One of the earliest enterprises in China to deploy AI for Science, it pioneered the “multi-scale modeling + machine learning + high-performance computing” paradigm, creating a new generation of multi-scale simulation platforms and industrial infrastructure for the biopharmaceutical, energy, and new materials industries. Its core large models, such as Uni-Mol and Uni-RNA, have significantly enhanced the accuracy and efficiency of drug target prediction and RNA design.

Relying on“DeepModeling·YuZhi” Scientific Discovery Intelligent EngineBuilding on this foundation, DP Technology has constructed a capability system centered around the “Bohr Scientific Research Space Station,” encompassing four core competencies: “Read, Compute, Execute, and Intelligence.” Its platforms have served over 1,000 universities worldwide and more than 3 million researchers, covering key fields such as life sciences and physical sciences. The company empowers the intelligent R&D transformation of more than 150 industry leaders, including Fosun Pharma, CATL, and BYD.

In 2025, the company completed a Series C financing round exceeding RMB 800 million, with investors including Hillhouse, Baidu Venture Capital, Lenovo Capital, and the Beijing Artificial Intelligence Industry Fund, among other industrial and state-owned capital firms, as it continues to iterate its intelligent engine for scientific discovery.

Molecular Mind — Biomacromolecule Infrastructure

Founded by internationally renowned computational biologist Jinbo Xu, it is a globally leading AI-powered protein design platform company. ItsSelf-developed AI-native bioeconomy operating system MoleculeOS,Built on the NewOrigin (Darwin) large language model, it deeply integrates AI with first-principles thinking to establish an engineered closed loop of “AI design–wet lab validation–data feedback–model evolution.”

On June 4, the company launched MMDesign, a novel AI-powered platform for de novo design of biologic drugs. The platform enables practical-level de novo design and experimental validation of nanobodies targeting high-value targets such as cytokines, immune checkpoints, and GPCR multi-transmembrane receptors. It achieves a target-binding success rate exceeding 90%, with optimal affinity reaching the picomolar level.

As of now, Molecular Mind has completed its Series A financing round, with cumulative funding exceeding USD 100 million. New investors introduced include Blue Bridge Capital, Pudong Venture Capital, and COFCO Emerging Industries Fund, making it a highly favored premium target in the AI + biotechnology sector.

Hongzhiwei Technology — Physics-AI Intelligent R&D Platform for Materials

A pioneer in domestic multi-scale material simulation technology, with genetic engineering at its core and the large-model agent “Hongyuan” as its vehicle,Build a transferable, replicable, and scalable infrastructure for intelligent materials R&D, with benchmark cases established in new energy, semiconductors, and high-end manufacturing.

Its flagship product, Phase Lab, leverages authoritative thermodynamic and industrial material databases to enable end-to-end analysis spanning “composition design–microstructure evolution–property prediction,” thereby significantly enhancing R&D efficiency for advanced high-end materials such as steel alloys, nickel-based superalloys, aluminum alloys, and copper alloys.

On the commercial front, the company’s platform processes an average of over 1 million computational tasks per month, serving nearly 800 universities and more than 50 leading enterprises, including CATL, Yangtze Memory Technologies, and Dongfang Electric.

To date, the company has secured investments from multiple institutions, including Guozhong Capital, Shenzhen Capital Group, Fortune Capital, and Shengyu Investment, emerging as a key gateway for the commercial deployment of AI for Science (AI4S).

Deep Principles—AI for Materials: Global Tech Pioneer

One of the key innovative enterprises in China’s AI for Science sector, centered on the core philosophy of “deep learning + first-principles,” it pioneered the dual-driven algorithmic approach of “LLM + Diffusion,” dedicated to revolutionizing chemical and materials R&D and enhancing energy conversion efficiency.

Leveraging this pathway to engineer deep mechanistic insightsAgent Mira for Materials Discovery and the ReactiveAI Platform,Six Proprietary Algorithm Modules Comprehensively Covering Six Core Stages: Material Generation, Precision Calculation, Process Control, Broad-Spectrum Screening, Synthesis Navigation, and High-Throughput Experimentation.

As of now,The company has completed nearly RMB 1 billion in Series A financing., secured support from top-tier investors including Hillhouse Venture Capital, Vertex Ventures, and Baidu Ventures (BV), and established deep collaborations with industry partners such as L'Oréal, Shanhai Innovation, and XtalPi.

Silang Technology – AI4S Computing Infrastructure

Spun out of the Institute of Automation, Chinese Academy of Sciences, it is one of the few companies globally to achieve full-stack independent research and development in scientific computing, withMaPU Architecture with 100% Independent Intellectual Property RightsCentered on the core, we build an integrated product system encompassing chips, complete devices, simulation software, and datasets.

Based on the MaPU, Silan Technology has developed "Tianqiong," China's first domestically produced 3D scientific computer. Its measured performance in molecular dynamics simulations reaches 5–10 microseconds per day, comparable to that of the US Anton2 supercomputer.

Currently, the company has deployed a number of “Tianqiong” 3D scientific computers across China. The Yangtze 3D Scientific Computing Center, jointly established with Yangtze Science and Computing, became operational in 2023, serving over 200 leading research teams and providing computing power and technical support for fields such as biopharmaceuticals and new materials.

On July 7, Silang Technology’s IPO on the STAR Market was accepted by the Shanghai Stock Exchange, with a planned fundraising of RMB 4.426 billion. If successfully listed, Silang Technology will become A-share’s “"The First Stock of Scientific Intelligence"。

BioMap — Foundation Large Model for Life Sciences

Founded by Robin Li, the founder of Baidu, it is dedicated to building next-generation AI infrastructure for the life sciences industry, comprehensively enhancing the efficiency of drug and biotechnology discovery. It currently serves over 800 institutional clients worldwide, covering multiple vertical sectors including antibodies and proteins, innovative therapies and precision medicine, synthetic biology, and cutting-edge scientific research.

Released in 2024xTrimo, the foundational large model for life sciences,Set a new global record for the largest life sciences AI foundation model at the time, with 210 billion parameters.xTrimoAchieved state-of-the-art (SOTA) performance in over 300 prediction tasks across more than 20 subfields of life sciences, and won the championship in the inaugural Global Virtual Cell Challenge.

In March this year, BioMap submitted its listing application to the Hong Kong Stock Exchange on a confidential basis. From large models to IPO, BioMap has carved out a unique path of “AI-defined life sciences.”

Longxun Kuangteng – Pioneer in Atomic-Precision Simulation

A leader in the research, development, and innovation of domestic computational materials simulation software tools, with its independently developedPWmat: Quantum Material Calculation SoftwareCapable of performing electronic structure calculations and ab initio molecular dynamics simulations, it is applicable to crystals, defect systems, semiconductor systems, metallic systems, nanosystems, quantum dots, clusters, and molecular systems, achieving internationally leading standards.

On July 7, the company launched MatPL-pro, a professional-grade GPU inference engine for NEP machine learning force fields. By reconstructing hardware-software co-designed low-level operators from the ground up, it achieves a 2–3 fold leap in inference speed while maintaining the accuracy of NEP models, thereby fully unlocking hardware potential.

To date, Longxun Kuangteng has completed multiple rounds of financing, including an RMB 100 million Series A round and a tens-of-millions RMB Series A++ round. The company now serves over 200 research and industrial clients, continuously driving the deep intelligent transformation of material R&D paradigms.