# Isaac

NVIDIA Isaac is an open robotics development platform comprising simulation and robot-learning frameworks, CUDA-accelerated libraries, AI models, and reference workflows for creating autonomous robots.

This is one page of public article previews, not the complete archive. Follow Next page to continue. Summaries are not the original full articles.

## Reconfiguring Procedural Knowledge for Compositional Robot Skill Adaptation

DevFeed: [Reconfiguring Procedural Knowledge for Compositional Robot Skill Adaptation](<https://devfeed.tech/articles/reconfiguring-procedural-knowledge-for-compositional-robot-skill-adaptation-139976.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.22174>)

Author: Daehee Lee, Dongsu Lee, Sanghyun Ahn, Minjong Yoo, Honguk Woo

Published: 2026-11-18T05:00:00Z

Content type: article

Language: en

Sources: [arXiv: Artificial Intelligence](<https://devfeed.tech/sources/arxiv-artificial-intelligence.md>)

Topics: [AI Chat](<https://devfeed.tech/topics/ai-chat.md>), [knowledge-engineering](<https://devfeed.tech/topics/knowledge-engineering.md>), [Isaac](<https://devfeed.tech/topics/isaac.md>)

Tags: [benchmark](<https://devfeed.tech/tags/benchmark.md>), [bind](<https://devfeed.tech/tags/bind.md>), [completion](<https://devfeed.tech/tags/completion.md>), [components](<https://devfeed.tech/tags/components.md>), [connectivity](<https://devfeed.tech/tags/connectivity.md>), [cs-ai](<https://devfeed.tech/tags/cs-ai.md>), [deployment](<https://devfeed.tech/tags/deployment.md>), [edits](<https://devfeed.tech/tags/edits.md>), [execution](<https://devfeed.tech/tags/execution.md>), [factor](<https://devfeed.tech/tags/factor.md>), [framework](<https://devfeed.tech/tags/framework.md>), [general](<https://devfeed.tech/tags/general.md>), [goals](<https://devfeed.tech/tags/goals.md>), [horizon](<https://devfeed.tech/tags/horizon.md>), [inference](<https://devfeed.tech/tags/inference.md>), [interfaces](<https://devfeed.tech/tags/interfaces.md>), [kitchen](<https://devfeed.tech/tags/kitchen.md>), [knowledge-base](<https://devfeed.tech/tags/knowledge-base.md>), [procedural-knowledge](<https://devfeed.tech/tags/procedural-knowledge.md>), [robot](<https://devfeed.tech/tags/robot.md>), [skill](<https://devfeed.tech/tags/skill.md>)

### AI overview

RecPo adapts robot skills by composing stored skill-induced transitions, rather than retraining for each new task. It represents skills through multiple procedure views that expose reusable factors and preserve local connectivity; at deployment, it interprets high-level task specifications through those views and resolves the needed skill bindings from stored relations. On long-horizon tasks in an extended Franka Kitchen benchmark, it reports 91.9% ordered completion and 1.0 ms task inference per episode across varying procedure support and task-record sparsity. Real-robot studies show adaptation through local memory edits without task-specific retraining.

### Source excerpt

arXiv:2611.22174v1 Announce Type: new Abstract: General-purpose behavior models increasingly internalize broad procedural knowledge during training, making their adaptation to new goals and skill interfaces less a problem of learning entirely new behaviors than of reconfiguring existing competence. We study this capability as reconfigurability, the ability to realize new behaviors by cost-efficiently restructuring existing components and relations under sparse governing information, where task specifications provide high-level goals and procedure specifications expose a limited set of reusable factors, leaving the concrete bindings and compositions required for execution underdetermined. We propose RecPo, a framework that represents skills task-agnostically through multiple procedure views and materializes executable skill sequences by composing stored transitions to realize task-specific changes. RecPo maintains a procedure knowledge base of skill-induced transitions, with each view exposing a reusable factor and preserving local skill connectivity. At deployment, it interprets task specifications through the same views and resolves the resulting compositional queries over the stored relations to bind skills for execution. On long-horizon tasks in an extended Franka Kitchen benchmark, RecPo achieves 91.9% ordered completion with 1.0 ms of task inference per episode across varying procedure support and task-record sparsity. Real-robot studies further demonstrate adaptation through local memory edits without task-specific retraining.

## AAEON BOXER-8741AI industrial Edge AI box PC NVIDIA Jetson Thor T5000 with dual 10GbE and four 5GbE LAN

DevFeed: [AAEON BOXER-8741AI industrial Edge AI box PC NVIDIA Jetson Thor T5000 with dual 10GbE and four 5GbE LAN](<https://devfeed.tech/articles/aaeon-boxer-8741ai-industrial-edge-ai-box-pc-nvidia-jetson-thor-t5000-with-dual-10gbe-and-four-5gbe-lan-135369.md>)

Original publisher: [Read original article](<https://www.cnx-software.com/2026/11/16/aaeon-boxer-8741ai-industrial-edge-ai-box-pc-nvidia-jetson-thor-t5000-with-dual-10gbe-and-four-5gbe-lan/>)

Author: Debashis Das

Published: 2026-11-16T04:00:27Z

Content type: news

Language: en

Sources: [CNX Software - Embedded Systems News](<https://devfeed.tech/sources/cnx-software-embedded-systems-news.md>)

Topics: [Thor](<https://devfeed.tech/topics/thor.md>), [Jetson](<https://devfeed.tech/topics/jetson.md>), [Isaac](<https://devfeed.tech/topics/isaac.md>), [Holoscan](<https://devfeed.tech/topics/holoscan.md>), [Metropolis](<https://devfeed.tech/topics/metropolis.md>), [Humanoid Robots](<https://devfeed.tech/topics/humanoid-robots.md>), [10 Gigabit Ethernet](<https://devfeed.tech/topics/10gbe-networking.md>)

Tags: [10gbe](<https://devfeed.tech/tags/10gbe.md>), [128gb-lpddr5x](<https://devfeed.tech/tags/128gb-lpddr5x.md>), [2-5gbe](<https://devfeed.tech/tags/2-5gbe.md>), [4g](<https://devfeed.tech/tags/4g.md>), [5g](<https://devfeed.tech/tags/5g.md>), [aaeon](<https://devfeed.tech/tags/aaeon.md>), [ai](<https://devfeed.tech/tags/ai.md>), [ai-performance](<https://devfeed.tech/tags/ai-performance.md>), [applications](<https://devfeed.tech/tags/applications.md>), [arm](<https://devfeed.tech/tags/arm.md>), [artificial-intelligence-ai](<https://devfeed.tech/tags/artificial-intelligence-ai.md>), [automotive](<https://devfeed.tech/tags/automotive.md>), [autonomous](<https://devfeed.tech/tags/autonomous.md>), [box](<https://devfeed.tech/tags/box.md>), [can-bus](<https://devfeed.tech/tags/can-bus.md>), [cuda](<https://devfeed.tech/tags/cuda.md>), [edge](<https://devfeed.tech/tags/edge.md>), [edge-ai](<https://devfeed.tech/tags/edge-ai.md>), [hardware](<https://devfeed.tech/tags/hardware.md>), [industrial](<https://devfeed.tech/tags/industrial.md>), [jetson](<https://devfeed.tech/tags/jetson.md>), [jetson-thor](<https://devfeed.tech/tags/jetson-thor.md>), [lan](<https://devfeed.tech/tags/lan.md>), [linux](<https://devfeed.tech/tags/linux.md>), [nvidia](<https://devfeed.tech/tags/nvidia.md>), [nvidia-jetson](<https://devfeed.tech/tags/nvidia-jetson.md>), [pc](<https://devfeed.tech/tags/pc.md>), [robotics](<https://devfeed.tech/tags/robotics.md>), [rs485](<https://devfeed.tech/tags/rs485.md>), [sensors](<https://devfeed.tech/tags/sensors.md>), [ubuntu](<https://devfeed.tech/tags/ubuntu.md>)

### AI overview

AAEON's BOXER-8741AI is a Jetson Thor T5000 industrial edge computer aimed at robotics and multi-sensor workloads. It is specified with 128GB LPDDR5X memory and up to 2070 sparse FP4 TFLOPS, plus two 10GbE and four 5GbE ports, two CAN FD interfaces, RS-232/RS-485, and optional GPIO. M.2 slots support NVMe storage, optional Wi-Fi/Bluetooth, and optional 4G/5G; power input is 19-48V DC, and operating temperature is -20°C to 60°C. The system supports JetPack 7.1 or later and NVIDIA tools including CUDA and TensorRT. The listed configuration includes a 128GB NVMe SSD; the 240W adapter and power cord are separate. The article lists a $10,219 price and expected availability around December 29, 2026, with preorders open at the time of writing.

### Source excerpt

AAEON BOXER-8741AI is an industrial edge AI box PC built around the NVIDIA Jetson Thor T5000 module, delivering up to 2070 FP4 TFLOPS (sparse) of AI performance and featuring 128GB LPDDR5X memory. Designed for humanoid robotics, autonomous mobile robots (AMRs), and multi-sensor perception workloads, it features two 10GbE and four 5GbE ports, two CAN FD interfaces, RS-232/RS-485, optional GPIO, and four USB 3.2 Gen 2 ports. Expansion options include M.2 slots for NVMe storage, Wi-Fi/Bluetooth, and 4G/5G cellular connectivity. The computer supports 19V to 48V DC input and operates from -20°C to 60°C for industrial applications. AAEON BOXER-8741AI specifications: SoM - NVIDIA Jetson Thor T5000 module CPU - 14-core 64-bit ARM Neoverse-V3AE @ up to 2.6GHz GPU - 2560-core NVIDIA Blackwell GPU with 96 fifth-generation Tensor Cores, Multi-Instance GPU (MIG) support with 10 Texture Processing Clusters (TPCs) VPU Video Decode - 4x 8Kp30 (H.265), 10x 4Kp60 (H.265), 22x 4Kp30 (H.265), [...] The post AAEON BOXER-8741AI industrial Edge AI box PC NVIDIA Jetson Thor T5000 with dual 10GbE and four 5GbE LAN appeared first on CNX Software - Embedded Systems News.

## AI Day Seoul Brings Together Korea's Industry Leaders, Startups and Developer Communities

DevFeed: [AI Day Seoul Brings Together Korea's Industry Leaders, Startups and Developer Communities](<https://devfeed.tech/articles/ai-day-seoul-brings-together-korea-s-industry-leaders-startups-and-developer-communities-134747.md>)

Original publisher: [Read original article](<https://blogs.nvidia.com/blog/ai-day-seoul-wrapup/>)

Author: NVIDIA Writers

Published: 2026-11-16T00:00:28Z

Content type: news

Language: en

Sources: [NVIDIA](<https://devfeed.tech/sources/nvidia-blog.md>)

Topics: [Developer Events](<https://devfeed.tech/topics/developer-events.md>), [Nvidia](<https://devfeed.tech/topics/nvidia.md>), [Isaac](<https://devfeed.tech/topics/isaac.md>), [AI Chat](<https://devfeed.tech/topics/ai-chat.md>), [Bootcamp](<https://devfeed.tech/topics/bootcamp.md>), [DGX Spark](<https://devfeed.tech/topics/dgx-spark.md>)

Tags: [agentic](<https://devfeed.tech/tags/agentic.md>), [agentic-ai](<https://devfeed.tech/tags/agentic-ai.md>), [agents](<https://devfeed.tech/tags/agents.md>), [ai](<https://devfeed.tech/tags/ai.md>), [ai-agents](<https://devfeed.tech/tags/ai-agents.md>), [ai-days](<https://devfeed.tech/tags/ai-days.md>), [ai-developers](<https://devfeed.tech/tags/ai-developers.md>), [ai-inference](<https://devfeed.tech/tags/ai-inference.md>), [ai-model](<https://devfeed.tech/tags/ai-model.md>), [communities](<https://devfeed.tech/tags/communities.md>), [deep-learning-institute](<https://devfeed.tech/tags/deep-learning-institute.md>), [developer](<https://devfeed.tech/tags/developer.md>), [developer-communities](<https://devfeed.tech/tags/developer-communities.md>), [edge](<https://devfeed.tech/tags/edge.md>), [edge-ai](<https://devfeed.tech/tags/edge-ai.md>), [events](<https://devfeed.tech/tags/events.md>), [hyundai](<https://devfeed.tech/tags/hyundai.md>), [inception](<https://devfeed.tech/tags/inception.md>), [industry](<https://devfeed.tech/tags/industry.md>), [korea](<https://devfeed.tech/tags/korea.md>), [nvidia](<https://devfeed.tech/tags/nvidia.md>), [physical-ai](<https://devfeed.tech/tags/physical-ai.md>), [robotics](<https://devfeed.tech/tags/robotics.md>), [startups](<https://devfeed.tech/tags/startups.md>)

### AI overview

AI Day Seoul brought more than 2,000 attendees to COEX for two days focused on physical AI and AI agents. Sessions featured Korean companies and startups using NVIDIA platforms for robotics, and NVIDIA's Jonathan Cohen presented on building agents that enterprises own and control. The event also included a community day, a Korean-language agent hackathon whose winning team received an NVIDIA DGX Spark, and workshops and Korean-language certification exams attended by more than 500 developers. Startup pitches and an AI for Science forum rounded out the program. The supplied text is an event summary and does not provide technical comparisons or detailed session content.

