Micron develops edge inference memory for AI robots — and more robotics news today

Vídeo em Inglês, publicado pelo canal Robotics News Daily.

About this content: The summary, key points and topics below are generated automatically from the video transcript. They reflect the content, opinions, claims and predictions presented in the video and do not represent a position, validation or fact-check by IA Dots.

Video summary

The video discusses recent advancements and developments in robotics and artificial intelligence, covering topics such as Micron's new edge inference memory, AI learning to dream, humanoid robots entering manufacturing, China's policies on embodied AI, Rivian's strategic moves, Nvidia's endorsement of Open Claw, and new hardware platforms from companies like Hesai, ABB, Jodel, and Roboparty.

Key points

• Micron is developing edge inference memory for physical AI applications to reduce latency and boost bandwidth for real-time robotic decision-making.
• AI systems are implementing rest state processing techniques to review past performance, self-correct, and maintain accuracy.
• Humanoid robots are moving into manufacturing and warehouses, with companies like Dexforce using virtual simulations for training.
• China's Ministry of Commerce issued policy directives to accelerate embodied AI and established national standards for humanoid robots.
• Nvidia CEO Jensen Huang endorsed the open-source AI agent framework Open Claw to enhance security and robotics capabilities.
• Hesai launched Cosmo, a 3D spatial data platform utilizing LiDAR and 3D reconstruction to digitize environments for AI and robot training.

Topics

RoboticsArtificial IntelligenceHumanoid RobotsEdge MemorySpatial DataIndustrial Manufacturing

Video transcript

Transcript from available captions; it may contain errors.

