On December 10, leveraging the core capabilities and large-scale implementation achievements of the Sushine Yaoguang Industrial Large Model, Luculent was ranked 15th on the "2025 China Industrial AI Leading Enterprises" list.

Not long ago, the company was also ranked 9th among "Industrial Software Enterprises" and 14th among "Industrial Internet Solution Providers."

For twenty-six years, Luculent has evolved from deepening its expertise in the industrial software sector, to pioneering the industrial internet track, and now to leading the frontier of industrial AI. It has consistently "stayed the course" while actively "embracing change," carving a path of continuous technological innovation and advancement.

Three "Transformations", from Starting with Industrial Software to Being Led by AI. 

In the 1990s, China's industrial software market was almost monopolized by foreign giants, posing severe challenges to information security and autonomous control.

"Changing this situation of being constrained by others is the mission on the shoulders of China's industrial software professionals."

Driven by this ideal, Wu Aibin founded Luculent in 1999. In 2003, after 14 months of dedicated R&D, the technical team, with a tenacious "never-give-up" spirit, pioneered the development of the LiEMS intelligent management and control platform—the first in the industry to cover group-wide, full-process, integrated operations. Gradually, they built a complete domestic industrial software system including the investment-construction-operation integration platform, a real-time database, and Enterprise Asset Management (EAM) software. Notably, the Intelligent Power Generation Production Software System has been deployed in thousands of enterprises, including major central state-owned power generation groups and local energy groups, and has been recognized as a national-level "Manufacturing Single Champion" product.

In 2016, as the concept of industrial internet began to emerge, Luculent keenly recognized the trend and subsequently launched the Sushine Industrial Internet Platform. It creates economic and social value for enterprises in areas like safety production, predictive warning, equipment maintenance, and energy conservation, fostering an open and collaborative industrial ecosystem. It is precisely this philosophy of "adding value for customers and empowering the industry" that rapidly established Luculent Sushine as a member of the national-level "Cross-Industry & Cross-Sector" platform. It participates in national "Unveiling the List and Appointing the Commander" projects and leads major initiatives like the Ministry of Industry and Information Technology's Industrial Internet Innovation Project.

When the wave of artificial intelligence arrived, Luculent's choice remained "returning to the essence of industry." It deeply integrated AI with industrial internet technology, applying it to high-value scenarios and building a collaborative "cloud-edge-device" system. The company dedicated itself to creating its flagship—the world-renowned, top-15 ranked Sushine Yaoguang Industrial Large Model—along with AI-powered "killer app" industrial applications and agents for intelligent monitoring, equipment fault prediction and diagnosis, and energy consumption optimization, driving manufacturing up the value chain.

Hi, let's be friends~ I'm the Sushine Yaoguang Industrial Large Model

▲Click to watch the video

These three "transformations" have not only accumulated profound technological barriers in the industrial software industry but also provided a solid foundation for the leap into the high-end AI industry. Today, Luculent has transformed into a leading "head goose" in the industry, powered by the trio of industrial software, industrial internet platform, and industrial AI.

A "Full-Stack" Answer

Solving the Difficulties of Industrial AI Implementation

The path to implementing industrial AI is long and fraught with challenges.

As the 15th Five-Year Plan period begins, AI application is moving into the industrialization and scale phase. Industrial production places extremely high demands on reliability and robustness, and large models must also overcome "hallucinations." Enterprises are at varying stages of digital transformation, often with weak data governance foundations, severe data fragmentation and silos, making it difficult to unlock AI's potential value.

How are these problems solved?

Luculent's established industrial AI empowerment system—"One Foundation, Three Layers of Cognition, Four Types of Scenarios, Full-Stack Closed Loop"—has provided the "answer."

A Solid Intelligent Foundation, Rooted in Industry. The Sushine Yaoguang Large Model is not a simple fine-tune of a general model. It injects the "soul" of massive equipment parameters, process specifications, fault cases, and expert experience into the model, enabling it to "understand" equipment language and "comprehend" process logic. Through technologies like knowledge graphs and hybrid retrieval, it improves the accuracy and reliability of outputs from the source, suppressing "model hallucinations." "Yaoguang" has accumulated over 5,000 fault cases, more than 2,800 mechanism models, and tens of thousands of diagnostic reports, making it a veritable industrial "living dictionary" and "knowledge base."

Three Evolving Layers of Cognition, Forging Professional Insight. Enhance Enterprise Cognition:Utilizing enterprise organization data, user profiles, and business process data to strengthen model learning, becoming an internal "know-it-all." Enhance Professional Knowledge Cognition:Enabling the large model to master industry standards and process specifications, equipping it with professional auditing and judgment capabilities for work permits, reports, etc. Enhance Problem-Solving Cognition:Training and integrating expert chains-of-thought to handle complex tasks like fault diagnosis and root cause analysis, accomplishing understanding context, mastering rules, and autonomously executing tasks in one go.

Precise Empowerment in Four Scenario Types, Achieving Value Closure. In interactive Q&A, the intelligent assistant generates precise answers and professional reports, improving the efficiency of equipment fault Q&A and maintenance knowledge retrieval by 30%. In business auditing, it automatically reviews "two tickets" (work and operation permits) and contracts, identifying logical errors and risk clauses, improving audit efficiency by 50%. In operational optimization, agents analyze real-time data, pinpoint root causes, and automatically issue optimization commands to reduce costs and increase efficiency. In data governance, unstructured documents are automatically parsed into structured knowledge, feeding back into the entire system, forming a continuous enhancement loop from data to knowledge, and from knowledge to intelligence.

Full-Stack Technical Closed-Loop Support, Driving Large-Scale Implementation. A full-stack toolchain covering knowledge management, agent development, model fine-tuning, and application scenario construction has been developed, forming a collaborative model of "Platform + Large Model + Agents + Scenario-based Apps," bridging the "last mile" from intelligent decision-making to field execution. A comprehensive data set, model, and industrial application marketplace provide customers and developers with a flexible, one-stop "capability supermarket," promoting the sharing and adoption of digital technology applications.

Luculent's Industrial AI Empowerment System has been widely adopted by leading central and state-owned enterprises in sectors like power, coal, steel, and equipment manufacturing, providing a viable path for the deep implementation of industrial AI.

China's AI industry has entered a new stage of vigorous development. Luculent is once again charging at the forefront, promoting the deep integration of industrial AI with the real economy. The aim is to enable industrial systems to truly "know what you think, understand your concerns, and solve your problems," helping enterprises seize the initiative in the new round of industrial revolution.