NVIDIA Jetson Brings Generative AI to Handheld Robotics

Beijing-based AI startup Moonshot has trained its Kimi K3 model, the world's largest open-source model with 2.8 trillion parameters, using Nvidia's advanced Blackwell chips despite U.S. export restrictions on such technology to Chinese firms. The company is now seeking additional Blackwell chips to develop Kimi K4, a significantly larger successor model. The situation highlights the tension between U.S. chip export controls and Chinese AI development capabilities.

ChangXin Memory Technologies (CXMT), a China-based RAM manufacturer, debuted on the Shanghai stock exchange with shares surging 466 percent on its first day, reaching a $484 billion valuation and becoming the most valuable Chinese company listed there. The company aims to compete with Samsung, Micron, and SK Hynix in the global memory market, capitalizing on demand from AI companies straining the industry and pushing device makers to seek cost alternatives.

Nvidia is investing $1 billion in South Korean internet company Naver to expand the country's AI data center infrastructure. The investment will finance Naver's plans to increase its AI data center capacity from 55 megawatts to 200 megawatts. The deal reflects Nvidia's strategy to build AI compute capacity outside the United States and positions South Korea as a regional hub for AI infrastructure.

Frontier physical AI models are moving beyond video training data to incorporate multiple camera angles, dense annotation, and brain wave readings as training inputs. The shift reflects growing recognition that traditional video datasets alone are insufficient for training AI systems that interact with the physical world. Brain wave data represents an emerging frontier in multimodal training approaches for robotics and embodied AI.