XtalPi AI Autonomous Lab Station Deployed at Sinopec, Setting a New Benchmark for AI-Driven Materials Characterization

Recently, XtalPi, a globally leading artificial intelligence (AI) + robotics platform for drug and new material research and development,2228.HK) announced that the “Intelligent Autonomous Experimental Workstation for Physical/Chemical Adsorption Analysis,” jointly developed with Sinopec (Shanghai) Petrochemical Research Institute Co., Ltd. and Beijing Jingwei Gaobo Instrument Co., Ltd., has officially commenced operations. This workstation drives chemical R&D through intelligent and automated material characterization, establishing a critical physical carrier and embodied intelligence foundation for the “AI for Science” (AI4Science) material discovery engine. For the first time in an industrial-scale scenario, it fully integrates high-throughput data collection for material characterization and autonomous experimental decision-making into an intelligent system comprising “AI + Robotics + Multi-Agent,” marking a pivotal step toward Physical AI and laying the groundwork for building an intelligent, autonomous R&D system for materials that is “perceivable, operable, and evolvable.”

This achievement prominently embodies three core technological labels:Fully automated, unmanned autonomous experimentation; a closed-loop system for native high-quality data; and a modular intelligent foundation integrating hardware and software. This collaborative achievement transforms material characterization—a high-frequency, critical-need process—from experience-driven, manual workshop-style operations into an industrial-grade intelligent system driven by data and AI decision-making, achieving order-of-magnitude improvements in characterization throughput, data accuracy, and experimental safety.

Strong Alliance,

Setting the Industry Benchmark for AI + Materials Characterization

In trillion-dollar industries such as petrochemicals, new energy, and environmental protection, precise material characterization is a high-frequency, essential step in the R&D process that heavily relies on manual labor and trial-and-error based on individual experience. Porous materials (such as molecular sieves, activated carbon, and metal-organic frameworks), serving as the most widely used and largest-volume catalysts and adsorbent supports in the petrochemical industry, have properties like specific surface area, pore size distribution, and surface acidity that directly determine the efficiency and selectivity of core processes such as catalytic cracking, hydrotreating, and gas separation. These materials are thus regarded as the “chips” of the petrochemical industry. Traditional characterization methods rely heavily on manual operations, resulting in low efficiency and throughput. Moreover, data consistency is poor due to subjective individual experiences and operational variations, preventing massive experimental datasets from effectively feeding back into AI models. This has become the biggest data bottleneck constraining the intelligent R&D of new materials.

To address this challenge, XtalPi andSinopec (Shanghai) Research Institute of Petrochemical TechnologyIn close collaboration with Jingwei Gaobo, we have built a collaborative innovation ecosystem that creates a closed-loop process spanning from the definition of industrial problems and R&D of high-end instruments to AI-driven autonomous experimentation. This transforms traditional analysis processes, which rely on intuition and experience, into quantifiable and reproducible scientific methods, directly addressing core pain points in the industry.

In this collaboration,Sinopec (Shanghai) Research Institute of Petrochemical TechnologyLeveraging its extensive R&D expertise and deep understanding of industrial scenarios in the petrochemical sector, it precisely identifies the core industrial requirements for high throughput, high precision, and high safety, while providing realistic and rigorous validation scenarios. As a leading Chinese manufacturer of physical/chemical adsorption instruments, Jingwei Gaobo provides the workstation with a high-precision analytical instrument hardware foundation that serves as its “precise sensory system,” ensuring the accuracy and reliability of AI-driven autonomous experimental data. As a benchmark enterprise in the AI for Science field, XtalPiBy integrating an “AI Brain” and autonomous robotic control capabilities, and leveraging its closed-loop, autonomously operating intelligent system based on the “AI + Robotics + Multi-Agent” framework, standalone instruments are upgraded into “AI Materials Characterization Scientists” capable of independently planning workflows, analyzing data, and making decisions. Through deep, innovative collaboration among the three parties, a new benchmark for intelligent materials characterization is established for the industry.

AI-Driven Autonomous Experimentation,

Reshaping the Core Paradigm of Material Characterization

Built on an AI-native core, this workstation integrates flexible and scalable robotic experimental stations into a unified standard for intelligent algorithm experimentation, forming an integrated “hardware-software” intelligent system. It achieves a fully automated, closed-loop workflow from sample processing to data analysis, thereby establishing a “high-quality data foundation” that empowers AI for Science in the field of materials science:

• High-Throughput Autonomous Experimentation:

The workstation supports intelligent expansion of multi-channel physical/chemical adsorption modules, autonomously allocating tasks under unified AI scheduling to achieve unmanned, uninterrupted 24/7 operation. This results in an order-of-magnitude increase in daily sample throughput, completely liberating researchers from repetitive labor.

