"Previously, procurement relied on phone calls, experience, and personal connections. Now, it's driven by data, algorithms, and 'Kun'an,'" says Han Yaoxuan, a purchasing officer at Baotou Aluminum, smiling as he opens the "Lvxing Liantong" system, inputs his requirements, and an AI automatically generates the optimal procurement plan—prices, delivery times, and risks are all clearly visible. With just a click to confirm, the order is dispatched instantly without any human intervention.
In 2025, similar scenes quietly unfold across enterprises under China Aluminum Corporation (Chinalco Group): At Yunnan Aluminum Wenshan, electrolysis workers use AI prediction models for aluminum smelting process parameters, combined with slot control, measurement, and lab test data, to predict and output key control parameters, achieving real-time control over electrolytic cell production; In Liangshan Mining, intelligent mineral exploration models assist in deep exploration, achieving precise three-dimensional "treasure hunting"; In Beijing, Chinalco's smart monitoring system for related-party transactions monitors listed company deposits in real time, automatically warns, and completes fund transfers exceeding limits, realizing a full-process closed-loop management from monitoring to warning to allocation.
This is the most vivid illustration of Chinalco Group’s digital and intelligent transformation during the 14th Five-Year Plan period.
Over five years, facing the proposition of high-quality development, Chinalco Group has taken the path of "addressing weaknesses and strengthening advantages," positioning artificial intelligence as its core engine, driving a systemic reshaping from management processes to frontline production, from supply chains to data infrastructure.
Since the beginning of the 14th Five-Year Plan, Chinalco Group has focused on core businesses such as alumina, aluminum smelting, and copper smelting, developing a three-year advancement plan for critical intelligent production lines, promoting a shift from "experience-driven" to "data + algorithm-driven."
At Yunnan Aluminum Wenshan, leading smart factories have achieved self-adaptive adjustment of production processes, self-sensing warnings for equipment status, and autonomous optimization of energy consumption. A position that previously required 12 people working four shifts now only needs four remote operators. In 2024, the company was rated as an "Outstanding Smart Factory" by the Ministry of Industry and Information Technology (MIIT).
At Southwest Copper, an AI vision system operates around the clock, using high-precision 3D visual technology to scan cathode copper surfaces, precisely locking millimeter-scale defects within 1.5 seconds. At Chinalco Ruimin, the "AluZhiyun Pai" AI large model directly outputs near-optimal minute-level production plans, increasing on-time delivery rates from 94.1% to over 98%. At Liangshan Mining, the "Tongye ZhiLian" system achieves intelligent judgment of smelting endpoints, adding more than RMB 5 million in annual benefits and being selected as a high-value AI demonstration scenario among central SOEs.
To date, Chinalco Group has established 24 intelligent production lines and implemented tiered cultivation for smart factories. A scoring standard covering six basic requirements and 21 evaluation elements allows companies to "see the gap, find the direction."
In 2024, the "Kun'an" large AI model officially went online at Chinalco Group—the first large model in the non-ferrous metals industry. It can write reports, calculate risks, research materials, and optimize processes.
At Chinalco Research Institute, researchers input target performance requirements for copper alloys, and intelligent entities based on the "Kun'an" foundation reduce copper alloy composition and process screening time to hourly levels. The core forward prediction (composition → performance) takes about 6 to 10 seconds, while reverse design (performance → composition) takes about 60 seconds. "The laboratory R&D cycle for copper alloys has been shortened by approximately 60%," says Liu Tao, head of the Center for Materials Science and Engineering Computing and a senior researcher at the institute.
At Chinalco Financial Company, the "Caiyin Lingmou·AI Intelligent Transaction Matching System" instantly matches settlement vouchers with bank e-receipts upon launch, greatly liberating veteran finance personnel. The system, based on the "Kun'an" large model foundation and integrating natural language processing with neural network algorithms, boosts business processing efficiency by 93%, achieves a matching accuracy rate of 99.7%, and reduces receipt retrieval time from 15 minutes to 15 seconds. This system has been selected as a typical application case of AI empowering new industrialization by MIIT.
On the workshop floor, the "Xiao Kun" assistant helps new employees look up procedures and answer questions—just speak out your question, and the system responds instantly, accurately pushing operational guidelines and policy references. This 24/7 "AI colleague" also filters sensitive information according to job permissions to ensure compliance and security.
Behind "Kun'an" lies a solid digital foundation: Chinalco Group's cloud data center has replaced outdated server rooms; the group-wide AI computing center now provides 58 petaflops of computing power; the data platform goes live, integrating core financial and supply chain data, breaking down silos; and completing inventories of data resources across 23 business domains, making the organization's "data inventory" clearer.
More importantly, national recognition follows closely: Chinalco Material Institute's dataset was selected as outstanding achievements by SASAC and as a model case by the National Data Bureau; three projects including those from Chinalco Material and Chinalco Digital were included in the National Data Bureau's High-Quality Dataset Initiative and Data Infrastructure Pilot Program.
As Chinalco Group's sole strategic platform for digital and intelligent transformation, Chinalco Digital has achieved leapfrog growth: In 2024, it set record highs in revenue and profit, with contract values growing 187% year-on-year.
As the 14th Five-Year Plan comes to an end, Chinalco Group's digital and intelligent story is just beginning its most exciting chapter. The magic of this transformation lies in its grandeur yet detail-oriented nature, both "hardcore" and warm. It spares procurement officers late-night reconciliations, saves inspectors from running through sweltering heat, prevents researchers from repeating failed experiments, and ensures new hires never feel lost on day one. More importantly, it reshapes a culture: from "relying on experience" to "trusting data," from "waiting for instructions" to "proactively optimizing," from "fearing AI" to "embracing AI."
