AI for Materials，10大核心玩家！

TL;DR - An industry survey profiles 20 leading international and Chinese AI-for-materials companies developing foundation models, generative design tools, atomic simulations, and autonomous laboratories. Their work signals a shift from slow experimental trial-and-error toward computationally guided, closed-loop materials discovery.

• DeepMind’s GNoME and Microsoft’s MatterGen exemplify large-scale crystal prediction and property-conditioned generation, with selected candidates progressing to experimental synthesis.
• Platforms increasingly connect AI design, physics simulation, robotic experimentation, and feedback-driven model improvement into autonomous research loops.
• Commercial applications span batteries, semiconductors, catalysts, carbon capture, cooling systems, alloys, and advanced manufacturing.
• The competitive landscape includes model and cloud providers, materials-data platforms, simulation-software vendors, and vertically integrated companies pursuing discovery through industrialization.