### Source excerpt

Over two days at COEX, more than 2,000 attendees converged for AI Day Seoul, bringing together Korea's industry leaders, startups and developer communities advancing physical AI and AI agents. At the plenary sessions, Deepu Talla, vice president of robotics and edge AI at NVIDIA, said Korea is "punching above its weight," with Doosan Robotics, Hyundai [...]

## Build a Robot Pick-and-Place Simulation with SimReady Assets

DevFeed: [Build a Robot Pick-and-Place Simulation with SimReady Assets](<https://devfeed.tech/articles/build-a-robot-pick-and-place-simulation-with-simready-assets-133556.md>)

Original publisher: [Read original article](<https://huggingface.co/blog/nvidia/simready-assets-robot-pick-and-place>)

Author: Dillon Bailey; Pomi Lee; Rishabh Chadha

Published: 2026-11-13T18:00:39Z

Content type: tutorial

Language: en

Sources: [Hugging Face](<https://devfeed.tech/sources/hugging-face-blog.md>)

Topics: [Isaac Sim](<https://devfeed.tech/topics/isaac-sim.md>), [Robotics Simulation](<https://devfeed.tech/topics/robotics-simulation.md>), [Robot Manipulation](<https://devfeed.tech/topics/robot-manipulation.md>), [Omniverse](<https://devfeed.tech/topics/omniverse.md>), [Isaac](<https://devfeed.tech/topics/isaac.md>), [Agent Skills](<https://devfeed.tech/topics/agent-skills.md>), [AI Chat](<https://devfeed.tech/topics/ai-chat.md>), [Tutorial](<https://devfeed.tech/topics/tutorial.md>)

Tags: [agent](<https://devfeed.tech/tags/agent.md>), [arm](<https://devfeed.tech/tags/arm.md>), [asset](<https://devfeed.tech/tags/asset.md>), [assets](<https://devfeed.tech/tags/assets.md>), [behavior](<https://devfeed.tech/tags/behavior.md>), [build](<https://devfeed.tech/tags/build.md>), [can](<https://devfeed.tech/tags/can.md>), [compose](<https://devfeed.tech/tags/compose.md>), [controls](<https://devfeed.tech/tags/controls.md>), [example](<https://devfeed.tech/tags/example.md>), [isaac-sim](<https://devfeed.tech/tags/isaac-sim.md>), [openusd](<https://devfeed.tech/tags/openusd.md>), [pick-and-place](<https://devfeed.tech/tags/pick-and-place.md>), [properties](<https://devfeed.tech/tags/properties.md>), [robot](<https://devfeed.tech/tags/robot.md>), [robotics](<https://devfeed.tech/tags/robotics.md>), [run](<https://devfeed.tech/tags/run.md>), [simulation](<https://devfeed.tech/tags/simulation.md>), [use-case](<https://devfeed.tech/tags/use-case.md>), [uses](<https://devfeed.tech/tags/uses.md>)

### AI overview

The walkthrough uses SimReady Central assets to build and test a scripted beverage-can pick-and-place task in Isaac Sim with a UR17G arm and Robotiq 2F-140 gripper. SimReady profiles define use-case features and requirements; validation checks authored settings such as joint drives, while the assembled simulation tests whether the robot can grasp, move, and release the can. The workflow downloads complete asset folders, assembles and validates the arm and gripper, then runs agent-generated controls, checking joint settings, motion, contact, and whether the can stays held. Isaac Sim 6.2 and an agent able to use its skills are required for the described agent-assisted setup; Hugging Face CLI is optional for command-line downloads. The example uses a fixed setup, scripted waypoints, and a 2 N-m drive-effort cap selected through simulation tests. It omits vision, limits collision checks, and does not model can deformation or liquid motion. Results do not establish robustness to uncertain positions, other payloads, or real hardware.

### Source excerpt

This walkthrough showcases a beverage can pick-and-place task that uses the UR17G arm, Robotiq 2F-140 gripper, and beverage can to compose a scene in Isaac Sim, check their behavior, and run a pick-and-place example with agent-generated controls. For robotics workflows, SimReady means an OpenUSD asset has been prepared with the properties needed for a defined simulation use case, such as collisions, mass, joints, and joint drives.

## Adapt Human Motion for Robots with NVIDIA SOMA Retargeter

DevFeed: [Adapt Human Motion for Robots with NVIDIA SOMA Retargeter](<https://devfeed.tech/articles/adapt-human-motion-for-robots-with-nvidia-soma-retargeter-132446.md>)

Original publisher: [Read original article](<https://developer.nvidia.com/blog/adapt-human-motion-for-robots-with-nvidia-soma-retargeter/>)

Author: John Malaska

Published: 2026-11-13T17:00:00Z

Content type: tutorial

Language: en

Sources: [NVIDIA Developer](<https://devfeed.tech/sources/nvidia-developer.md>), [NVIDIA Technical Blog](<https://devfeed.tech/sources/nvidia-technical-blog.md>)

Topics: [Humanoid Robots](<https://devfeed.tech/topics/humanoid-robots.md>), [Isaac](<https://devfeed.tech/topics/isaac.md>), [Machine learning](<https://devfeed.tech/topics/machine-learning.md>), [Isaac Sim](<https://devfeed.tech/topics/isaac-sim.md>), [3D](<https://devfeed.tech/topics/3d.md>)

Tags: [3d](<https://devfeed.tech/tags/3d.md>), [actions](<https://devfeed.tech/tags/actions.md>), [actuator](<https://devfeed.tech/tags/actuator.md>), [after](<https://devfeed.tech/tags/after.md>), [agents](<https://devfeed.tech/tags/agents.md>), [agibot](<https://devfeed.tech/tags/agibot.md>), [ai](<https://devfeed.tech/tags/ai.md>), [ai-agents](<https://devfeed.tech/tags/ai-agents.md>), [great](<https://devfeed.tech/tags/great.md>), [how-to](<https://devfeed.tech/tags/how-to.md>), [humanoid](<https://devfeed.tech/tags/humanoid.md>), [humanoid-robots](<https://devfeed.tech/tags/humanoid-robots.md>), [isaac-lab](<https://devfeed.tech/tags/isaac-lab.md>), [motion](<https://devfeed.tech/tags/motion.md>), [newton](<https://devfeed.tech/tags/newton.md>), [nvidia](<https://devfeed.tech/tags/nvidia.md>), [nvidia-research](<https://devfeed.tech/tags/nvidia-research.md>), [offers](<https://devfeed.tech/tags/offers.md>), [physical-ai](<https://devfeed.tech/tags/physical-ai.md>), [point](<https://devfeed.tech/tags/point.md>), [reinforcement-learning](<https://devfeed.tech/tags/reinforcement-learning.md>), [robotics](<https://devfeed.tech/tags/robotics.md>), [robotics-simulation](<https://devfeed.tech/tags/robotics-simulation.md>), [robots](<https://devfeed.tech/tags/robots.md>), [simulation-modeling-design](<https://devfeed.tech/tags/simulation-modeling-design.md>), [teaching](<https://devfeed.tech/tags/teaching.md>), [timing](<https://devfeed.tech/tags/timing.md>), [tutorial](<https://devfeed.tech/tags/tutorial.md>), [warp](<https://devfeed.tech/tags/warp.md>), [with](<https://devfeed.tech/tags/with.md>)

### AI overview

NVIDIA SOMA Retargeter adapts human motion to a robot's proportions and range of motion, then exports per-frame root poses and joint angles as CSV for downstream training or control. It accepts BVH animation on the SOMA skeleton; motion can come from capture, video reconstruction, or generation, but must be represented compatibly. The viewer supports configuring a robot, retargeting motion through weighted inverse kinematics, and applying post-processing such as foot-plant correction. Sample configurations and reference results cover five humanoid robots. Developers can refine weights on representative motions, then batch-retarget larger collections. For policy training, SONIC includes a SOMA CSV converter for G1, while ProtoMotions provides another physics-based tracking framework; the input representation and policy must match the target robot. The excerpt also describes simulation evaluation and a MicroDuck example, where trained controllers were compared with reference motions.

### Source excerpt

Human motion offers roboticists a great starting point for teaching humanoid robots how to move. A human demonstration captures the coordination and timing...

## Agile Robots raises $500M funding led by SoftBank at $8.5B valuation

DevFeed: [Agile Robots raises $500M funding led by SoftBank at $8.5B valuation](<https://devfeed.tech/articles/agile-robots-raises-500m-funding-led-by-softbank-at-8-5b-valuation-130575.md>)

Original publisher: [Read original article](<https://siliconangle.com/2026/11/12/agile-robots-raises-500m-funding-led-by-softbank-at-8-5b-valuation/>)

Author: Kyt Dotson

Published: 2026-11-12T16:10:57Z

Content type: news

Language: en

Sources: [SiliconANGLE](<https://devfeed.tech/sources/siliconangle.md>)

Topics: [Physical AI](<https://devfeed.tech/topics/physical-ai.md>), [Humanoid Robots](<https://devfeed.tech/topics/humanoid-robots.md>), [Robot Manipulation](<https://devfeed.tech/topics/robot-manipulation.md>), [Artificial Intelligence](<https://devfeed.tech/topics/ai.md>), [Isaac](<https://devfeed.tech/topics/isaac.md>)

Tags: [agile-robots](<https://devfeed.tech/tags/agile-robots.md>), [ai](<https://devfeed.tech/tags/ai.md>), [artificial-intelligence](<https://devfeed.tech/tags/artificial-intelligence.md>), [emerging-tech](<https://devfeed.tech/tags/emerging-tech.md>), [funding](<https://devfeed.tech/tags/funding.md>), [news](<https://devfeed.tech/tags/news.md>), [robotics](<https://devfeed.tech/tags/robotics.md>), [softbank](<https://devfeed.tech/tags/softbank.md>), [the-latest](<https://devfeed.tech/tags/the-latest.md>), [top-story-6](<https://devfeed.tech/tags/top-story-6.md>)

### AI overview

Agile Robots raised $500 million in a SoftBank-led Series F at an $8.5 billion valuation, bringing total capital invested to $2 billion. Nvidia, Hitachi and T Capital also participated. The company combines robotic hardware, sensing, computer vision and AI, with force-sensitive manipulation as a core capability. Its products include the Diana 7 arm and Agile One humanoid platforms; the Agile One SN is preconfigured for Nvidia Isaac GR00T models. Agile says the funding will support U.S. expansion, engineering hires, application development and increased production. Its Fürstenfeldbruck facility, opened in early 2026, is described as capable of producing up to 10,000 legged and wheeled humanoid robots annually. The excerpt also reports around 20,000 industrial robotic solutions deployed.

### Source excerpt

German artificial intelligence robotics startup Agile Robots SE announced today that it raised $500 million in new funding at an $8.5 billion valuation. Japanese multinational investment holding company SoftBank led the Series F funding round. Other participants included AI chip leader Nvidia, Japanese group Hitachi and T Capital. According to the company, the financing brings [...] The post Agile Robots raises $500M funding led by SoftBank at $8.5B valuation appeared first on SiliconANGLE.

