Industry Briefing
A single destination for timely, editor-curated robotics news from around the world.
Hugging Face Introduces Open-Source Humanoid HOPEJr and Desktop Robot Reachy Mini
Hugging Face has unveiled two innovative open-source robots designed to enhance accessibility in AI development. The first, HOPEJr, is a fully humanoid robot priced under $3,000, developed in collaboration with The Robot Studio. The second, Reachy Mini, is a desktop head created by the Pollen Robotics team. This launch, which aims to democratize robotics technology, reflects Hugging Face's commitment to making advanced robotics more affordable and user-friendly for developers and researchers alike.
open-source pollen-robotics hugging-face the-robot-studio
Hugging Face Acquires Pollen Robotics, Eyes Open Source Future for Humanoid Hardware
AI platform Hugging Face has made its inaugural foray into hardware by acquiring French startup Pollen Robotics. This acquisition, announced recently, includes Pollen's open-source humanoid robot, Reachy, along with its development team. The move reflects Hugging Face's strategic expansion into the robotics sector, reinforcing its commitment to open-source principles that have traditionally guided its AI software initiatives. By integrating Pollen's technology and expertise, Hugging Face aims to enhance its offerings and innovate within the robotics field.
open-source pollen-robotics hugging-face
NVIDIA and Hugging Face integrate AI model "GR00T 1.7" into open-source robot development platform "LeRobot."
NVIDIA and Hugging Face have announced the integration of their cutting-edge technologies aimed at enhancing humanoid robots. The collaboration will see the incorporation of NVIDIA's visual language action (VLA) model, known as NVIDIA Isaac GR00T 1.7, along with the remote operation framework, NVIDIA Isaac Teleop, into Hugging Face's open-source robot development library, LeRobot. This initiative is set to advance the capabilities of humanoid robots, enabling more sophisticated interactions and functionalities. The announcement highlights a significant step in the ongoing evolution of robotics, reflecting both companies' commitment to fostering innovation in the field.
NVIDIA and Hugging Face Enhance LeRobot with New AI Models and Frameworks
NVIDIA has expanded its collaboration with Hugging Face to enhance the LeRobot open-source robotics platform with new AI models and frameworks. This integration includes the NVIDIA Isaac GR00T 1.7 vision-language-action model and the Isaac Teleop framework, aimed at streamlining robot development. The partnership seeks to make advanced robotics tools more accessible to developers and researchers, with plans to incorporate NVIDIA Cosmos 3 in the future. This collaboration is significant as it addresses the fragmented nature of robotics development by providing standardized workflows for data collection, model training, and robot deployment. The introduction of the Isaac Teleop framework allows for high-quality training data collection through human demonstrations, which can be shared within the LeRobot ecosystem. By lowering barriers to entry, NVIDIA and Hugging Face aim to foster broader collaboration in the robotics community. Looking ahead, NVIDIA plans to integrate the Cosmos 3 model into LeRobot, which will generate synthetic robotics data and assist in policy development. The collaboration builds on existing resources, including a dataset with over 350,000 robot trajectories and 57 million grasp examples. No further timeline was disclosed at the time of publication.
AI and Robotics
NVIDIA and Hugging Face Collaborate to Enhance LeRobot Open Source Community with Physical AI
NVIDIA and Hugging Face have joined forces to enhance the LeRobot open-source robotics library by integrating NVIDIA's Isaac GR00T 1.7 and Isaac Teleop frameworks. This partnership, announced recently, seeks to streamline the robot development process for developers by offering a comprehensive, standardized open-source pathway that encompasses everything from data collection to deployment. The initiative is designed to significantly reduce the barriers to entry for developers venturing into physical AI development, making it more accessible and efficient.
Robotics Open Source AI Machine Learning Simulation
NVIDIA and Hugging Face bring new models and frameworks to LeRobot
Hugging Face has announced an update to its open-source robotics library, LeRobot, which is designed for training, running, and sharing robot datasets, models, policies, and workflows. This development comes in collaboration with NVIDIA, which is contributing new models and frameworks to enhance the capabilities of LeRobot. The partnership aims to streamline the process of robotics development and foster innovation in the field. This initiative is part of a broader effort to advance artificial intelligence and machine learning applications in robotics, making it easier for developers and researchers to access and utilize sophisticated tools. The announcement was made public recently, highlighting the growing synergy between AI and robotics as industries seek to leverage these technologies for improved automation and efficiency.
