[DigitalToday reporter Jinju Hong (홍진주)] Alibaba Group Holding's research organisation DAMO Academy unveiled an artificial intelligence agent, Elements Clo, to discover superconducting materials and said it found four previously unknown candidate superconducting compounds and verified them in experiments.

According to the South China Morning Post on July 3 (local time), Elements Clo is a specialised system that applies generative AI to scientific research. It is designed to analyse vast numbers of papers and crystal-structure data to identify new materials with a high likelihood of superconductivity. DAMO Academy introduced it as the industry's first AI agent for searching for superconductors.

Superconductors are materials whose electrical resistance disappears and which expel magnetic fields in cryogenic environments. These properties are seen as key for next-generation technologies such as improving power-grid efficiency, quantum computing and high-speed magnetic levitation trains. But because a theory that accurately explains superconductivity is not yet complete, discovering new materials has long relied on trial-and-error experiments over extended periods.

DAMO Academy said it developed Elements Clo to reduce these limitations. The system is based on a specialised foundation model with 1 billion parameters trained on 125 million molecular and crystal structures. Renmin University of China and the University of Chinese Academy of Sciences also participated in the research.

DAMO Academy also stressed the speed of the AI search. Elements Clo screened about 2.4 million stable crystal structures during 28 hours of GPU computing and narrowed them to about 68,000 candidate materials with potential superconductivity. After additional analysis, it selected candidates with high experimental value, and four previously unreported compounds were ultimately confirmed in experiments, DAMO Academy explained.

The case shows that technology companies are rapidly expanding how they use AI. Companies are no longer limited to text generation or writing software code, and are extending AI to search scientific literature and analyse vast datasets to present hypotheses for researchers to test in experiments. Alibaba has also produced an example of applying AI to on-the-ground scientific research in line with this trend.

New materials development is seen as one of the areas where AI can have the greatest effect. Finding new materials requires reviewing countless compounds, and AI can reduce bottlenecks in the process to shorten research time. The widely used superconductor database SuperCon contains only about 2,000 registered materials.

Therefore, the key point of this announcement is how much AI can reduce bottlenecks in the candidate-search stage. Alibaba presented a structure in which Elements Clo handles paper searches, structure screening and candidate compression, while researchers focus on experimental verification. Attention is focused on whether AI can shorten search time in areas like superconductors, where theoretical prediction is difficult.