## Mobile Robot Full-stack Development In Practice Ⅳ: How Does a Robot Understand What It Sees? YOLO, ByteTrack, SegFormer & FoundationPose

DevFeed: [Mobile Robot Full-stack Development In Practice Ⅳ: How Does a Robot Understand What It Sees? YOLO, ByteTrack, SegFormer & FoundationPose](<https://devfeed.tech/articles/mobile-robot-full-stack-development-in-practice-iv-how-does-a-robot-understand-what-it-sees-yolo-bytetrack-segformer-foundationpose-129674.md>)

Original publisher: [Read original article](<https://www.seeedstudio.com/blog/2026/11/11/mobile-robot-full-stack-development-in-practice-m4-jetson-ai-vision/>)

Author: Frank

Published: 2026-11-11T06:22:18Z

Content type: tutorial

Language: en

Sources: [Seeed Studio](<https://devfeed.tech/sources/latest-news-from-seeed-studio.md>)

Topics: [Isaac](<https://devfeed.tech/topics/isaac.md>), [Robot Navigation](<https://devfeed.tech/topics/robot-navigation.md>), [Robot Manipulation](<https://devfeed.tech/topics/robot-manipulation.md>), [Documentation](<https://devfeed.tech/topics/documentation.md>), [ultralytics](<https://devfeed.tech/topics/ultralytics.md>), [tegra](<https://devfeed.tech/topics/tegra.md>), [seeed studio](<https://devfeed.tech/topics/seeed-studio.md>)

Tags: [2d](<https://devfeed.tech/tags/2d.md>), [3d](<https://devfeed.tech/tags/3d.md>), [3d-reconstruction](<https://devfeed.tech/tags/3d-reconstruction.md>), [6d-pose-estimation](<https://devfeed.tech/tags/6d-pose-estimation.md>), [bytetrack](<https://devfeed.tech/tags/bytetrack.md>), [detection](<https://devfeed.tech/tags/detection.md>), [development](<https://devfeed.tech/tags/development.md>), [edge-ai](<https://devfeed.tech/tags/edge-ai.md>), [feature](<https://devfeed.tech/tags/feature.md>), [foundationpose](<https://devfeed.tech/tags/foundationpose.md>), [jetson-ai-vision](<https://devfeed.tech/tags/jetson-ai-vision.md>), [mobile](<https://devfeed.tech/tags/mobile.md>), [mobile-robot](<https://devfeed.tech/tags/mobile-robot.md>), [mobile-robotics](<https://devfeed.tech/tags/mobile-robotics.md>), [news](<https://devfeed.tech/tags/news.md>), [nvidia](<https://devfeed.tech/tags/nvidia.md>), [object](<https://devfeed.tech/tags/object.md>), [object-detection](<https://devfeed.tech/tags/object-detection.md>), [object-tracking](<https://devfeed.tech/tags/object-tracking.md>), [recomputer-robotics-j501](<https://devfeed.tech/tags/recomputer-robotics-j501.md>), [robot](<https://devfeed.tech/tags/robot.md>), [ros-2](<https://devfeed.tech/tags/ros-2.md>), [seeed-studio](<https://devfeed.tech/tags/seeed-studio.md>), [segformer](<https://devfeed.tech/tags/segformer.md>), [semantic](<https://devfeed.tech/tags/semantic.md>), [tensorrt](<https://devfeed.tech/tags/tensorrt.md>), [tracking](<https://devfeed.tech/tags/tracking.md>), [yolo](<https://devfeed.tech/tags/yolo.md>), [yolo11](<https://devfeed.tech/tags/yolo11.md>)

### AI overview

The guide shows how to inspect four camera perception outputs on a Jetson J501 with ROS 2: YOLO11 detections, ByteTrack IDs across frames, SegFormer pixel classes and a selected ground-candidate mask, and FoundationPose estimates of a known object's 6D pose. Each output answers a different question and needs separate validation. It gives TensorRT deployment and ROS message checks for YOLO; explains how detector confidence filtering can deprive ByteTrack's second association stage of low-score candidates; and shows how SegFormer's class selection changes the mask without changing model predictions. FoundationPose needs aligned RGB-D, intrinsics, a target-instance mask and mesh, with a camera-to-robot transform for base-frame use. The course compares native NVlabs and Isaac ROS routes as separate implementations; its recorded-data timing is not a full live-pipeline rate, and continuous physical-camera tracking and accuracy acceptance remain integration work. The guide recommends fixed inputs, one-setting-at-a-time comparisons, and labeled data or ground truth for quality evaluation.

### Source excerpt

From object detection and multi-object tracking to semantic segmentation and 6D pose estimation on NVIDIA The post Mobile Robot Full-stack Development In Practice Ⅳ: How Does a Robot Understand What It Sees? YOLO, ByteTrack, SegFormer & FoundationPose appeared first on Latest News from Seeed Studio.

## NVIDIA Physical AI in Korea: Samsung Cuts a Chip Thermal Solve From 18 Hours to 41 Seconds, Kakao Targets Level 4, LG Trains CLOiD

DevFeed: [NVIDIA Physical AI in Korea: Samsung Cuts a Chip Thermal Solve From 18 Hours to 41 Seconds, Kakao Targets Level 4, LG Trains CLOiD](<https://devfeed.tech/articles/nvidia-physical-ai-in-korea-samsung-cuts-a-chip-thermal-solve-from-18-hours-to-41-seconds-kakao-targets-level-4-lg-trains-cloid-124326.md>)

Original publisher: [Read original article](<https://www.storagereview.com/news/nvidia-physical-ai-korea-samsung-physicsnemo-kakao-level-4-lg-cloid>)

Author: Harold Fritts

Published: 2026-11-09T16:09:45Z

Content type: news

Language: en

Sources: [StorageReview.com](<https://devfeed.tech/sources/storagereview-com.md>)

Topics: [Physical AI](<https://devfeed.tech/topics/physical-ai.md>), [Robotics](<https://devfeed.tech/topics/robotics.md>), [Isaac](<https://devfeed.tech/topics/isaac.md>), [autonomous vehicles](<https://devfeed.tech/topics/autonomous-vehicles.md>), [Cosmos](<https://devfeed.tech/topics/cosmos.md>), [Simulation and Design](<https://devfeed.tech/topics/simulation-and-design.md>), [Jetson AGX Thor](<https://devfeed.tech/topics/jetson-agx-thor.md>), [NVIDIA DRIVE](<https://devfeed.tech/topics/nvidia-drive.md>)

Tags: [ai](<https://devfeed.tech/tags/ai.md>), [chip](<https://devfeed.tech/tags/chip.md>), [enterprise](<https://devfeed.tech/tags/enterprise.md>), [korea](<https://devfeed.tech/tags/korea.md>), [level-4](<https://devfeed.tech/tags/level-4.md>), [lg](<https://devfeed.tech/tags/lg.md>), [nvidia](<https://devfeed.tech/tags/nvidia.md>), [physical-ai](<https://devfeed.tech/tags/physical-ai.md>), [samsung](<https://devfeed.tech/tags/samsung.md>), [thermal](<https://devfeed.tech/tags/thermal.md>), [trains](<https://devfeed.tech/tags/trains.md>)

### AI overview

Samsung reports one chip-scale thermal and mechanical analysis taking 41 seconds on a GPU versus about 18 hours on a 32-core CPU, using AI-predicted effective material properties on a coarser grid. The roughly 1,600x comparison reflects the changed method and grid as well as GPU acceleration; the GPU model is unspecified. Samsung also uses NVIDIA cuISS and cuDSS for TCAD sparse linear systems and reports selected sMOL quantum chemistry calculations up to 5x faster end to end with cuEST. Kakao Mobility plans a Level 4 system combining its driving software and Seoul operating experience with NVIDIA Hyperion, DRIVE AGX Thor, and Halos; initial development is planned around Kia purpose-built vehicles, with no deployment timeline or vehicle count given. LG and NVIDIA demonstrated CLOiD robot tasks in a manufacturing environment, using a cycle of real-world data, Cosmos augmentation, GR00T training, Jetson AGX Thor computing, and simulation evaluation. The excerpt describes ongoing development, not production deployment.

### Source excerpt

NVIDIA used AI Day Seoul to lay out physical AI work with three Korean manufacturers: Samsung Electronics on chip simulation, Kakao Mobility on a Level 4 autonomous driving system, and LG Electronics on factory robots. The headline figure belongs to Samsung, which cut one chip-scale thermal and mechanical analysis from about 18 hours on a The post NVIDIA Physical AI in Korea: Samsung Cuts a Chip Thermal Solve From 18 Hours to 41 Seconds, Kakao Targets Level 4, LG Trains CLOiD appeared first on StorageReview.com.

## Practice Makes Precise: At CoRL, NVIDIA and Partners Reveal How Robots Rehearse Delicate Work in Simulation

DevFeed: [Practice Makes Precise: At CoRL, NVIDIA and Partners Reveal How Robots Rehearse Delicate Work in Simulation](<https://devfeed.tech/articles/practice-makes-precise-at-corl-nvidia-and-partners-reveal-how-robots-rehearse-delicate-work-in-simulation-123634.md>)

Original publisher: [Read original article](<https://blogs.nvidia.com/blog/robotics-simulation-advancements-corl/>)

Author: Sasa Docca

Published: 2026-11-09T14:00:01Z

Content type: news

Language: en

Sources: [NVIDIA](<https://devfeed.tech/sources/nvidia-blog.md>)

Topics: [Robot Manipulation](<https://devfeed.tech/topics/robot-manipulation.md>), [Robotics Simulation](<https://devfeed.tech/topics/robotics-simulation.md>), [Isaac](<https://devfeed.tech/topics/isaac.md>), [Isaac Sim](<https://devfeed.tech/topics/isaac-sim.md>), [Simulation and Design](<https://devfeed.tech/topics/simulation-and-design.md>), [Omniverse](<https://devfeed.tech/topics/omniverse.md>), [World models](<https://devfeed.tech/topics/world-models.md>), [Humanoid Robots](<https://devfeed.tech/topics/humanoid-robots.md>), [Artificial Intelligence](<https://devfeed.tech/topics/ai.md>)

Tags: [agentic-ai](<https://devfeed.tech/tags/agentic-ai.md>), [artificial-intelligence](<https://devfeed.tech/tags/artificial-intelligence.md>), [cables](<https://devfeed.tech/tags/cables.md>), [components](<https://devfeed.tech/tags/components.md>), [conference](<https://devfeed.tech/tags/conference.md>), [cosmos](<https://devfeed.tech/tags/cosmos.md>), [events](<https://devfeed.tech/tags/events.md>), [industrial-and-manufacturing](<https://devfeed.tech/tags/industrial-and-manufacturing.md>), [isaac](<https://devfeed.tech/tags/isaac.md>), [learning](<https://devfeed.tech/tags/learning.md>), [manipulation](<https://devfeed.tech/tags/manipulation.md>), [nvidia](<https://devfeed.tech/tags/nvidia.md>), [nvidia-research](<https://devfeed.tech/tags/nvidia-research.md>), [omniverse](<https://devfeed.tech/tags/omniverse.md>), [open-source](<https://devfeed.tech/tags/open-source.md>), [partners](<https://devfeed.tech/tags/partners.md>), [physical-ai](<https://devfeed.tech/tags/physical-ai.md>), [robotics](<https://devfeed.tech/tags/robotics.md>), [robots](<https://devfeed.tech/tags/robots.md>), [simulation](<https://devfeed.tech/tags/simulation.md>), [simulation-and-design](<https://devfeed.tech/tags/simulation-and-design.md>), [tasks](<https://devfeed.tech/tags/tasks.md>)

### AI overview

NVIDIA and research partners presented simulation tools, robot models, benchmarks and agent workflows at CoRL to train and evaluate precise manipulation. GR00T 2, trained on robot and human data, completed 40 of 50 tasks at least once without task-specific demonstrations and averaged 71% success with one demonstration per task, rising to 95% with five; agents can sequence skills and recover from errors. Isaac Lab 3.0 integrates the Newton physics engine to simulate deformable objects and articulated hands, and achieved up to 2.5x the training throughput of Isaac Lab 2.3 with PhysX on a vision-based dexterity benchmark. AlphaNIST used Success-Guided Sampling to train one policy across 20 millimeter-scale assembly tasks, reporting over 98% training success, versus 81% for separate policies, with roughly 39x less GPU time. Isaac Lab-Arena evaluates policies across varied conditions; new benchmarks DexBench and VABAR target industrial dexterity and adaptability. UC Berkeley's Graph-as-Policy uses agents to compose and refine skill sequences in simulation, producing interpretable programs for engineering review.