• Native High-Quality Data Closed-Loop:

The unified experimental standards for robotic operations fundamentally eliminate human error. The system integrates core intelligent algorithms to ensure ultra-high data consistency and reproducibility, providing high-quality data support for AI model training, making every set of data traceable and reusable.

• Intrinsic Safety and Intelligent Analysis:

Unmanned operation significantly reduces operational risks associated with traditional experiments, such as liquid nitrogen handling, thereby enhancing the intrinsic safety of the laboratory. The system also automates data analysis, result normalization, and trend modeling, laying the foundation for the digital and intelligent transformation of materials research and development.

This system is built upon XtalPi’s mature, customizable AI-driven autonomous experimentation platform. Featuring over 30 modular functional components that can be flexibly combined, the platform has earned recognition from more than 300 leading enterprises and research institutions worldwide.

Building the Data Foundation,

Advancing the Industry Toward Physical AI

“Intelligent Physical/Chemical Adsorption Analysis Autonomous Experimental Workstation” achieves complete capture and structured accumulation of experimental data for the first time through full-process digitization and autonomous operation, transforming originally scattered and ephemeral personal experience into accumulative and reusable data assets. Its significance lies not only in improving efficiency at individual points but also in providing high-quality “data fuel” for constructing AI models capable of understanding and interacting with the physical world, thereby enabling rational materials design to truly uncover structure–property relationships from massive experimental datasets and establish a closed-loop evolutionary capability of “experiment–data–model–prediction.”

Business Leader of Sinopec (Shanghai) Research Institute of Petrochemical Technology Co., Ltd.stated: “By integrating AI and autonomous experimentation technologies into the core stages of catalyst development, this intelligent workstation not only achieves a significant leap in efficiency and data accuracy, but also provides a novel research paradigm for exploring high-performance, low-energy-consumption catalysts. It supports the industry’s green and low-carbon transition under the ‘Dual Carbon’ goals and serves as a benchmark practice for the intelligent transformation of chemical R&D.”

Head of Automated Innovation Business, XtalPistated, “This represents a significant breakthrough for AI for Science in the petrochemical industry. What we are building is a system that enables AI to continuously generate and digest high-qualityquantitative data, and drive self-iterating R&D infrastructure. Starting from this point, this model will expand from adsorption characterization to broader materials R&D scenarios, promoting the transition of materials development from ‘experience-driven’ to ‘data-driven and intelligence-driven’, thereby empowering source innovation across more industries.”

Relevant Person in Charge at Jingwei Gaobostated: “This collaboration serves as a paradigm for the integration of domestically produced instruments with AI-driven autonomous experimental technologies. We will continue to enhance the intelligence level of our products, providing global users with more comprehensive digital and intelligent solutions for material characterization.”

In the future, XtalPi will continue to deepen its AI for Science platform technologies, driving materials R&D from empirical exploration toward deeper mechanistic understanding and rational design. By continuously refining its R&D flywheel—centered on vertical-specific AI models, large-scale robotic laboratories, and Multi-Agent systems—in industrial settings, XtalPi aims to accelerate the comprehensive intelligent transformation of China’s materials R&D sector, injecting sustained momentum into the national strategy for building a scientific and technological powerhouse and fostering high-quality industrial development.

Sinopec (Shanghai) Research Institute of Petrochemical Technology Co., Ltd. is a comprehensive petrochemical research institution directly affiliated with Sinopec Group. It is dedicated to the technological R&D and industrial application in fields such as petrochemicals, synthetic resins, synthetic fibers, and catalysts, providing core support for Sinopec’s technological advancement and industrial upgrading.

● About Jingwei Gaobo●

Founded in 2004, JWGB (Jingwei Gaobo) is a scientific instrument manufacturer deeply engaged in the global market, with its headquarters and R&D center in Beijing, production bases in Tianjin, and subsidiaries in the United States and Germany. The company’s product portfolio covers key areas such as adsorption instruments, thermal analysis instruments, X-ray diffractometers, and reaction systems. As an innovator driving progress in the global field of material analysis instruments, JWGB consistently adheres to its mission of “providing high-quality, user-friendly, and cost-effective advanced measurement instruments for the research and manufacturing of new materials.” By continuously exploring frontier technologies in material characterization and accelerating technological iteration, JWGB delivers diversified products and service solutions to customers worldwide.