China's latest list for the fourth-level (CMMM) certification of intelligent manufacturing capability maturity has been released, with the Wuhu production base of LONGi Green Energy Technology Co Ltd (SHA: 601012) making the grade as the first photovoltaic module factory in the industry to attain this credential. As China accounts for over 80% of global PV module output, its sheer scale is well established, yet the transition from being "large" to being "strong" has become an essential task for the sector. Behind this certification lies a key question: how far has the intelligent manufacturing level of PV factories actually advanced? An exclusive interview was conducted with Hu Zhifeng, head of lean and intelligent manufacturing at LONGi Green Energy, and Deng Shengxiang, head of the Anhui base's Wuhu factory, to explore this topic in depth.
What does Level 4 CMMM signify?
The assessment is conducted in accordance with two national standards: the "Intelligent Manufacturing Capability Maturity Model" and the "Intelligent Manufacturing Capability Maturity Assessment Method." Maturity levels are divided into five grades from low to high: planning level, standardized level, integrated level, optimized level, and leading level. According to information from the public service platform for intelligent manufacturing evaluation and assessment, no enterprise has yet achieved the fifth-level "leading" designation, making Level 4 the highest level actually attained by domestic companies at present. As one of the participants in drafting this national standard and a trainer for intelligent manufacturing capability maturity assessors, Hu Zhifeng noted that the capability sub-domains in the upcoming new version of the national standard will expand from 20 to 24, with broader coverage and higher capability requirements. The guidance for enterprises building intelligent manufacturing capabilities will become more targeted, with heightened emphasis on applying technologies such as big data analytics, AI, and digital twins in manufacturing scenarios. Additionally, in terms of sustainable development capability requirements, energy management will be upgraded to energy-carbon management, underscoring the confidence and determination in the sustainable development of Chinese manufacturing enterprises.
Addressing the core demarcation between Levels 3 and 4, Hu Zhifeng explained: "Level 3 resolves data silos and process bottlenecks by requiring equipment and system integration to achieve data sharing across business domains, but it still relies on manual decision-making and parameter adjustment. Level 4, by contrast, treats data as a core production factor, building model algorithms based on big data analytics to achieve precise business forecasting and optimization—such as predictive equipment maintenance, AI-driven automatic parameter adjustment, intelligent dynamic production scheduling, and automated energy consumption optimization. Decision-making shifts from 'humans making decisions based on data' to 'systems optimizing autonomously.' While advancing from Level 3 to Level 4, what holds equal importance to data governance is the establishment of knowledge bases for each capability domain through a comprehensive knowledge governance system, thereby forming platform-level capabilities that support the invocation of intelligent agents."
Hu Zhifeng further noted: "At present, the automation rate in LONGi Green Energy's cell and module manufacturing processes has reached 90%, making it a highly automated factory. If we liken automation equipment to the human body, we need to implant a soul into that body, endowing each one with human-like intelligence—where the soul controls the body, and the body uses perception to feed data back to the soul. Many enterprises currently refer to their 'smart factories' as those applying artificial intelligence, big data, and the Internet of Things in localized business scenarios to achieve intelligence in certain areas. However, CMMM is a standardized capability evaluation system covering the entire value chain of research, production, supply, and sales. It can comprehensively measure the maturity of an enterprise's manufacturing system and assess the true level of intelligence in a factory. Of course, evaluating the degree of intelligent transformation in a manufacturing enterprise can also reference the smart factory gradient cultivation evaluation criteria, which correspond to CMMM levels (the four-tier smart factory gradient cultivation system includes: foundation level and advanced level corresponding to CMMM Level 2, excellence level corresponding to Level 3, and pilot level corresponding to Level 4 and above). The evaluation sets clear hard indicators—for instance, the excellence level requires AI application scenarios to account for no less than 30% and mandates the implementation of workshop-level digital twins; the pilot level requires AI application scenarios to account for no less than 60% and mandates the implementation of a complete factory-wide digital twin."
In Hu Zhifeng's view, this is the core yardstick for distinguishing genuine intelligent manufacturing. "Data, computing power, and algorithms are the three essential elements of artificial intelligence, with data serving as the prerequisite," Hu said. "Data is the foundational production factor for intelligent manufacturing. We must identify core influencing factors based on business needs, carry out targeted data collection, and conduct data governance." According to reports, LONGi Green Energy has built a unified data middle platform to support data consumption demands across various business units.
How the production line achieves "foreknowledge"
Stepping into the workshop at the Wuhu base, 16 fully automatic intelligent module production lines are arranged in sequence, with glass, solar cells, and encapsulant film flowing along conveyor belts while robotic arms rise and fall, gradually shaping a PV module. This is the core production base for LONGi Green Energy's BC modules, equipped with seven major AI engines covering areas such as scheduling, quality inspection, operations and maintenance, process control, and energy management. The certification covers 18 key links, including supply chain procurement, intelligent scheduling, equipment maintenance, safety and environmental protection, and logistics distribution.