### Source excerpt

Simulation, which has long been the proving ground for robots learning to walk, roll and hop, is now helping researchers teach robots tasks that require a finer touch: aligning gears, connecting cables and assembling small components. The shift from locomotion to manipulation is taking center stage at this week's Conference on Robot Learning (CoRL) in [...]

## Korea Advances Physical AI Across Industries With NVIDIA

DevFeed: [Korea Advances Physical AI Across Industries With NVIDIA](<https://devfeed.tech/articles/korea-advances-physical-ai-across-industries-with-nvidia-123629.md>)

Original publisher: [Read original article](<https://blogs.nvidia.com/blog/korea-physical-ai-ecosystem/>)

Author: NVIDIA Writers

Published: 2026-11-09T05:00:40Z

Content type: news

Language: en

Sources: [NVIDIA](<https://devfeed.tech/sources/nvidia-blog.md>)

Topics: [Physical AI](<https://devfeed.tech/topics/physical-ai.md>), [Isaac](<https://devfeed.tech/topics/isaac.md>), [autonomous vehicles](<https://devfeed.tech/topics/autonomous-vehicles.md>), [Cosmos](<https://devfeed.tech/topics/cosmos.md>), [Jetson AGX Thor](<https://devfeed.tech/topics/jetson-agx-thor.md>), [Robotics](<https://devfeed.tech/topics/robotics.md>), [NVIDIA DRIVE](<https://devfeed.tech/topics/nvidia-drive.md>), [Chip design](<https://devfeed.tech/topics/chip-design.md>), [Robotics Simulation](<https://devfeed.tech/topics/robotics-simulation.md>), [Nvidia](<https://devfeed.tech/topics/nvidia.md>)

Tags: [accelerated-computing](<https://devfeed.tech/tags/accelerated-computing.md>), [ai](<https://devfeed.tech/tags/ai.md>), [autonomous](<https://devfeed.tech/tags/autonomous.md>), [autonomous-vehicles](<https://devfeed.tech/tags/autonomous-vehicles.md>), [cuda](<https://devfeed.tech/tags/cuda.md>), [development](<https://devfeed.tech/tags/development.md>), [driving](<https://devfeed.tech/tags/driving.md>), [ecosystem](<https://devfeed.tech/tags/ecosystem.md>), [engineering](<https://devfeed.tech/tags/engineering.md>), [industrial-and-manufacturing](<https://devfeed.tech/tags/industrial-and-manufacturing.md>), [industries](<https://devfeed.tech/tags/industries.md>), [industry](<https://devfeed.tech/tags/industry.md>), [isaac](<https://devfeed.tech/tags/isaac.md>), [korea](<https://devfeed.tech/tags/korea.md>), [manufacturing](<https://devfeed.tech/tags/manufacturing.md>), [nvidia](<https://devfeed.tech/tags/nvidia.md>), [nvidia-halos](<https://devfeed.tech/tags/nvidia-halos.md>), [nvidia-jetson](<https://devfeed.tech/tags/nvidia-jetson.md>), [nvidia-physicsnemo](<https://devfeed.tech/tags/nvidia-physicsnemo.md>), [physical-ai](<https://devfeed.tech/tags/physical-ai.md>), [production](<https://devfeed.tech/tags/production.md>), [real-world](<https://devfeed.tech/tags/real-world.md>), [robotics](<https://devfeed.tech/tags/robotics.md>), [samsung](<https://devfeed.tech/tags/samsung.md>), [semiconductor](<https://devfeed.tech/tags/semiconductor.md>), [semiconductor-engineering](<https://devfeed.tech/tags/semiconductor-engineering.md>), [semiconductor-manufacturing](<https://devfeed.tech/tags/semiconductor-manufacturing.md>)

### AI overview

NVIDIA-backed physical AI work in Korea spans semiconductor engineering, autonomous driving and factory robotics. Samsung reports selected quantum-chemistry calculations up to 5x faster with cuEST; in one AI homogenization case, coarser grids, AI-predicted material properties and GPU acceleration reduced a solve from about 18 hours on a 32-core CPU to 41 seconds. Kakao Mobility plans to combine its driving software and Gangnam operating experience with NVIDIA Hyperion, DRIVE AGX Thor and Halos as it develops a Level 4 system, initially for Kia purpose-built vehicles. LG Electronics and NVIDIA demonstrated CLOiD robot tasks in manufacturing using a development loop linking real-world data, Cosmos data augmentation, GR00T training, Jetson AGX Thor computing and simulation evaluation. The excerpt describes ongoing plans and demonstrations; it does not establish deployment of the planned mobility services.

### Source excerpt

Tapping into the country's strengths in manufacturing and engineering, NVIDIA is helping advance Korea's physical AI ecosystem. The country's industry leaders are applying accelerated computing, AI and simulation to everything from semiconductor manufacturing to robotics and autonomous vehicles. These efforts are helping bring physical AI from development into real-world production. Samsung Advances Semiconductor Engineering With [...]

## VLA Handover: Failure-Prediction-Based Switching from On-Robot to Server VLA Models

DevFeed: [VLA Handover: Failure-Prediction-Based Switching from On-Robot to Server VLA Models](<https://devfeed.tech/articles/vla-handover-failure-prediction-based-switching-from-on-robot-to-server-vla-models-123335.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.08987>)

Author: Takayuki Nishio

Published: 2026-11-09T05:00:00Z

Content type: article

Language: en

Sources: [arXiv: Machine Learning](<https://devfeed.tech/sources/arxiv-machine-learning.md>), [arXiv: Networking](<https://devfeed.tech/sources/arxiv-networking.md>)

Topics: [Isaac](<https://devfeed.tech/topics/isaac.md>), [smolvla](<https://devfeed.tech/topics/smolvla.md>)

Tags: [5g](<https://devfeed.tech/tags/5g.md>), [6g](<https://devfeed.tech/tags/6g.md>), [completion](<https://devfeed.tech/tags/completion.md>), [control](<https://devfeed.tech/tags/control.md>), [cost](<https://devfeed.tech/tags/cost.md>), [cs-lg](<https://devfeed.tech/tags/cs-lg.md>), [cs-ni](<https://devfeed.tech/tags/cs-ni.md>), [cs-ro](<https://devfeed.tech/tags/cs-ro.md>), [fail](<https://devfeed.tech/tags/fail.md>), [failure](<https://devfeed.tech/tags/failure.md>), [fewer](<https://devfeed.tech/tags/fewer.md>), [framework](<https://devfeed.tech/tags/framework.md>), [inference](<https://devfeed.tech/tags/inference.md>), [language](<https://devfeed.tech/tags/language.md>), [large](<https://devfeed.tech/tags/large.md>), [latency](<https://devfeed.tech/tags/latency.md>), [less](<https://devfeed.tech/tags/less.md>), [models](<https://devfeed.tech/tags/models.md>), [prediction](<https://devfeed.tech/tags/prediction.md>), [robot](<https://devfeed.tech/tags/robot.md>), [server](<https://devfeed.tech/tags/server.md>), [switching](<https://devfeed.tech/tags/switching.md>), [vla](<https://devfeed.tech/tags/vla.md>)

### AI overview

VLA handover uses an on-robot small VLA (S-VLA) for routine actions and calls a server-hosted large VLA (L-VLA) when the small model is likely to fail. The approach frames switching as minimizing server requests while meeting a required success probability, then uses a heuristic that retrains neither model. In simulations using SmolVLA and π₀.₅ on three LIBERO suites, it improved success by 11-22 percentage points over the S-VLA alone. Compared with server-only control, it sent 49-77% fewer requests with success 2-4 percentage points lower, and completion time increased less as network latency rose. These findings are limited to the reported simulation suites.

### Source excerpt

arXiv:2611.08987v1 Announce Type: cross Abstract: Large vision-language-action (VLA) models generalize well but are difficult to run in real time on a robot, and small VLAs that run there fail more often. Remote inference over beyond-5G and 6G networks lets a robot use a large VLA on a server, but querying the server for every action chunk limits the number of robots per server. We propose VLA handover: the robot runs a small VLA (S-VLA) and switches to a large VLA (L-VLA) on a server only when the S-VLA is likely to fail. We formulate the switching as minimizing the request cost under a required success probability and, since this problem cannot be solved directly, propose a heuristic framework that retrains neither VLA. In simulation with SmolVLA and $\pi_{0.5}$ on three of the four LIBERO suites, VLA handover raised the success rate by 11-22 percentage points over the S-VLA alone and sent 49-77% fewer requests than server-only control at a 2-4 percentage-point lower success rate, and its completion time grew less with the network latency.

## Which Robot Arm Should You Buy for VLA Research?

DevFeed: [Which Robot Arm Should You Buy for VLA Research?](<https://devfeed.tech/articles/which-robot-arm-should-you-buy-for-vla-research-124304.md>)

Original publisher: [Read original article](<https://www.seeedstudio.com/blog/2026/11/09/choose-a-robot-arm-to-buy-for-vla-research-seeed-rebot-arm-is-built-for-it-industrial-precision-with-nvidia-isaac-and-hf-lerobot/>)

Author: Jennie Wang

Published: 2026-11-09T03:07:44Z

Content type: opinion

Language: en

Sources: [Seeed Studio](<https://devfeed.tech/sources/latest-news-from-seeed-studio.md>)

Topics: [Robot Manipulation](<https://devfeed.tech/topics/robot-manipulation.md>), [Isaac](<https://devfeed.tech/topics/isaac.md>), [Isaac Sim](<https://devfeed.tech/topics/isaac-sim.md>), [Python](<https://devfeed.tech/topics/python.md>)

Tags: [6-dof](<https://devfeed.tech/tags/6-dof.md>), [actions](<https://devfeed.tech/tags/actions.md>), [adapter](<https://devfeed.tech/tags/adapter.md>), [ai](<https://devfeed.tech/tags/ai.md>), [ai-robotics](<https://devfeed.tech/tags/ai-robotics.md>), [arm](<https://devfeed.tech/tags/arm.md>), [camera](<https://devfeed.tech/tags/camera.md>), [can-bus](<https://devfeed.tech/tags/can-bus.md>), [data](<https://devfeed.tech/tags/data.md>), [data-augmentation](<https://devfeed.tech/tags/data-augmentation.md>), [data-collection](<https://devfeed.tech/tags/data-collection.md>), [dataset](<https://devfeed.tech/tags/dataset.md>), [deploy](<https://devfeed.tech/tags/deploy.md>), [deployment](<https://devfeed.tech/tags/deployment.md>), [digital-twin](<https://devfeed.tech/tags/digital-twin.md>), [drivers](<https://devfeed.tech/tags/drivers.md>), [physical-ai](<https://devfeed.tech/tags/physical-ai.md>), [rebot](<https://devfeed.tech/tags/rebot.md>), [research](<https://devfeed.tech/tags/research.md>), [robot](<https://devfeed.tech/tags/robot.md>), [robot-arm](<https://devfeed.tech/tags/robot-arm.md>), [tech](<https://devfeed.tech/tags/tech.md>), [vla](<https://devfeed.tech/tags/vla.md>)

### AI overview

The article presents Seeed's reBot Arm B601-RS as one desktop platform for collecting demonstrations, training VLA policies in simulation, and testing them on hardware. It describes a 6-DOF arm with 0.1 mm repeat positioning accuracy, 2.5 kg rated payload, and public hardware designs, BOM, Python SDK, ROS 2 drivers, LeRobot integration, and Isaac Sim assets. The stated workflow combines leader-follower teleoperation and synchronized RGB-D/joint data, Isaac Sim digital-twin training and augmentation, Isaac GR00T fine-tuning, and Jetson deployment. The article contrasts this with industrial arms it calls precise but closed and expensive, and hobby servo arms it describes as affordable but too imprecise. It also notes a lower-cost B601-DM variant with 0.2 mm repeatability and dual-arm configurations. These are the article's claims; the excerpt provides no independent comparison or performance results beyond the described benchmark task.

