Alibaba Group's research arm, Damo Academy, has developed Elements Claw, an artificial intelligence agent that discovered four previously unknown superconducting compounds verified through lab experiments, marking a significant advancement in materials science.
- Elements Claw scanned 2.4 million stable crystal structures in just 28 hours.
- Four new superconductors proposed by AI were experimentally confirmed.
- AI-driven materials discovery aims to revolutionize energy and quantum tech.
What happened
Alibaba Group Holding’s Damo Academy introduced an innovative AI agent called Elements Claw designed specifically for discovering superconducting materials. This system leverages a powerful foundation model trained on 125 million molecular and crystal structures, enabling it to scan and analyze large scientific datasets rapidly.
By processing 2.4 million stable crystal structures within 28 hours of GPU computing time, Elements Claw identified approximately 68,000 potential superconducting candidates. The AI then refined these options to highlight the most promising materials, leading to the successful experimental verification of four previously unknown superconductors.
Why it matters
Superconductors are crucial because they conduct electricity with zero resistance and expel magnetic fields at low temperatures, properties that underpin the next generation of power grids, quantum computers, and high-speed maglev trains. Traditional discovery methods depend on slow trial-and-error experiments with limited theoretical guidance, which constrain progress.
Alibaba’s AI-driven approach could drastically accelerate the pace of discovery by efficiently searching vast numbers of crystalline compounds and proposing viable candidates for laboratory testing. This aligns with a broader trend of applying AI beyond text and code generation, extending into scientific research to tackle complex problems and resource-intensive experimental workflows.
What to watch next
With Elements Claw's initial success, the next steps involve scaling and applying this AI methodology to uncover more superconducting materials and potentially other advanced materials with transformative properties. Continued collaboration with research institutions will be critical for validating AI-driven predictions and integrating findings into practical applications.
Globally, companies like Google DeepMind and Microsoft are also pushing AI in materials science and biology, indicating a competitive, innovation-driven landscape. Alibaba’s advancement may intensify efforts to harness AI for scientific breakthroughs, fundamentally changing how new technologies are developed and commercialized in China and worldwide.