"The base focuses on three priorities: product reliability, customer delivery assurance, and cost reduction with efficiency gains under the premise of guaranteed quality and delivery," Deng Shengxiang said. Digital transformation has provided strong support across all three areas. Module manufacturing involves numerous complex processes, with screen printing being particularly delicate. Cell welding requires printing insulating paste at designated positions, where uneven thickness or missed printing can affect product yield. At the Wuhu base, parameters such as squeegee pressure, operating speed, and paste wet weight are uploaded in real time, and the system monitors not just a single point but the overall trend. "When the trend begins to shift, we intervene and adjust in advance, eliminating problems at their inception and preventing defective products from the source," Deng noted.
The logic of quality management has also changed accordingly. "Previously, we relied on post-production inspection of finished products; now, with full-process AI, we identify risks during the production process in advance," Deng explained. The end-to-end AI quality inspection platform has increased defect detection efficiency by 80% while reducing the occurrence of quality issues across the entire line by 40%. The base has also introduced mature technologies from other industries, adapting facial recognition technology originally used in the security sector to the cell traceability environment, enabling precise traceability of every single cell throughout the entire process. In the past, equipment failures relied on manual point-to-point notification of engineers; now, the SCADA system automatically captures anomalies and quickly dispatches work orders. The system has also accumulated a comprehensive fault knowledge base, automatically pushing matched resolution plans when similar anomalies occur, supporting "predictive maintenance" and eliminating hidden failure risks before shutdowns. Data shows that the intelligent dispatch system has improved dispatch efficiency by 60%, significantly enhanced overall equipment effectiveness across the line, and driven production efficiency up by more than 20%.
Energy consumption management has also become more refined. For high-energy environments such as compressed air and HVAC systems, an AI intelligent control system integrating "air and water linkage" has been deployed, reducing overall energy consumption by 7%, with compressed air energy consumption down by 12%. "Customers place the highest value on product reliability and on-time delivery," Deng said. Digital tools have made the manufacturing process more transparent, noticeably improving customer recognition during factory audits and supervision, and boosting confidence in the shipped products.
Where does smart manufacturing head next?
Discussions of smart manufacturing inevitably touch on the industry's current realities. "The PV sector is an extremely 'involuted' industry," Hu Zhifeng candidly admitted. In July 2025, the sixth meeting of the Central Financial and Economic Affairs Commission called for "governing low-price disorderly competition among enterprises in accordance with laws and regulations." According to data from the China Photovoltaic Industry Association, module production in the first half of 2026 declined by 35.1% year-on-year. As the industry shifts from "comparing scale and competing on price" to "value competition," smart manufacturing is emerging as one of the key directions for breaking the deadlock.
"PV products carry a quality warranty cycle of up to 25 years—a constraint that few other manufacturing industries face," Hu noted. This extended warranty period forces full-process parameter traceability across the manufacturing chain, while next-generation cell-module technologies such as BC inherently have relatively narrow process windows. These multiple constraints impose high demands on intelligent manufacturing capabilities. LONGi Green Energy benchmarks its smart manufacturing against mature industries such as automotive and consumer electronics, while bringing in a large number of high-end manufacturing talents who draw on decades of technological accumulation from other sectors to make innovative adaptations. The continuous build-out of intelligent manufacturing capabilities has strongly supported the large-scale, stable production of high-difficulty BC modules.
This exploration extends beyond a single factory. In December 2023, LONGi Green Energy's Jiaxing base became the world's first "lighthouse factory" in the PV industry, and now both the Wuhu and Jiaxing bases have successively passed CMMM Level 4 certification. "The 'lighthouse factory' designation reflects the height of advanced technology application, while CMMM measures the breadth of systematic capability building—the two are complementary. Lighthouse factories explore cutting-edge technologies, while CMMM Level 4 transforms individual innovations into a replicable, comprehensive manufacturing system standard," Hu said. The 14 digital use cases accumulated at Jiaxing have been fully rolled out across the group, with some use cases having undergone multiple rounds of iterative upgrades as application needs have deepened—such as the MPT human resource transparency management system, which has already been updated to version 3.0.
Looking further ahead, the ultimate goal is the fifth-level "leading" designation, which no enterprise has yet attained. A few leading domestic companies have begun capability cultivation, and LONGi Green Energy plans to focus on this advancement in 2027. "The core logic of Level 5 is no longer the local optimization of a single factory, but rather the end-to-end global optimization of the industrial chain—extending upward to suppliers by integrating production, inventory, and delivery data across the supply chain; and connecting downward to customers by linking product operational data with service feedback, thereby maximizing resource efficiency across the entire chain," Hu explained. He drew an analogy: in the future, PV modules could transmit data back in real time just like smartphones, feeding quality management requirements and service optimization back into the manufacturing side.
Looking forward, Hu expressed strong confidence: "The progression logic remains the same—it's the replication of capabilities 'from 1 to N.' Last year, seven factories passed Level 3 certification. This year's goal is to have 3 to 4 factories achieve Level 4 certification, with Wuhu being among the first batch and the first module factory in the PV industry." He described industry competition as "tigers ahead and wolves behind," noting that LONGi is both opening up a time gap over its peers and simultaneously driving industry-wide smart manufacturing upgrades through shared standards and experience. This aligns closely with the smart factory gradient cultivation initiative, which encourages leading enterprises to drive "chain-style transformation" across the industrial chain. As "talking modules" emerge from these production lines, they connect not just a single factory but also the path of China's PV industry from scale leadership to leadership in quality and intelligent manufacturing.