### Source excerpt

If you're asking "which robot arm should I buy for VLA research?", the honest answer The post Which Robot Arm Should You Buy for VLA Research? appeared first on Latest News from Seeed Studio.

## ToyBlocks: A controlled world for robot learning and evaluation

DevFeed: [ToyBlocks: A controlled world for robot learning and evaluation](<https://devfeed.tech/articles/toyblocks-a-controlled-world-for-robot-learning-and-evaluation-121632.md>)

Original publisher: [Read original article](<https://huggingface.co/blog/allenai/toyblocks>)

Author: Kyle Wiggers

Published: 2026-11-06T14:00:48Z

Content type: article

Language: en

Sources: [Hugging Face](<https://devfeed.tech/sources/hugging-face-blog.md>)

Topics: [Robot Manipulation](<https://devfeed.tech/topics/robot-manipulation.md>), [Dataset development](<https://devfeed.tech/topics/dataset-development.md>), [Artificial Intelligence](<https://devfeed.tech/topics/ai.md>), [Reinforcement learning](<https://devfeed.tech/topics/reinforcement-learning.md>), [World models](<https://devfeed.tech/topics/world-models.md>), [Open Source Models & Datasets](<https://devfeed.tech/topics/open-source-models-datasets.md>), [llm-reasoning](<https://devfeed.tech/topics/llm-reasoning.md>), [Frontier AI](<https://devfeed.tech/topics/frontier-ai.md>), [Predictive AI](<https://devfeed.tech/topics/predictive-ai.md>), [image recognition](<https://devfeed.tech/topics/image-recognition.md>), [Isaac](<https://devfeed.tech/topics/isaac.md>), [robot sense of touch](<https://devfeed.tech/topics/robot-sense-of-touch.md>)

Tags: [code](<https://devfeed.tech/tags/code.md>), [data](<https://devfeed.tech/tags/data.md>), [evaluation](<https://devfeed.tech/tags/evaluation.md>), [learning](<https://devfeed.tech/tags/learning.md>), [models](<https://devfeed.tech/tags/models.md>), [robot](<https://devfeed.tech/tags/robot.md>), [world](<https://devfeed.tech/tags/world.md>)

### AI overview

ToyBlocks is an open simulation framework for studying robot learning in a controlled tabletop world, combining automatic task and demonstration generation with training and evaluation tools. A robot arm learns to push, flip, pick, place, and stack blocks; researchers can vary training data and scene properties, then test generalization on held-out tasks. Simulation ground truth helps isolate planning, execution, and prediction failures. Initial results show popular action models generalize poorly across spatial changes; ManiFlow's atomic-skill success was similar in simulation and reality (80.0% and 81.1%), while the Molmo 2-ManiFlow system completed 55.5% of simulated and 50.0% of real block-spelling episodes, with further trials needed for statistical confidence. The release includes code, models, evaluation assets, and 85,336 long-horizon demonstrations. The excerpt also reports distinct world-model results: Wan 2.2 predicted block motion and outcomes most accurately, Astra tracked the gripper but struggled with contact, and LeWM was faster and more accurate on short interactions but drifted over full trajectories.

### Source excerpt

💻 Code | 📊 Data | 🤖 Models | 📖 Documentation Robots are learning to fold laundry, sort parts, and handle unfamiliar objects. But those demonstrations are often rather narrow and don't explain what a robot has learned or why it fails. To study a complex system, scientists often begin with a small, controlled environment, such as a petri dish or terrarium, where they can vary conditions and observe the effects.

## ToyBlocks: A controlled world for robot learning and evaluation

DevFeed: [ToyBlocks: A controlled world for robot learning and evaluation](<https://devfeed.tech/articles/toyblocks-a-controlled-world-for-robot-learning-and-evaluation-122104.md>)

Original publisher: [Read original article](<https://allenai.org/blog/toyblocks>)

Published: 2026-11-06T08:00:00Z

Content type: article

Language: en

Sources: [Ai2 Blog](<https://devfeed.tech/sources/ai2-blog.md>)

Topics: [Robot Manipulation](<https://devfeed.tech/topics/robot-manipulation.md>), [Physical AI](<https://devfeed.tech/topics/physical-ai.md>), [Dataset development](<https://devfeed.tech/topics/dataset-development.md>), [Reinforcement learning](<https://devfeed.tech/topics/reinforcement-learning.md>), [World models](<https://devfeed.tech/topics/world-models.md>), [Artificial Intelligence](<https://devfeed.tech/topics/ai.md>), [llm-reasoning](<https://devfeed.tech/topics/llm-reasoning.md>), [Isaac](<https://devfeed.tech/topics/isaac.md>), [image recognition](<https://devfeed.tech/topics/image-recognition.md>), [Fine-tuning](<https://devfeed.tech/topics/fine-tuning.md>), [robot sense of touch](<https://devfeed.tech/topics/robot-sense-of-touch.md>)

Tags: [automated](<https://devfeed.tech/tags/automated.md>), [evaluation](<https://devfeed.tech/tags/evaluation.md>), [framework](<https://devfeed.tech/tags/framework.md>), [generation](<https://devfeed.tech/tags/generation.md>), [learning](<https://devfeed.tech/tags/learning.md>), [open](<https://devfeed.tech/tags/open.md>), [robot](<https://devfeed.tech/tags/robot.md>), [simulation](<https://devfeed.tech/tags/simulation.md>), [tabletop](<https://devfeed.tech/tags/tabletop.md>), [training](<https://devfeed.tech/tags/training.md>), [world](<https://devfeed.tech/tags/world.md>)

### AI overview

ToyBlocks is a controlled tabletop simulation for studying how robot training data and methods affect performance. It automatically generates task demonstrations with reinforcement learning, motion planning, and a task-and-motion planner, and supports evaluation beyond success rates, including speed, force, and task progress. Researchers can vary demonstration mixtures and test held-out tasks, spatial generalization, and long-horizon planning; privileged simulator state helps generate skills, while perception models act from sensor observations. Reported tests include 63% success or correct infeasibility judgments in search-and-spell tasks, and zero-shot transfer of ManiFlow atomic skills with similar success in simulation and reality (80.0% and 81.1%). The hierarchical Molmo 2 + ManiFlow system succeeded in 55.5% of simulated and 50.0% of real episodes, with the authors noting more trials are needed for statistical confidence. The initial release includes 85,336 long-horizon demonstrations, training and evaluation code, simulation assets, and two trained models; integrated RL training is described as a future addition.

### Source excerpt

ToyBlocks is an open simulation framework for studying robot learning through automated task generation, training demonstrations, and evaluation in a controlled tabletop world.

## How NVIDIA Factory Operations Blueprint Coordinates AI Agents to Investigate Production Issues

DevFeed: [How NVIDIA Factory Operations Blueprint Coordinates AI Agents to Investigate Production Issues](<https://devfeed.tech/articles/how-nvidia-factory-operations-blueprint-coordinates-ai-agents-to-investigate-production-issues-117641.md>)

Original publisher: [Read original article](<https://developer.nvidia.com/blog/how-nvidia-factory-operations-blueprint-coordinates-ai-agents-to-investigate-production-issues/>)

Author: Samuel Ochoa

Published: 2026-11-04T17:00:00Z

Content type: article

Language: en

Sources: [NVIDIA Developer](<https://devfeed.tech/sources/nvidia-developer.md>), [NVIDIA Technical Blog](<https://devfeed.tech/sources/nvidia-technical-blog.md>)

Topics: [Multi Agent Systems](<https://devfeed.tech/topics/multi-agent-systems.md>), [agentic workflows](<https://devfeed.tech/topics/agentic-workflows.md>), [Agent Harness](<https://devfeed.tech/topics/agent-harness.md>), [human in the loop agent](<https://devfeed.tech/topics/human-in-the-loop-agent.md>), [Local AI](<https://devfeed.tech/topics/local-ai.md>), [NemoClaw](<https://devfeed.tech/topics/nemoclaw.md>), [OpenShell](<https://devfeed.tech/topics/openshell.md>), [Image processing](<https://devfeed.tech/topics/image-processing.md>), [Nemotron](<https://devfeed.tech/topics/nemotron.md>), [Isaac](<https://devfeed.tech/topics/isaac.md>)

Tags: [2](<https://devfeed.tech/tags/2.md>), [2026](<https://devfeed.tech/tags/2026.md>), [3](<https://devfeed.tech/tags/3.md>), [about](<https://devfeed.tech/tags/about.md>), [actions](<https://devfeed.tech/tags/actions.md>), [adapters](<https://devfeed.tech/tags/adapters.md>), [agent-toolkit](<https://devfeed.tech/tags/agent-toolkit.md>), [agentic-ai-generative-ai](<https://devfeed.tech/tags/agentic-ai-generative-ai.md>), [agents](<https://devfeed.tech/tags/agents.md>), [ai](<https://devfeed.tech/tags/ai.md>), [ai-agents](<https://devfeed.tech/tags/ai-agents.md>), [blueprint](<https://devfeed.tech/tags/blueprint.md>), [build-ai-agents](<https://devfeed.tech/tags/build-ai-agents.md>), [computer-vision-video-analytics](<https://devfeed.tech/tags/computer-vision-video-analytics.md>), [costs](<https://devfeed.tech/tags/costs.md>), [downtime](<https://devfeed.tech/tags/downtime.md>), [edge-computing](<https://devfeed.tech/tags/edge-computing.md>), [factory](<https://devfeed.tech/tags/factory.md>), [failed](<https://devfeed.tech/tags/failed.md>), [featured](<https://devfeed.tech/tags/featured.md>), [large](<https://devfeed.tech/tags/large.md>), [llms](<https://devfeed.tech/tags/llms.md>), [nemo](<https://devfeed.tech/tags/nemo.md>), [nemoclaw](<https://devfeed.tech/tags/nemoclaw.md>), [nvidia](<https://devfeed.tech/tags/nvidia.md>), [operations](<https://devfeed.tech/tags/operations.md>), [production](<https://devfeed.tech/tags/production.md>), [report](<https://devfeed.tech/tags/report.md>), [robotics](<https://devfeed.tech/tags/robotics.md>), [siemens](<https://devfeed.tech/tags/siemens.md>)

### AI overview

NVIDIA Factory Operations Blueprint (FOX) is a reference architecture for a factory manager agent that connects existing factory systems and delegates work to specialist agents. In the SMT example, an AOI alert prompts a second image inspection; confirmed defects can be sent for rework while a root cause agent uses the board's production history and machine readings to investigate likely causes and produce a sourced report for a line manager. Inconclusive inspections and initial handling require operator review. FOX uses an LLM, an agent harness (NemoClaw by default, with alternatives supported), and NVIDIA OpenShell to isolate execution and enforce access policies. Developers connect MES, ERP, machine databases and procedures; fully on-premises operation requires all models and service endpoints to stay within the factory network, while external models require assessment of data sent to them. Physical robot and AMR handling needs factory-specific integrations; the example demonstrates transport in Isaac Sim and ROS 2. Separately, Siemens describes rush-order simulations through Intelligence Center X and virtual AOI preparation using Process Preparation X, Cosmos and Omniverse, but

### Source excerpt

A Siemens report estimates that unplanned downtime costs an average large plant in the surveyed sectors $253 million per year. Whether investigating a failed...