Artificial Intelligence Artificial Intelligence / Cognition Design / Development Development Tools / SDKs / Libraries Humanoids News
Hugging Face launches agentic toolkit for Reachy Mini
Hugging Face has introduced a new agentic toolkit designed for the Reachy Mini robot, enabling users to specify desired behaviors in straightforward language. This innovative tool allows users to articulate their requests, which the agent then translates into code, tests, and deploys to the robot. The launch of this toolkit marks a significant advancement in making robotic programming more accessible and user-friendly, catering to both developers and enthusiasts. By simplifying the coding process, Hugging Face aims to enhance interaction with robotics, fostering greater creativity and experimentation in the field.
Academia / Research Artificial Intelligence Artificial Intelligence / Cognition Design / Development Human Robot Interaction / Haptics News
The app store for robots has arrived: Hugging Face launches open-source Reachy Mini App Store with 200+ apps
Hugging Face, a New York City-based startup known for its open-source AI models, has launched the Reachy Mini App Store, a platform designed for its low-cost physical robot, Reachy Mini, which debuted in July 2025. Priced at $299, the robot aims to make robotics accessible to hobbyists and developers. The App Store features over 200 community-built applications, allowing users to download and create custom apps without needing a technical background. This initiative is part of Hugging Face's broader goal to democratize robotics, enabling individuals to develop functional software for the robot using simple English commands. The launch comes as Hugging Face has sold around 10,000 Reachy Mini units, with 3,000 sold in the past two weeks alone. The platform supports various AI models, including Hugging Face's own ML Intern, and offers tools for users to simulate and test their apps in a virtual environment. By removing traditional barriers to robotics development, Hugging Face aims to shift the focus from technical expertise to creativity in programming robotic behaviors. The App Store is part of Hugging Face's ongoing "Le Robot" project, which seeks to provide open-source resources for robotics development.
Technology
Humanoid Robotics Momentum: London and Shanghai Host Major Industry Summits Concurrently
On May 29-30, 2025, two significant robotics events are set to occur concurrently in London and Shanghai, marking a pivotal moment in the humanoid robotics sector. The Humanoids Summit in London will convene industry leaders to discuss mass production and investment strategies, aiming to drive economic growth and innovation in robotics. Meanwhile, the Embodied Intelligence Developer Conference in Shanghai will emphasize the importance of policy support, talent development, and practical applications, reflecting China's commitment to advancing humanoid robotics. This dual event highlights a global push towards accelerating advancements in humanoid technology, showcasing the diverse approaches taken by different regions to foster growth in this rapidly evolving field.
Unitree Robotics Neura Robotics kepler-robotics pollen-robotics AGIBOT hugging-face
NVIDIA and Hugging Face Bring New Models and Frameworks to LeRobot for the Open Robotics Community
New LeRobot integrations give developers open access to NVIDIA Isaac GR00T 1.7, Isaac Teleop, datasets and robotics workflows, with NVIDIA Cosmos 3 integration planned to bring frontier world models to open robotics development.
NVIDIA and Hugging Face Bring New Models and Frameworks to LeRobot for the Open Robotics Community
Open source AI has shown how quickly developers can innovate when models, data and tools are shared. Robotics has the same opportunity, but advancements in physical AI development can still be gated by costly and fragmented resources, from large ...
Hugging Face’s new $2.5k LeRobot Humanoid brings 3D-printed robotics within reach
Humanoid robots continue to be largely inaccessible to the general public due to their prohibitively high costs. As of October 2023, advancements in robotics have not yet translated into affordable options for consumers, leaving many interested in the technology unable to purchase or utilize these sophisticated machines. The ongoing research and development in the field aim to enhance the capabilities of humanoid robots, but the financial barriers remain a significant hurdle. Experts suggest that until production costs decrease and more efficient manufacturing processes are developed, widespread adoption of humanoid robots will remain unlikely. The situation highlights the need for innovation in both technology and economic models to make these advanced robots available to a broader audience.
Open-Source Software Is Starting to Help Robots Think
A new wave of open-source robotics is transforming the field, enabling a broader range of individuals to develop advanced robotic systems. This movement, which gained momentum in the last two years, has seen major companies like Hugging Face, Nvidia, and Alibaba invest in tools and models aimed at enhancing robots' cognitive abilities. The initiative builds on the success of the Robot Operating System (ROS), introduced in 2007, which standardized robotics software and significantly reduced the time researchers needed to focus on their core projects. Nvidia has developed a comprehensive open-source robotics stack that includes simulation tools and pre-trained models, making it easier for developers to create sophisticated robots without needing extensive expertise. Hugging Face has furthered this effort with its LeRobot platform, which has rapidly expanded its dataset offerings, now boasting over 58,000 robotics datasets. While the open-source approach democratizes access to robotics technology, concerns arise regarding the influence of commercial interests in shaping the field. Experts warn that as companies drive the movement, there is a risk of duplicating existing solutions rather than innovating. Nevertheless, the growing community of contributors—from industry leaders to hobbyists—signals a promising future for open-source robotics, fostering collaboration and creativity in a field that is becoming increasingly accessible.