Welcome back. Hey Race, what's been happening in robotics this afternoon? Micron's development of new memory, AI's ability to dream, humanoid robots entering manufacturing, China's policy boost for embodied AI, Rivian's strategic moves, Nvidia's endorsement of Open Claw, Heysea's spatial data platform, ABB's precision robots, Jodel's industrial execution tools, and Roboparties open-source humanoids. >> Sponsor this [music] excellent channel. pronetgrowth.com/robotics >> Micron is developing edge inference memory specifically for physical AI applications. This innovative technology aims to significantly reduce latency and boost bandwidth between AI processors and memory, which are critical for real-time decision-making in robots. Humanoid robots, with their increasing complexity and operational demands, are expected to drive a substantial new wave of memory demand in the coming years. Adding further context to this burgeoning market, Morgan Stanley forecasts that global humanoid robot deliveries could exceed 1 million units by the year 2030. This development is significant as it directly supports the hardware requirements for the [music] next generation of intelligent physical machines, potentially impacting companies like Figure and Tesla as they scale their humanoid production. AI is now learning to dream by reflecting on its mistakes through new rest state processing techniques. This allows sophisticated systems to review past [music] performance during designated downtime and self-correct, effectively moving AI from being merely reactive text generators to evolving >> [music] >> adaptive entities capable of continuous improvement. Developers are actively building systems that incorporate scheduled downtime for autonomous reflection cycles, where the AI [music] meticulously reviews its past inputs and outputs. This fundamental architectural change is designed [music] to break potentially toxic feedback loops and maintain a higher degree of accuracy even amidst complex digital noise. This advancement is crucial for ensuring the reliability and safety of AI systems operating in critical applications, potentially benefiting sectors like healthcare and autonomous driving. AI humanoid robots are steadily moving beyond the confines of research laboratories [music] and into the demanding environments of manufacturing facilities and large-scale warehouses. Consequently, human-robot collaboration is emerging as a profoundly important component of the future landscape of robotics and industrial operations. >> [music] >> Developers are actively exploring innovative ways for robots to assist human workers with tasks that are often repetitive, physically demanding, or ergonomically challenging. However, the practical, widespread deployment of these advanced robots still hinges on achieving higher levels of safety, reliability, and [music] cost-effectiveness for businesses. A notable example of progress in this area comes from Chinese startup Dexforce, which is utilizing virtual simulations to extensively train humanoid robots before transferring their learned capabilities [music] to physical machines. This simulated training approach can significantly reduce the substantial [music] time and expense typically associated with collecting real-world training data. The International Federation of Robotics reported approximately 7,000 humanoid robot sales for industrial and professional service applications in 2025, a number expected to grow as these technologies mature. Shifting gears, China's Ministry of Commerce, along with eight other government departments, has jointly issued a significant opinion document focused on accelerating technology-driven commerce. This directive explicitly calls for a substantial raising of the application level of embodied AI and other cutting-edge new technologies across the nation's [music] economy. This forward-thinking policy, coupled with the introduction of new national standards specifically for humanoid robots, has already had a notable impact on the market, leading to a surge in Unitree Robotics shares and reflecting growing investor confidence in the embodied AI sector. These concerted developments signal a clear strategic move towards the broader integration of humanoid robots into consumer and service scenarios, extending their utility far beyond traditional industrial [music] settings. is something that just came in. >> [music] >> Rivian Automotive has managed to land back on investor dashboards, sparking renewed interest. A key factor in [music] this resurgence is the appointment of veteran operator Doug Field as president of Mind Robotics, an entity in which Rivian holds a substantial stake of roughly 1/3. Complementing this leadership development, Rivian also reported record third-quarter vehicle deliveries, indicating strong operational performance. The company's commitment to vertical integration across autonomy, [music] battery technology, and software development is intended to unlock new, high-margin revenue streams in the long term, aiming to capitalize on these core competencies. >> [music] >> This is important for the automotive sector, as it highlights potential new directions for vehicle manufacturers beyond traditional passenger cars. Nvidia is backing Open Claw, a move that is putting significant robotics potential into sharp focus. Nvidia CEO Jensen Huang himself has publicly endorsed the open-source AI agent framework, describing it as a crucial foundation for future technological advancements, particularly within the realm of robotics. The framework is specifically designed to enhance agent security and to unlock new, advanced robotics capabilities that were previously difficult to achieve. This development has the potential to significantly accelerate the creation of more sophisticated, capable, and crucially, more secure robotic systems across a variety of applications. The potential applications span diverse industries, ranging from advanced manufacturing processes to intricate logistics and supply chain management. >> [music] >> Hesai has launched Cosmo, a groundbreaking 3D spatial data platform designed to revolutionize how environments are digitized. This advanced system is capable of converting real-world environments into highly detailed, accurate digital models, which are essential for training artificial intelligence systems and developing next-generation intelligent robots. Cosmo achieves this by seamlessly combining LiDAR technology, high-resolution imaging, and sophisticated 3D reconstruction techniques [music] to create these precise digital environments. The platform directly addresses the persistent challenge of obtaining accurate, realistic, and comprehensive representations of physical spaces for robot navigation and interaction. By extending its established LiDAR capabilities into the burgeoning physical AI market, Hesai is positioning itself to significantly accelerate the creation, testing, and eventual deployment of advanced robotic systems. ABB Robotics is introducing two new high-performance robots engineered for precision manufacturing applications. These compact IRB 1700 and IRB 1700 ID robots are designed to deliver up to 20% shorter cycle times while simultaneously offering enhanced motion accuracy and [music] increased payload capacities. They are specifically built to meet the growing demand for high-speed and precision manufacturing processes [music] across critical sectors like electronics and automotive manufacturing. Furthermore, the new robots boast a the total cost of ownership, achieved through reductions in power consumption and increased service intervals, meaning less downtime and maintenance. The robots are particularly well-suited for demanding tasks such as material handling, arc welding, and the increasingly [music] complex assembly of AI server racks. Jodel Robotics has unveiled a new lineup of products specifically designed for industrial embodied execution. This development is particularly significant as it addresses practical challenges faced by industrial robots and is expected to enable the more reliable deployment of embodied intelligence in precision manufacturing environments, enhancing task capabilities for companies like Jodel. [music] Roboparty has introduced open-source humanoid robots through its innovative RP1 platform. >> Inquire about sponsoring. pronetgrowth.com/robotics >> more before the day is out. Subscribe for more news on robotics.

Assistir ao vídeo original no YouTube

Assuntos: Agentes de IA, Robótica e humanoides, Assistente pessoal. Publicado no YouTube em 2026-10-19.

Seleção automática por assunto e idioma. O vídeo pertence ao canal de origem.