## ANYbotics begins delivery of legged inspection robot built for explosive atmospheres

DevFeed: [ANYbotics begins delivery of legged inspection robot built for explosive atmospheres](<https://devfeed.tech/articles/anybotics-begins-delivery-of-legged-inspection-robot-built-for-explosive-atmospheres-102307.md>)

Original publisher: [Read original article](<https://siliconangle.com/2026/10/27/anybotics-begins-delivery-of-legged-inspection-robot-built-for-explosive-atmospheres/>)

Author: Kyt Dotson

Published: 2026-10-27T16:30:21Z

Content type: news

Language: en

Sources: [SiliconANGLE](<https://devfeed.tech/sources/siliconangle.md>)

Topics: [Isaac](<https://devfeed.tech/topics/isaac.md>), [Robot Navigation](<https://devfeed.tech/topics/robot-navigation.md>)

Tags: [ai-robotics](<https://devfeed.tech/tags/ai-robotics.md>), [anybotics](<https://devfeed.tech/tags/anybotics.md>), [anymal](<https://devfeed.tech/tags/anymal.md>), [delivery](<https://devfeed.tech/tags/delivery.md>), [emerging-tech](<https://devfeed.tech/tags/emerging-tech.md>), [heavy-industry](<https://devfeed.tech/tags/heavy-industry.md>), [inspection](<https://devfeed.tech/tags/inspection.md>), [legged-robots](<https://devfeed.tech/tags/legged-robots.md>), [news](<https://devfeed.tech/tags/news.md>), [peter-fankhauser](<https://devfeed.tech/tags/peter-fankhauser.md>), [robot](<https://devfeed.tech/tags/robot.md>), [robotics](<https://devfeed.tech/tags/robotics.md>), [the-latest](<https://devfeed.tech/tags/the-latest.md>), [thierry-obede](<https://devfeed.tech/tags/thierry-obede.md>), [top-story-6](<https://devfeed.tech/tags/top-story-6.md>)

### AI overview

ANYbotics is shipping ANYmal X, a legged inspection robot engineered for oil, gas and chemical facilities where flammable atmospheres make ignition a risk. The company says it assessed ignition risks and added safeguards across the machine, including its electrical components and hot surfaces, while preserving the cameras, sensors and docking ports needed for operation. Energy and chemical companies tested ANYmal robots across hundreds of onshore and offshore facilities from 2023, and their feedback helped refine the platform. The text names Equinor and DSM-Firmenich among current recipients; it does not provide independent certification details or performance results for ANYmal X.

### Source excerpt

ANYbotics AG, a leading provider of four-legged inspection robots for heavy industry, today announced it's shipping the world's first legged robot designed to operate in dangerous environments where a spark or fire could trigger an explosion. The new robot, named ANYmal X, is a dog-like robot that can patrol and inspect equipment where specialist technicians [...] The post ANYbotics begins delivery of legged inspection robot built for explosive atmospheres appeared first on SiliconANGLE.

## Hottest toy for Christmas? Nvidia CEO plays with adorable Microduck robots, 20,000 sell

DevFeed: [Hottest toy for Christmas? Nvidia CEO plays with adorable Microduck robots, 20,000 sell](<https://devfeed.tech/articles/hottest-toy-for-christmas-nvidia-ceo-plays-with-adorable-microduck-robots-20-000-sell-98531.md>)

Original publisher: [Read original article](<https://techcrunch.com/2026/10/23/hottest-toy-for-christmas-nvidia-ceo-plays-with-adorable-microduck-robots-20000-sell/>)

Author: Julie Bort

Published: 2026-10-23T21:59:19Z

Content type: news

Language: en

Sources: [TechCrunch | AI News & Artificial Intelligence](<https://devfeed.tech/sources/techcrunch-ai-news-artificial-intelligence.md>)

Topics: [Isaac](<https://devfeed.tech/topics/isaac.md>), [Isaac Sim](<https://devfeed.tech/topics/isaac-sim.md>)

Tags: [ai](<https://devfeed.tech/tags/ai.md>), [ceo](<https://devfeed.tech/tags/ceo.md>), [christmas](<https://devfeed.tech/tags/christmas.md>), [gadgets](<https://devfeed.tech/tags/gadgets.md>), [microduck](<https://devfeed.tech/tags/microduck.md>), [nvidia](<https://devfeed.tech/tags/nvidia.md>), [pollen-robotics](<https://devfeed.tech/tags/pollen-robotics.md>), [robots](<https://devfeed.tech/tags/robots.md>), [tc](<https://devfeed.tech/tags/tc.md>), [toy](<https://devfeed.tech/tags/toy.md>), [toys](<https://devfeed.tech/tags/toys.md>)

### AI overview

After a viral video of Jensen Huang playing with Microducks, preorders passed 20,000 units, with delivery expected before Christmas. The open-source robots from Pollen Robotics ship with basic tricks learned in simulation using Nvidia's technology stack. Owners can train them to perform new tricks and share those with the community. The excerpt says the robots appeal to adults as well as children, but provides no pricing or technical specifications.

### Source excerpt

The robots, stomping their little feet and making tiny R2-D2 noises, sat, fell down dead for a "nap," and summersaulted on voice command.

## Using AI Agents to Teach Robots Like Microduck to Walk in Simulation

DevFeed: [Using AI Agents to Teach Robots Like Microduck to Walk in Simulation](<https://devfeed.tech/articles/using-ai-agents-to-teach-robots-like-microduck-to-walk-in-simulation-97810.md>)

Original publisher: [Read original article](<https://developer.nvidia.com/blog/using-ai-agents-to-teach-robots-like-microduck-to-walk-in-simulation/>)

Author: Gabriele Leone

Published: 2026-10-23T14:58:32Z

Content type: tutorial

Language: en

Sources: [NVIDIA Developer](<https://devfeed.tech/sources/nvidia-developer.md>), [NVIDIA Technical Blog](<https://devfeed.tech/sources/nvidia-technical-blog.md>)

Topics: [Reinforcement learning](<https://devfeed.tech/topics/reinforcement-learning.md>), [Isaac](<https://devfeed.tech/topics/isaac.md>), [Isaac Sim](<https://devfeed.tech/topics/isaac-sim.md>), [Robotics Simulation](<https://devfeed.tech/topics/robotics-simulation.md>), [Robot Navigation](<https://devfeed.tech/topics/robot-navigation.md>), [Machine Perception](<https://devfeed.tech/topics/machine-perception.md>), [Physical AI](<https://devfeed.tech/topics/physical-ai.md>), [Simulation](<https://devfeed.tech/topics/simulation.md>), [Artificial Intelligence](<https://devfeed.tech/topics/ai.md>), [3D](<https://devfeed.tech/topics/3d.md>), [3D visualization](<https://devfeed.tech/topics/3d-visualization.md>)

Tags: [actuator](<https://devfeed.tech/tags/actuator.md>), [agentic-ai-generative-ai](<https://devfeed.tech/tags/agentic-ai-generative-ai.md>), [agents](<https://devfeed.tech/tags/agents.md>), [ai](<https://devfeed.tech/tags/ai.md>), [ai-agents](<https://devfeed.tech/tags/ai-agents.md>), [ai-assisted](<https://devfeed.tech/tags/ai-assisted.md>), [algorithm](<https://devfeed.tech/tags/algorithm.md>), [apis](<https://devfeed.tech/tags/apis.md>), [asset](<https://devfeed.tech/tags/asset.md>), [assets](<https://devfeed.tech/tags/assets.md>), [astra](<https://devfeed.tech/tags/astra.md>), [base](<https://devfeed.tech/tags/base.md>), [blender](<https://devfeed.tech/tags/blender.md>), [cad](<https://devfeed.tech/tags/cad.md>), [camera](<https://devfeed.tech/tags/camera.md>), [central](<https://devfeed.tech/tags/central.md>), [cloud](<https://devfeed.tech/tags/cloud.md>), [gtc-berlin-2026](<https://devfeed.tech/tags/gtc-berlin-2026.md>), [microduck](<https://devfeed.tech/tags/microduck.md>), [policy](<https://devfeed.tech/tags/policy.md>), [robot-navigation](<https://devfeed.tech/tags/robot-navigation.md>), [robotics](<https://devfeed.tech/tags/robotics.md>), [robots](<https://devfeed.tech/tags/robots.md>), [simulation](<https://devfeed.tech/tags/simulation.md>), [simulation-modeling-design](<https://devfeed.tech/tags/simulation-modeling-design.md>)

### AI overview

The workflow uses AI agents to plan and prepare a Microduck obstacle course, then combines editable assets made with FreeCAD and Blender with Isaac Sim and Newton physics. Isaac Lab trains a locomotion policy using parallel reinforcement-learning simulations; the described ramp and obstacle run starts from a pretrained Microduck policy, uses head-camera depth and robot-state observations, and runs 4,096 environments on an RTX PRO 6000 Blackwell Workstation Edition GPU. The selected policy is evaluated locally in Isaac Sim on RTX Spark, with course completion, falls, collisions, and stalled progress recorded to guide iteration. Physical deployment is outside the post's detailed scope. It describes preparing for sim-to-real transfer through actuator calibration, including delays, backlash, and voltage drops, plus realistic domain randomization and subsequent hardware testing.

### Source excerpt

Teaching a robot to navigate an obstacle course requires more than a walking policy. It must perceive its surroundings, maintain balance, and adjust its actions...

## Mobile Robot Full-Stack Development in Practice Ⅲ: How Does a Robot Know It Is Moving? LiDAR, IMU and Sensor Fusion

DevFeed: [Mobile Robot Full-Stack Development in Practice Ⅲ: How Does a Robot Know It Is Moving? LiDAR, IMU and Sensor Fusion](<https://devfeed.tech/articles/mobile-robot-full-stack-development-in-practice-iii-how-does-a-robot-know-it-is-moving-lidar-imu-and-sensor-fusion-97449.md>)

Original publisher: [Read original article](<https://www.seeedstudio.com/blog/2026/10/23/mobile-robot-full-stack-development-m3-robot-sensor-fusion-with-ros-2/>)

Author: Frank

Published: 2026-10-23T09:17:21Z

Content type: article

Language: en

Sources: [Seeed Studio](<https://devfeed.tech/sources/latest-news-from-seeed-studio.md>)

Topics: [Robot Navigation](<https://devfeed.tech/topics/robot-navigation.md>), [Isaac](<https://devfeed.tech/topics/isaac.md>), [Error Propagation](<https://devfeed.tech/topics/error-propagation.md>)

Tags: [alignment](<https://devfeed.tech/tags/alignment.md>), [development](<https://devfeed.tech/tags/development.md>), [feature](<https://devfeed.tech/tags/feature.md>), [feedback](<https://devfeed.tech/tags/feedback.md>), [first](<https://devfeed.tech/tags/first.md>), [full](<https://devfeed.tech/tags/full.md>), [full-stack](<https://devfeed.tech/tags/full-stack.md>), [full-stack-development](<https://devfeed.tech/tags/full-stack-development.md>), [fusion](<https://devfeed.tech/tags/fusion.md>), [imu](<https://devfeed.tech/tags/imu.md>), [jetson-course](<https://devfeed.tech/tags/jetson-course.md>), [know](<https://devfeed.tech/tags/know.md>), [lidar](<https://devfeed.tech/tags/lidar.md>), [mobile](<https://devfeed.tech/tags/mobile.md>), [mobile-robot](<https://devfeed.tech/tags/mobile-robot.md>), [moving](<https://devfeed.tech/tags/moving.md>), [news](<https://devfeed.tech/tags/news.md>), [nvidia-jetson](<https://devfeed.tech/tags/nvidia-jetson.md>), [point](<https://devfeed.tech/tags/point.md>), [recomputer-robotics](<https://devfeed.tech/tags/recomputer-robotics.md>), [robot](<https://devfeed.tech/tags/robot.md>), [ros-2](<https://devfeed.tech/tags/ros-2.md>), [seeed-studio](<https://devfeed.tech/tags/seeed-studio.md>), [sensor](<https://devfeed.tech/tags/sensor.md>), [sensor-fusion](<https://devfeed.tech/tags/sensor-fusion.md>), [stack](<https://devfeed.tech/tags/stack.md>)

### AI overview

Reliable robot motion estimates depend on making sensor measurements comparable in time, coordinate frame and uncertainty before fusing them. The article's minimum ROS 2 `robot_localization` experiment combines wheel forward velocity with IMU yaw rate in a two-dimensional EKF; raw LiDAR points are not direct inputs to that filter. It outlines validating LiDAR message types and timing, preprocessing clouds while preserving geometry and per-point timing, decoding chassis CAN feedback into odometry, and checking IMU units, bias, axes and covariance. Sensor timestamps need a common time basis, and measured sensor-to-body transforms and a single `odom -> base_link` publisher help avoid misalignment and TF conflicts. Covariances should reflect observed behavior; smoother output alone does not prove accuracy. GNSS is a separate global-positioning extension requiring valid fixes, heading, local odometry and antenna extrinsics. Stationary, straight, turning and rectangular-route tests provide repeatable comparisons of raw and fused estimates.

### Source excerpt

From point clouds and chassis feedback to time alignment and EKF state estimation. In the The post Mobile Robot Full-Stack Development in Practice Ⅲ: How Does a Robot Know It Is Moving? LiDAR, IMU and Sensor Fusion appeared first on Latest News from Seeed Studio.