Open-robotics Robot-ai Huggingface Ros Nvidia
NVIDIA Open Sources Embodied Intelligence Toolchain to Enhance Robotics Development
On July 6, NVIDIA integrated three key components into Hugging Face's open-source robotics library, LeRobot: the GR00T N1.7 model, Isaac Teleop framework, and the upcoming Cosmos 3. This collaboration connects NVIDIA's 3 million robot developers with Hugging Face's 16 million AI builders, facilitating access to pre-trained models and data. This initiative is significant as it shifts NVIDIA's focus from merely creating models to building an ecosystem that addresses data bottlenecks in embodied intelligence development. The Isaac Teleop framework standardizes data collection, allowing for easier sharing and reuse within the community, which is crucial for advancing robotics. Looking ahead, the integration of GR00T N1.7 and Isaac Teleop into the LeRobot workflow marks a pivotal moment for robotics developers. No further timeline was disclosed at the time of publication.
Robotics Open Source AI Development Data Collection Machine Learning
Ant Group Launches LingBot-VA 2.0 with Advanced Spatial Awareness and Vision Models
On July 10, Ant Group introduced LingBot-VA 2.0, the first embodied native action model in the industry. This release marks the culmination of their full-stack 2.0 model series, featuring advancements in spatial perception and real-time interaction. The launch has generated significant buzz across international platforms like Reddit and Hugging Face, indicating a strong interest in their technology. The significance of LingBot-VA 2.0 lies in its comprehensive technology stack, which includes LingBot-Depth 2.0 and LingBot-Vision. LingBot-Depth 2.0 enhances depth perception with a training dataset expanded from 3 million to 150 million, achieving top scores in 12 out of 16 benchmarks. Meanwhile, LingBot-Vision introduces a novel pre-training target, improving depth estimation accuracy despite using a smaller dataset compared to competitors. Looking ahead, the next steps for Ant Group involve further collaboration with industry partners, as LingBot-Depth 2.0 has already received professional certification from Orbbec. The company is also focusing on integrating their models into the industry, with no further timeline disclosed at the time of publication for upcoming releases or partnerships.
Robotics AI Machine Learning Computer Vision
Ant Group Releases Open Source LingBot-VLA 2.0 with 60,000 Hours of Real-World Data
On July 8, Ant Group unveiled the upgraded LingBot-VLA 2.0 model, which has been enhanced through the integration of 60,000 hours of high-quality real-world data. This latest version offers improved support for a variety of robot configurations and degrees of freedom, leading to increased efficiency and performance in both dual-arm operations and mobile tasks. The model has been made accessible to developers via platforms such as Hugging Face and GitHub, promoting wider collaboration and innovation in robotics.
Robotics AI Open Source Machine Learning
ABB Robotics completes its AI-powered Visual SLAM AMR portfolio with new autonomous forklift
ABB Robotics completes its AI-powered Visual SLAM AMR portfolio with new autonomous forklift Visit http://go.abb/robotics for further information -The new Flexley Stack F712 extends ABB Robotics’ AI-powered Visual SLAM technology to autonomous forklifts, enabling pallet transport and high-density storage. -Customers can now deploy mixed fleets of Visual SLAM-powered tugs, movers and forklifts on a common navigation, fleet management and software platform. -Powered by ABB Robotics' AMR Studio, the portfolio enables up to 20% faster commissioning while ensuring seamless interoperability and safe, reliable operation. 07/07/26, 07:10 AM | Industrial Robotics, Mobile Robots | ABB Inc. ABB Robotics is expanding its Autonomous Mobile Robotics (AMR) portfolio with the launch of the Flexley® Stack F712, creating a complete interoperable ecosystem across all major Visual SLAM AMR types. Combining autonomous forklifts, tugs and movers on one platform, ABB Robotics enables customers to automate a broader range of material-handling and intralogistics processes. Offering market-leading accuracy, the F712 is designed for demanding material handling, end-of-line storage and warehouse operations across industries including automotive manufacturing, helping increase efficiency, flexibility and scalability. More Headlines A3's Automate 2026 Breaks Records as Demand for Robotics, AI and Automation Grows NVIDIA and Hugging Face Bring New Models and Frameworks to LeRobot for the Open Robotics Community Palladyne AI Executes $4.2 Million U.S. Air Force Contract to Advance Swarming Capabilities for Integrated Cross-Domain