日本政府计划向辉达（Nvidia）采购27,500颗新一代Rubin晶片，用于打造本土机器人基础AI模型。新成立的Noetra公司获得政府拨款3,873亿日圆（约24亿美元），将负责建设一座约140兆瓦的资料中心，预计于2028年6月上线。索尼、软银、丰田旗下Preferred Networks及NEC等数十家企业参与了Noetra的设立与运营。

Noetra的成立旨在整合日本国内分散的AI研发力量，降低对外国技术的依赖并强化国家安全。该公司由曾主导软银大型语言模型开发的丹波裕信担任社长，计划在明年三月前发布首个AI模型，并在数年内推出专为机器人应用设计的模型。日本政府的更大目标是到2040年占据全球预估60兆日圆机器人市场的30%以上份额。

随著全球开发能驱动精密机器人动作的AI模型竞争日趋白热化，日本面临一代产业工匠退休、技艺难以传承的挑战。辉达执行长黄仁勋表示，AI将使小型制造商能够教导机器人学习新流程，把师傅的技艺传承给下一代。经济产业大臣赤泽亮正也强调，建构强大的AI生态系统对资源匮乏且易受灾害影响的日本至关重要。

Japan plans to purchase 27,500 next-generation Rubin chips from Nvidia to develop a domestic foundational AI model for robotics. The newly established Noetra Corp., backed by ¥387.3 billion (approximately $2.4 billion) in government funding, will oversee the project and construct a roughly 140-megawatt data center expected to come online in June 2028. Dozens of major companies, including Sony, SoftBank, Preferred Networks, and NEC, are involved in setting up and operating Noetra.

Noetra was created to consolidate Japan's fragmented AI development efforts, reduce reliance on foreign technology, and strengthen national security. Led by President Hironobu Tamba, who previously spearheaded SoftBank's large language model, the venture aims to release an initial AI model by March of next year, with a robotics-specific model to follow within a few years. The Japanese government's broader ambition is to capture over 30% of the estimated ¥60 trillion global robotics market by 2040.

The global race to develop AI models capable of governing complex robotic movements is intensifying, while Japan faces the challenge of a retiring generation of industrial craftsmen whose skills risk being lost. Nvidia CEO Jensen Huang emphasized that AI can empower small manufacturers to teach robots new processes and preserve master craftsmanship for future generations. Economy Minister Ryosei Akazawa underscored that building a robust AI ecosystem is vital for resource-scarce and disaster-prone Japan.