On October 15, Luculent was invited to attend the 2026 Power Industry Digital & Intelligent Transformation and Smart Grid Industry Ecosystem Launch Conference (Nanjing), hosted by the Technology Center of China Electricity Council. Together with central SOEs in the energy and power sector, key industrial chain enterprises, as well as experts and scholars from universities and research institutions, Luculent gathered to explore cutting-edge technologies, share industry achievements and jointly plan the future of the industry.
At the main forum, Luculent delivered a special report on the thinking and practices of digital and intelligent transformation in the power generation industry.
“Currently, the digital and intelligent development of large power generation groups has shifted from the platform construction phase to a new stage of value-oriented application deployment. However, many projects fail to deliver expected outcomes after going live. The major pain points include the inability of models to perform autonomous iterative training, lack of effectiveness evaluation criteria, unexplainable black-box models, and difficulties in promoting and reusing standalone applications,” noted the company’s expert.
To address these challenges, Luculent focuses on four dimensions: safety, efficiency, low carbon and value. Leveraging the integrated cloud-edge-end collaborative architecture, it has built a highly resilient new-generation industrial AI real-time operating system.
Powered by a computing center foundation, the platform takes data centers and knowledge centers as high-quality industrial corpora. Using model algorithms and the intelligent agent development platform as production tools, it centers on industry models and various industrial intelligent agents. It drives business operations from process-driven to data-driven modes, and project delivery from static deployment toward dynamic iteration, converting cloud-based AI capabilities into tangible, usable productivity at energy sites.
Edge Devices (End Layer) — the “limbs” at the equipment endpoints, responsible for perception and execution. Traditionally, sensors only collected data and PLCs merely executed commands. Now upgraded into intelligent terminals embedded with lightweight AI algorithms, they are no longer just the “sense organs” for on-site information perception. They can independently make preliminary judgments and carry out on-site responses.
Site Edge Layer — the nerve center of power plants, delivering rapid on-site responses. Acting as an “on-site commander” with autonomous disposal capabilities, it performs real-time computing locally at the plant where data is generated, runs local AI models for inference and judgment, and directly implements fast closed-loop control of field equipment. Meanwhile, it feeds back accumulated on-site experience and knowledge to the cloud.
Cloud Layer — the global brain and ecosystem hub of the group, tasked with holistic oversight, trend analysis and decision-making. It aggregates data from all production lines and power plants for correlation analysis, generating strategic algorithm models to optimize the entire production value chain. A group-level intelligent agent library covering scenarios such as intelligent monitoring, fault diagnosis, operation optimization, risk identification and knowledge Q&A is established to enable continuous iteration: “learn while using, grow more accurate with usage and become increasingly intelligent”.
The seamless three-in-one closed loop — cloud training, edge inference and direct industrial control execution — supports penetrating supervision, intensive operation and unified group management. This enables AI to truly “sink down” to the field, helping enterprises including China Datang Corporation and China Energy Group strengthen intrinsic safety, achieve cost reduction, efficiency improvement and green low-carbon development.
Beyond breakthroughs in technical architecture, Luculent has upgraded from a tool-based platform to an ecosystem platform, and transformed from a platform builder into a platform operator, bringing forth four major model innovations:
New paradigm for capability trading: Build a model market operation system, accumulate high-quality datasets to continuously amplify model value and realize commercial monetization of model capabilities;
New paradigm for open ecosystem: Internally activate enterprise knowledge and enable experience reuse; externally cooperate with enterprises, research institutes, industry experts and diverse third-party partners to co-build the ecosystem;
New paradigm for crowdsourced research: Regularly hold modeling competitions to develop AI models targeting high-value business scenarios and elevate the overall industry modeling capability;
New paradigm for talent empowerment: Develop comprehensive professional training courses for different job levels, cultivating a large pool of interdisciplinary talents within groups who master both business and digital-intelligent technologies.
With dual innovations in technical architecture and operation models, Luculent leverages cloud-edge-end collaborative industrial AI capabilities to provide a replicable, high-value practical path for the digital and intelligent implementation of the power generation industry.