On June 27, the China Optics Valley AI Innovation Conference and the unveiling of the 2025 Forbes China AI Technology Enterprise TOP50 took place in Wuhan's Optics Valley. The event gathered over 50 prominent AI talents from around the world, including Turing Award winners, professors from Tsinghua University, foreign academicians of the European Academy of Sciences, renowned scholars, and founders of leading enterprises, who shared valuable insights on AI implementation. Among them, Dr. Li Nan, General Manager of the Solutions and Product Center at Griddynamics East Smart, delivered a keynote speech titled "The Breakthrough Formula and Practice of Industrial AI Super Productivity." In his address, he systematically dissected the core methodology for implementing industrial AI, drawing comparisons with leading AI enterprises such as Tesla, Baidu Intelligent Cloud, Lingyiwanwu, and Shengtong Technology.

Industrial AI has demonstrated disruptive advantages over traditional digitization. Dr. Li Nan highlighted that while significant strides have been made, the accuracy of AI outputs still requires improvement, making scenarios with human backup or non-core production the primary focus of exploration. With the maturation of model capabilities and small model tools, AI is poised to usher in new breakthroughs in decision-making and control, gradually achieving full-process intelligence. Additionally, the accumulation of industry-specific knowledge is crucial for value creation, and the input of high-quality industry data and industrial corpus can enable AI models to internalize expert experience, enhancing output accuracy and developing practical AI solutions.

Li Nan believes that industrial AI is accelerating its implementation through four paradigms: application intelligence, Agent/intelligent agent, industrial AI middleware, and intelligent equipment/embodied intelligence. During his discussion on the implementation path, Li Nan unveiled the "three-step" formula for the first time. The first step involves scenario screening, focusing on three key points: prioritizing high-value links such as research, production, supply, sales, and service based on value chain monetization; customizing solutions for five scenario types—knowledge-intensive, data-intensive, change-intensive, interaction-intensive, and creation-intensive—based on large model adaptability; and reusing existing IoT and data governance foundations to lower scenario investment thresholds. The second step is model collaboration, establishing a deep system of cooperation between large and small models, where large models handle macro decisions and small models perform high-precision tasks. The third step is capability integration, merging the five capabilities of Algorithm, Big Data, Computing, Domain knowledge, and Equipment (ABCDE) to expedite the bridging of the AI application implementation gap.

Li Nan's insights are grounded in the practical wisdom accumulated through Griddynamics East Smart's in-depth customer service. Taking TCL Huaxing's full-link intelligent service practice as an example, Griddynamics East Smart not only systematically connects key business scenarios but also achieves deep collaboration between large and small models. For instance, the "Equipment Knowledge Base Agent" small model deployed at TCL Huaxing conducts real-time diagnosis, supported by the "Xingzhi X-Intelligence Vertical Large Model" for R&D knowledge Q&A. Furthermore, the integration of the five capabilities and deep industry know-how in the pan-semiconductor field enhances system deployment effectiveness, helping TCL Huaxing achieve Level 4 in smart manufacturing maturity, the highest level nationwide.

As a leader in industrial intelligence, Griddynamics East Smart upholds the strategic pillars of AI + industrial software + intelligent equipment, deeply empowering the digital transformation of the pan-semiconductor manufacturing industry through core capabilities such as independent breakthroughs, platform-based empowerment, and intelligent transformation. During the conference, Dr. Li Nan was interviewed by Forbes China, sharing in-depth insights into Griddynamics East Smart's systematic implementation of industrial AI.

At the award ceremony, Griddynamics East Smart was honored as a 2025 Forbes China Emerging AI Enterprise, recognizing the effectiveness of its deep technical expertise and broad scenario coverage. Looking ahead, Griddynamics East Smart will continue to deepen its empowerment of industrial AI scenarios, contributing more replicable productivity leap formulas to the industry.