## FluxVLA Engine: A One-Stop VLA Engineering Platform for Embodied Intelligence

DevFeed: [FluxVLA Engine: A One-Stop VLA Engineering Platform for Embodied Intelligence](<https://devfeed.tech/articles/fluxvla-engine-a-one-stop-vla-engineering-platform-for-embodied-intelligence-97973.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2609.17210>)

Author: Yinhao Li, Weixin Mao, Zihan Lan, Jikun Rong, Qirui Hu, Yiming Zhang, Weipeng Deng, Bowen Shen, Minzhao Zhu, Yiming Mao, Yan Yang, Chenguang Cui, Hongyuan Chen, Xu Huang, Zheyi Zhao, Pinxi Shen, Bozhe

Published: 2026-10-23T04:00:00Z

Content type: article

Language: en

Sources: [arXiv: Artificial Intelligence](<https://devfeed.tech/sources/arxiv-artificial-intelligence.md>)

Topics: [Isaac](<https://devfeed.tech/topics/isaac.md>), [AI Engineering](<https://devfeed.tech/topics/ai-engineering.md>), [Artificial Intelligence](<https://devfeed.tech/topics/ai.md>), [Model Development](<https://devfeed.tech/topics/model-development.md>), [World models](<https://devfeed.tech/topics/world-models.md>), [Humanoid Robots](<https://devfeed.tech/topics/humanoid-robots.md>)

Tags: [automatic](<https://devfeed.tech/tags/automatic.md>), [cs-ai](<https://devfeed.tech/tags/cs-ai.md>), [cs-ro](<https://devfeed.tech/tags/cs-ro.md>), [deployment](<https://devfeed.tech/tags/deployment.md>), [embodied-learning](<https://devfeed.tech/tags/embodied-learning.md>), [engineering](<https://devfeed.tech/tags/engineering.md>), [evaluation](<https://devfeed.tech/tags/evaluation.md>), [gpu](<https://devfeed.tech/tags/gpu.md>), [human-in-the-loop](<https://devfeed.tech/tags/human-in-the-loop.md>), [inference](<https://devfeed.tech/tags/inference.md>), [offline-reinforcement-learning](<https://devfeed.tech/tags/offline-reinforcement-learning.md>), [platform](<https://devfeed.tech/tags/platform.md>), [policies](<https://devfeed.tech/tags/policies.md>), [real-time](<https://devfeed.tech/tags/real-time.md>), [simulation](<https://devfeed.tech/tags/simulation.md>), [training](<https://devfeed.tech/tags/training.md>), [vla](<https://devfeed.tech/tags/vla.md>)

### AI overview

FluxVLA Engine is an open, configuration-driven platform for connecting embodied-policy research to robot deployment; it presents an engineering workflow rather than a new policy model. It standardizes interfaces for datasets, visual-language and world models, action heads, reward- or advantage-weighted learning, distributed training, simulation evaluation, inference, and robot operators. The abstract also describes compositional dual-arm simulation, automatic data generation, and human-in-the-loop rollouts, takeovers, correction collection, and reward annotation. For physical execution, it combines Real-Time Chunking with accelerated inference, remote GPU serving, and configurable trajectory post-processing. These shared, auditable interfaces are intended to link offline learning, simulation validation, online correction, and real-robot execution. The supplied evidence is an abstract and does not report comparative performance results.

### Source excerpt

arXiv:2609.17210v2 Announce Type: replace-cross Abstract: Vision-language-action (VLA) models, world-action models (WAMs), and offline reinforcement learning methods are rapidly expanding the design space of embodied policies, yet turning these algorithms into reliable robot systems remains constrained by fragmented data formats, training stacks, evaluation protocols, inference runtimes, and embodiment-specific interfaces. We present $\mathrm{FluxVLA}$ Engine, an open, configuration-driven platform that turns heterogeneous embodied-policy components into a reproducible data-to-deployment workflow. Rather than introducing another policy model, $\mathrm{FluxVLA}$ standardizes interfaces for datasets, visual-language and world models, action heads, reward- or advantage-weighted learning, distributed training, simulation evaluation, optimized inference, and robot operators. The engine further integrates compositional dual-arm simulation, scalable automatic data generation, and model-decoupled human-in-the-loop rollout, takeover, correction collection, and reward annotation. For responsive physical execution, it combines Real-Time Chunking (RTC) with accelerated inference backends, lightweight remote GPU serving, and configurable trajectory post-processing. Together, these capabilities connect offline learning, simulation validation, online correction, and real-robot execution through shared and auditable contracts. $\mathrm{FluxVLA}$ therefore targets the engineering bottlenecks separating promising embodied-learning algorithms from reproducible evaluation and dependable deployment. Code is available at https://github.com/FluxVLA/FluxVLA

## RoboParty RP1 (ROBOTO 01) - A full-stack open-source humanoid robot with up to 30 degrees of freedom

DevFeed: [RoboParty RP1 (ROBOTO 01) - A full-stack open-source humanoid robot with up to 30 degrees of freedom](<https://devfeed.tech/articles/roboparty-rp1-roboto-01-a-full-stack-open-source-humanoid-robot-with-up-to-30-degrees-of-freedom-96498.md>)

Original publisher: [Read original article](<https://www.cnx-software.com/2026/10/23/roboparty-rp1-roboto-01-a-full-stack-open-source-humanoid-robot-with-up-to-30-degrees-of-freedom/>)

Author: Debashis Das

Published: 2026-10-23T00:00:41Z

Content type: news

Language: en

Sources: [CNX Software - Embedded Systems News](<https://devfeed.tech/sources/cnx-software-embedded-systems-news.md>)

Topics: [Humanoid Robots](<https://devfeed.tech/topics/humanoid-robots.md>), [Isaac](<https://devfeed.tech/topics/isaac.md>), [Open Source](<https://devfeed.tech/topics/open-source.md>)

Tags: [ai](<https://devfeed.tech/tags/ai.md>), [ai-development](<https://devfeed.tech/tags/ai-development.md>), [allwinner](<https://devfeed.tech/tags/allwinner.md>), [allwinner-a-series](<https://devfeed.tech/tags/allwinner-a-series.md>), [artificial-intelligence-ai](<https://devfeed.tech/tags/artificial-intelligence-ai.md>), [battery](<https://devfeed.tech/tags/battery.md>), [computer-vision](<https://devfeed.tech/tags/computer-vision.md>), [computing](<https://devfeed.tech/tags/computing.md>), [control](<https://devfeed.tech/tags/control.md>), [development](<https://devfeed.tech/tags/development.md>), [edge-ai](<https://devfeed.tech/tags/edge-ai.md>), [education](<https://devfeed.tech/tags/education.md>), [embedded](<https://devfeed.tech/tags/embedded.md>), [embodied-ai](<https://devfeed.tech/tags/embodied-ai.md>), [experiments](<https://devfeed.tech/tags/experiments.md>), [hardware](<https://devfeed.tech/tags/hardware.md>), [humanoid](<https://devfeed.tech/tags/humanoid.md>), [jetson](<https://devfeed.tech/tags/jetson.md>), [jetson-orin](<https://devfeed.tech/tags/jetson-orin.md>), [jetson-orin-nx](<https://devfeed.tech/tags/jetson-orin-nx.md>), [linux](<https://devfeed.tech/tags/linux.md>), [mcu](<https://devfeed.tech/tags/mcu.md>), [nvidia-jetson](<https://devfeed.tech/tags/nvidia-jetson.md>), [open](<https://devfeed.tech/tags/open.md>), [open-source](<https://devfeed.tech/tags/open-source.md>), [roboparty](<https://devfeed.tech/tags/roboparty.md>), [robot](<https://devfeed.tech/tags/robot.md>), [robotics](<https://devfeed.tech/tags/robotics.md>), [rockchip-rk35xx](<https://devfeed.tech/tags/rockchip-rk35xx.md>), [rp1](<https://devfeed.tech/tags/rp1.md>), [steam](<https://devfeed.tech/tags/steam.md>), [video](<https://devfeed.tech/tags/video.md>)

### AI overview

RoboParty's RP1 is a research-oriented open-source humanoid offered as a 24-DoF model and a 30-DoF RP1 Pro. Both use an RK3588 processor plus a separate MCU for real-time motion control; the Pro adds an Allwinner A733, Jetson Orin NX, interaction screen, camera and more capable arms and hands. The robot uses ROMOMO actuators with local control loops, and a dual-battery system supports hot-swapping while maintaining critical control power. RoboParty publishes mechanical CAD, PCB files, firmware components, ROS 2 deployment software, simulation models and Isaac Lab training workflows. The 24-axis configuration has a training-to-deployment workflow; the Pro body configuration lacks dedicated training tasks and default policies. The article notes that actuator firmware source was not found, some software remains in development, and price and shipping timing are unannounced.

### Source excerpt

Shanghai-based RoboParty has introduced the RP1 (ROBOTO 01), an open-source humanoid robot designed for robotics research, embodied AI development, and motion-control experiments. The robot is available in two configurations: the standard RP1 with 24 active degrees of freedom (DoF) and the RP1 Pro with 30 DoF. The robot stands 1.35 meters (~4.43 feet) tall and is built around a Rockchip RK3588 processor, with the Pro model adding an Allwinner A733 processor and an NVIDIA Jetson Orin NX development computing module. Other features include in-house ROMOMO joint actuators with up to 160 N-m peak torque, interchangeable head modules, optional dexterous hands, and a 10Ah battery rated for up to three hours of operation. A separate 480MHz MCU handles real-time motion control, while a dual-voltage battery system supports hot-swapping without shutting down the robot. RoboParty RP1 specifications: Computing RP1 - Rockchip RK3588 processor with a 6 TOPS NPU, 8GB LPDDR4x RAM, and [...] The post RoboParty RP1 (ROBOTO 01) - A full-stack open-source humanoid robot with up to 30 degrees of freedom appeared first on CNX Software - Embedded Systems News.

## Nvidia's Berlin GTC: Faster AI factories, lower power use and more secure agents

DevFeed: [Nvidia's Berlin GTC: Faster AI factories, lower power use and more secure agents](<https://devfeed.tech/articles/nvidia-s-berlin-gtc-faster-ai-factories-lower-power-use-and-more-secure-agents-96168.md>)

Original publisher: [Read original article](<https://www.networkworld.com/article/4239248/nvidias-berlin-gtc-faster-ai-factories-lower-power-use-and-more-secure-agents.html>)

Author: Andy Patrizio

Published: 2026-10-22T17:51:50Z

Content type: news

Language: en

Sources: [Network World](<https://devfeed.tech/sources/network-world.md>)

Topics: [NVIDIA DSX](<https://devfeed.tech/topics/nvidia-dsx.md>), [OpenShell](<https://devfeed.tech/topics/openshell.md>), [Isaac](<https://devfeed.tech/topics/isaac.md>), [clean energy data centers](<https://devfeed.tech/topics/clean-energy-data-centers.md>)

Tags: [accelerate](<https://devfeed.tech/tags/accelerate.md>), [agents](<https://devfeed.tech/tags/agents.md>), [ai](<https://devfeed.tech/tags/ai.md>), [ai-agents](<https://devfeed.tech/tags/ai-agents.md>), [ai-compute](<https://devfeed.tech/tags/ai-compute.md>), [ai-factories](<https://devfeed.tech/tags/ai-factories.md>), [ai-factory](<https://devfeed.tech/tags/ai-factory.md>), [ai-infrastructure](<https://devfeed.tech/tags/ai-infrastructure.md>), [architecture](<https://devfeed.tech/tags/architecture.md>), [artificial-intelligence](<https://devfeed.tech/tags/artificial-intelligence.md>), [artificial-intelligence-data-center-robotics](<https://devfeed.tech/tags/artificial-intelligence-data-center-robotics.md>), [autonomous](<https://devfeed.tech/tags/autonomous.md>), [autonomous-ai-agents](<https://devfeed.tech/tags/autonomous-ai-agents.md>), [berlin](<https://devfeed.tech/tags/berlin.md>), [bluefield-4](<https://devfeed.tech/tags/bluefield-4.md>), [compute](<https://devfeed.tech/tags/compute.md>), [conference](<https://devfeed.tech/tags/conference.md>), [cooling](<https://devfeed.tech/tags/cooling.md>), [data-center](<https://devfeed.tech/tags/data-center.md>), [edition-us](<https://devfeed.tech/tags/edition-us.md>), [faster](<https://devfeed.tech/tags/faster.md>), [gtc](<https://devfeed.tech/tags/gtc.md>), [language-en](<https://devfeed.tech/tags/language-en.md>), [nvidia](<https://devfeed.tech/tags/nvidia.md>), [power](<https://devfeed.tech/tags/power.md>), [robotics](<https://devfeed.tech/tags/robotics.md>), [storytype-news](<https://devfeed.tech/tags/storytype-news.md>)

### AI overview

At its Berlin GTC, Nvidia announced a modular data center system intended to shorten AI infrastructure deployment from months to weeks. Its DSX reference architecture combines liquid cooling, dry coolers, and power-management software; Nvidia claims DSX MaxLPS can produce the same number of AI tokens with up to 30% less power, and estimates dry cooling at 2 gigawatts could save 54 million liters of water daily versus cooling towers. OpenShell combines Vera Rubin compute, Vera CPUs, BlueField-4 DPUs, and security software to build, monitor, and constrain agents under security policies; the reference architecture is available now. Nvidia also expanded Isaac robotics tools, while Agile Robots' Jetson Thor-powered humanoid platform is planned for early 2027. In telecom, T-Mobile completed what the companies called the first AI-RAN call on a commercial network, and BT, Deutsche Telekom, and Elisa announced pilots. Efficiency figures and deployment benefits are vendor claims.

