What if a metal could become smarter? A research laboratory in southwest China's Chongqing is exploring how AI and cutting-edge technologies can reshape magnesium production and push the lightweight material beyond traditional processing.

AI takes the guesswork out of magnesium innovation

Magnesium is the lightest structural metal currently used in industry. China holds about 70% of the world's magnesium reserves and produces roughly 90% of global output. However, Chongqing's magnesium industry has historically focused on lower-value raw materials and preliminary processing.

Mingyue Lake Laboratory is working to increase the value of the resource by developing magnesium alloys, batteries, hydrogen storage systems and automotive components.

One of the lab's key innovations is an AI-powered platform for the design of magnesium materials. It draws on more than 500,000 scientific publications and patents, 3 million alloy-structure datasets and a lightweight-alloy large language model trained on 140 million scientific documents.

Researchers previously relied heavily on repeated physical experiments to identify suitable material compositions. The platform allows engineers to enter performance requirements and receive recommended alloy formulations, increasing research efficiency by more than 50%.

The facility has also developed an ampere-hour level magnesium-ion battery. The lab said its production cost could be about 4% of a lithium-ion battery's and that its chemistry reduces the risk of fire. After 19,000 charge-and-discharge cycles, the battery retained 88.3% of its capacity, according to lab data.

Another research direction is solid-state magnesium hydrogen storage. The lab said magnesium particles weighing about 100 grams can store hydrogen at a high density, with the technology intended to reduce reliance on large high-pressure tanks in hydrogen-powered vehicles.

The breakthrough could make lighter cars a reality

Automotive lightweighting is currently among the lab's closest links to industrial production.

Its researchers have developed the world's largest magnesium-alloy automotive die-casting component. The part is 32% lighter than a comparable aluminum-alloy component and has been installed in mass-produced vehicles made by Chongqing-based automaker SERES.

The lab is also developing magnesium-alloy wheels using semi-solid injection molding, a process in which partially solidified metal is injected into a mold to produce a part close to its final shape.

Jiang Bin, deputy director of the National Engineering Research Center for Magnesium Alloys and head of the lab's Institute of Lightweight Materials and Engineering, said the method could address the cost and efficiency limitations of forged magnesium wheels.

"Forging is a form of plastic processing with a long production process and high costs," Jiang said. "With semi-solid injection molding, the material cost can be about 25% lower than that of aluminum alloy, while the production process is simpler and allows near-net shaping."

Jiang said the research team had worked on both alloy composition and manufacturing parameters and could now produce wheels ranging from 16 to 21 inches.

A 16-inch magnesium wheel developed by the team weighs about 8 kilograms and is 30% lighter than an aluminum wheel of the same size. Four such wheels could reduce an electric vehicle's weight by at least 10 kilograms, according to project data.

Following engineering trials, the technology could enter mass production in new models from SERES, Changan Automobile and FAW as early as next year.

China's average use of magnesium alloy per vehicle is currently about 5 kilograms. Changan vehicles use approximately 10 kilograms, while SERES models use more than 20 kilograms, according to the lab.

"If semi-solid injection-molded magnesium wheels can be used in complete vehicles, the level of lightweighting will improve further," Jiang said. With additional applications in electric-drive housings, inner door panels and rear floor structures, he added, magnesium use in Chongqing-brand vehicles could eventually exceed 100 kilograms per vehicle and potentially reach 150 kilograms.

Mingyue Lake Laboratory has a research team of 719 people, led by academician Pan Fusheng of the Chinese Academy of Engineering. It operates eight major research platforms and has produced six original research outcomes that are particularly relevant to Chongqing.

The lab has received 59 invention patents and participated in the development or revision of six international standards and three national standards. It has also established six joint research centers with 10 companies, including China Baowu Steel Group and SERES.

A new-materials industrial fund of more than 1.2 billion yuan (about $177 million) has also been established with partners, including Minmetals Jintong, to support the commercialization of research results.

Global EV lightweighting fueling magnesium demand

For automakers around the world, reducing vehicle weight has become a key strategy to improve efficiency and extend electric vehicle range. As battery-powered cars become heavier due to larger battery packs, lightweight materials such as aluminum, carbon fiber and magnesium alloys are gaining attention across the industry.

The International Energy Agency identified vehicle lightweighting as an important pathway for improving EV efficiency in its Global EV Outlook 2024. The US Department of Energy has also listed magnesium alloys as a key material for future lightweight vehicles.

Magnesium remains far less widely used than aluminum, but adoption is increasing. Industry data showed that average magnesium use per vehicle in China reached about 10 kilograms in 2023, while the figure in North America was around 18 kilograms.

Demand for magnesium is increasing partly because of its use in vehicle and aircraft lightweighting, said the US Department of Energy. In Europe, increasingly stringent carbon dioxide standards for new cars and vans are also adding pressure on automakers to improve vehicle efficiency, with lightweight materials offering one potential route.

(Ran Zheng, Huxin Luo, Xudong Yang, Vivian Yan, Kenny Dong and Dan Liu contributed to this story.)