### Source excerpt

Nvidia is hosting its GTC conference in Berlin this week, and while the bulk of the news is European-oriented, there are several stories of interest around the world. They include the usual stories of Nvidia: AI compute platforms, data center infrastructure, and networking. Nvidia Unveils DSX POD Modular AI Factories Nvidia introduced the DSX POD, a factory-built, modular data center approach designed to accelerate AI infrastructure deployment. The concept packages compute and supporting infrastructure into integrated systems rather than requiring every deployment to be built from scratch. Nvidia is really positioning this as a turnkey solution, stating that it could reduce deployment timelines from months to weeks. "An AI factory creates value when it starts producing tokens -- not when construction is complete," said Jensen Huang, founder and CEO of Nvidia in a statement. "With Nvidia DSX POD, we're bringing factory-built precision and speed to AI infrastructure construction -- reducing deployment time from months to weeks so customers can scale capacity and bring revenue-generating production online faster." Nvidia Targets Lower Power Consumption and Water Use Nvidia's DSX reference architecture combines closed-loop liquid cooling, dry coolers and power-management software. The company says its DSX MaxLPS software can generate the same number of AI tokens using up to 30% less power. Nvidia also estimates that a 2-gigawatt deployment using its dry-cooling approach could save 54 million liters of cooling water per day compared with conventional cooling-tower systems. OpenShell Brings Security Controls to Autonomous AI Agents Nvidia announced its OpenShell Reference Architecture, combining Vera Rubin compute, Vera CPUs, BlueField-4 data processing units (DPUs) and security software to build, monitor and constrain AI agents. The OpenShell Reference Architecture builds on the foundation of Nvidia OpenShell secure runtime software, extending it into full-stack, industria

## Nvidia debuts new AI reference architectures, development tools

DevFeed: [Nvidia debuts new AI reference architectures, development tools](<https://devfeed.tech/articles/nvidia-debuts-new-ai-reference-architectures-development-tools-93940.md>)

Original publisher: [Read original article](<https://siliconangle.com/2026/10/21/nvidia-debuts-new-ai-reference-architectures-development-tools/>)

Author: Maria Deutscher

Published: 2026-10-21T22:36:53Z

Content type: news

Language: en

Sources: [SiliconANGLE](<https://devfeed.tech/sources/siliconangle.md>)

Topics: [NVIDIA DSX](<https://devfeed.tech/topics/nvidia-dsx.md>), [Isaac](<https://devfeed.tech/topics/isaac.md>), [Nvidia](<https://devfeed.tech/topics/nvidia.md>)

Tags: [accelerators](<https://devfeed.tech/tags/accelerators.md>), [agent](<https://devfeed.tech/tags/agent.md>), [agents](<https://devfeed.tech/tags/agents.md>), [ai](<https://devfeed.tech/tags/ai.md>), [ai-agent](<https://devfeed.tech/tags/ai-agent.md>), [architectures](<https://devfeed.tech/tags/architectures.md>), [berlin](<https://devfeed.tech/tags/berlin.md>), [build](<https://devfeed.tech/tags/build.md>), [chips](<https://devfeed.tech/tags/chips.md>), [customers](<https://devfeed.tech/tags/customers.md>), [data-center](<https://devfeed.tech/tags/data-center.md>), [data-centers](<https://devfeed.tech/tags/data-centers.md>), [development-tools](<https://devfeed.tech/tags/development-tools.md>), [dsx-pod](<https://devfeed.tech/tags/dsx-pod.md>), [gtc](<https://devfeed.tech/tags/gtc.md>), [infra](<https://devfeed.tech/tags/infra.md>), [isaac](<https://devfeed.tech/tags/isaac.md>), [jensen-huang](<https://devfeed.tech/tags/jensen-huang.md>), [llm](<https://devfeed.tech/tags/llm.md>), [nasdaq-nvda](<https://devfeed.tech/tags/nasdaq-nvda.md>), [news](<https://devfeed.tech/tags/news.md>), [nvidia](<https://devfeed.tech/tags/nvidia.md>), [open-source](<https://devfeed.tech/tags/open-source.md>), [openshell](<https://devfeed.tech/tags/openshell.md>), [reference](<https://devfeed.tech/tags/reference.md>), [reference-architecture](<https://devfeed.tech/tags/reference-architecture.md>), [robotics](<https://devfeed.tech/tags/robotics.md>), [robots](<https://devfeed.tech/tags/robots.md>), [the-latest](<https://devfeed.tech/tags/the-latest.md>), [tools](<https://devfeed.tech/tags/tools.md>)

### AI overview

Nvidia announced reference architectures aimed at speeding AI infrastructure deployment and containing agent workloads. DSX POD packages factory-integrated and tested Rubin GPUs, Vera CPUs, networking, cooling and power equipment in containers; Nvidia estimates deployment could shrink from six months to 14 days. A second architecture combines Vera, Rubin and BlueField-4 to run OpenShell, whose agents operate in isolated sandboxes with a supervisor monitoring outbound traffic. It also supports edge environments on DGX Station, DGX Spark and RTX Spark laptops. For robotics, Isaac's updated simulator adds APIs for configuring virtual robot sensors, alongside three workflows for customizing the GR00T foundation model. Nvidia also introduced Agile ONE SN, a humanoid robot powered by GR00T that it presents as a starting point for customer projects. The excerpt says DSX POD systems are expected to ship early next year; these timelines and performance estimates are Nvidia's claims.

### Source excerpt

Nvidia Corp. today debuted a set of technologies that will make it easier for customers to build data centers and robots using its chips. Chief Executive Officer Jensen Huang detailed the offerings at the company's annual GTC Berlin event. The first addition to Nvidia's data center portfolio, DSX POD, is a reference architecture. That's a [...] The post Nvidia debuts new AI reference architectures, development tools appeared first on SiliconANGLE.

## NVIDIA GTC Berlin: 14-Day DSX POD AI Factories, a 2.5x Substrate Buildout, and the OpenShell Reference Architecture

DevFeed: [NVIDIA GTC Berlin: 14-Day DSX POD AI Factories, a 2.5x Substrate Buildout, and the OpenShell Reference Architecture](<https://devfeed.tech/articles/nvidia-gtc-berlin-14-day-dsx-pod-ai-factories-a-2-5x-substrate-buildout-and-the-openshell-reference-architecture-93073.md>)

Original publisher: [Read original article](<https://www.storagereview.com/news/nvidia-gtc-berlin-2026-dsx-pod-14-day-ai-factory-substrate-openshell-ai-ran>)

Author: Harold Fritts

Published: 2026-10-21T18:04:42Z

Content type: news

Language: en

Sources: [StorageReview.com](<https://devfeed.tech/sources/storagereview-com.md>)

Topics: [Nvidia](<https://devfeed.tech/topics/nvidia.md>), [NVIDIA DSX](<https://devfeed.tech/topics/nvidia-dsx.md>), [AI Infrastructure](<https://devfeed.tech/topics/ai-infrastructure.md>), [OpenShell](<https://devfeed.tech/topics/openshell.md>), [NVIDIA Vera Rubin](<https://devfeed.tech/topics/nvidia-vera-rubin.md>), [Isaac](<https://devfeed.tech/topics/isaac.md>), [AI Architecture](<https://devfeed.tech/topics/ai-architecture.md>), [BioNeMo](<https://devfeed.tech/topics/bionemo.md>), [NVIDIA DRIVE](<https://devfeed.tech/topics/nvidia-drive.md>), [Thor](<https://devfeed.tech/topics/thor.md>), [NVIDIA Blackwell](<https://devfeed.tech/topics/nvidia-blackwell.md>), [DOCA](<https://devfeed.tech/topics/doca.md>)

Tags: [agent](<https://devfeed.tech/tags/agent.md>), [agent-security](<https://devfeed.tech/tags/agent-security.md>), [ai](<https://devfeed.tech/tags/ai.md>), [ai-factories](<https://devfeed.tech/tags/ai-factories.md>), [ai-factory](<https://devfeed.tech/tags/ai-factory.md>), [announcements](<https://devfeed.tech/tags/announcements.md>), [architecture](<https://devfeed.tech/tags/architecture.md>), [autonomous](<https://devfeed.tech/tags/autonomous.md>), [berlin](<https://devfeed.tech/tags/berlin.md>), [containerized](<https://devfeed.tech/tags/containerized.md>), [dsx](<https://devfeed.tech/tags/dsx.md>), [dsx-pod](<https://devfeed.tech/tags/dsx-pod.md>), [enterprise](<https://devfeed.tech/tags/enterprise.md>), [gtc](<https://devfeed.tech/tags/gtc.md>), [nvidia](<https://devfeed.tech/tags/nvidia.md>), [nvidia-gtc](<https://devfeed.tech/tags/nvidia-gtc.md>), [nvidia-gtc-berlin](<https://devfeed.tech/tags/nvidia-gtc-berlin.md>), [openshell](<https://devfeed.tech/tags/openshell.md>), [reference](<https://devfeed.tech/tags/reference.md>), [reference-architecture](<https://devfeed.tech/tags/reference-architecture.md>), [substrate](<https://devfeed.tech/tags/substrate.md>)

### AI overview

NVIDIA says its DSX POD reference design can reduce deployment at a prepared site from six months to as little as 14 days by completing rack integration and validation in factories; partner systems are targeted for early 2027, and the estimate excludes facility construction. The announcements also set planned capacity expansions, including a 2.5x increase in substrate capacity over three years. DSX software targets power allocation, grid coordination and infrastructure validation; NVIDIA reports efficiency and throughput gains in partner evaluations, which the source says were not independently audited. OpenShell pairs agent workloads and harnesses with BlueField DPUs running Sentry for out-of-band monitoring and hardware policy enforcement; NVIDIA claims millisecond quarantine and performance gains, but gives no comparison configurations or workload conditions. The excerpt also describes AI-RAN, vehicle safety and autonomy, robotics, protein design and healthcare programs, with many deployments, trials and capacity goals still planned or under development.

### Source excerpt

NVIDIA's GTC Berlin announcements today span containerized Vera Rubin infrastructure, a 2.5x expansion of its substrate supply, and software for AI factory power management and agent security, along with telecom, autonomous vehicle, robotics, and healthcare programs. The headline number is a deployment target: NVIDIA says its new DSX POD reference design cuts an AI factory's The post NVIDIA GTC Berlin: 14-Day DSX POD AI Factories, a 2.5x Substrate Buildout, and the OpenShell Reference Architecture appeared first on StorageReview.com.

[Next page](<https://devfeed.tech/topics/isaac.md?cursor=WyIyMDI2LTEwLTIxVDE4OjA0OjQyKzAwOjAwIiwgIjk4NTM0MDQwLTUwMTctNDY3YS1iZTY0LTY2Y2YzNmRiNTZiOSJd>)