Operations UMA Unveils Its Vision for the Next Generation of Humanoid Robots Robbyant Unveils LingBot-Depth 2.0 and LingBot-Vision to Redefine Robotic Spatial Perception Articles Unleash AI Innovation: The Power of NVIDIA RTX PRO 6000 Blackwell Workstation Edition Fueled by PNY-Supplied GPUs Automate 2026 Q&A with DESTACO Automate 2026 Q&A with Roboteon Advances in Robots to See & Interpret within Warehouse Environments Building Resilient Fulfillment Networks with Robotics and Real-Time Logistics Data "Across intralogistics operations, businesses are being asked to process greater volumes in less time, while working with increasingly limited resources," said Marc Segura, President, ABB Robotics. "They are under pressure to move goods faster and with greater flexibility, while labour availability is becoming a critical constraint. As part of our journey to more autonomous and versatile robotics (AVRTM), we have combined advanced vision, mobility and intelligence in the Flexley Stack F712 forklift AMR, completing our scalable, AI-powered AMR portfolio." F712 is versatile, capable of handling multiple load types and sizes - including open and closed pallets, containers or racks- up to 2,000 kg and reaching heights of 8.5 meters. The Flexley Stack AMR F712 joins the Flexley Tug and Flexley Mover in ABB Robotics' growing Visual SLAM AMR portfolio. Applications include intralogistics tasks such as warehouse storage and retrieval, as well as line supply, end-of-line handling, body- and press-shop and drive-in and light buffer in the automotive and industries sector. Unlike conventional AMR forklifts on the market, F712 uses Visual SLAM to map and navigate its environment, eliminating the need for pre-installed infrastructure like markers or reflectors. The AI-enabled Visual SLAM supports the autonomous decisions required to operate in complex, dynamic warehouse operations with a market-leading positional accuracy of ±10 mm. Together with AMR Studio®, this shortens commissioning times by up to 20 percent and creates a versatile and reliable system that can adapt instantly when a warehouse or production floor layout changes. Certified to the latest ISO and ANSI safety standards, Flexley Stack F712 can safely operate at class-leading speeds of up to 1.7 m/s while loaded. F712 is fully integrated with AMR Studio and is VDA5050 compatible, enabling seamless integration with ABB Robotics' Visual SLAM AMRs and existing systems within a unified project. This makes it easy to manage complex projects and integrate different types of mobile robots. The no-code, drag-and-drop software suite supports rapid setup, fleet coordination, traffic management and real-time visualization, allowing ABB Robotics' tugs, movers and forklifts to operate together in the same layout for scalable turnkey automation projects. ABB Robotics as one of the world's leading robotics companies, is the only company with a comprehensive and integrated AI-powered portfolio covering robots, cobots and Autonomous Mobile Robots (AMRs), designed and orchestrated by our value-creating software. We help companies of all sizes and sectors - from automotive to electronics and logistics - to outperform by becoming more resilient, flexible and efficient. ABB Robotics is at the forefront of developing and commercializing a new generation of Autonomous Versatile Robotics
Palladyne AI Executes $4.2 Million U.S. Air Force Contract to Advance Swarming Capabilities for Integrated Cross-Domain Operations
Palladyne AI Executes $4.2 Million U.S. Air Force Contract to Advance Swarming Capabilities for Integrated Cross-Domain Operations Visit http://www.palladyneai.com for further information Palladyne AI’s SwarmOS™ platform to support satellite integration, marking a major expansion of its multi-domain autonomy and ISR capabilities across space, air, maritime, and land 07/07/26, 06:15 AM | Mobile Robots, Other Topics | Palladyne AI Corp. Palladyne AI Corp. (NASDAQ: PDYN and PDYNW) ("Palladyne AI"), a developer of artificial intelligence software for robotic platforms in the defense and commercial sectors, today announced that it has executed the previously announced contract awarded by the Air Force Research Laboratory (AFRL) to solve one of the most persistent challenges in modern defense operations—how to make different autonomous systems work together as one coordinated team. The "Hierarchical Adaptive Networked Game-Theoretic Integration of Multiple Echelons (HANGTIME)" contract will address this need. More Headlines A3's Automate 2026 Breaks Records as Demand for Robotics, AI and Automation Grows NVIDIA and Hugging Face Bring New Models and Frameworks to LeRobot for the Open Robotics Community ABB Robotics completes its AI-powered Visual SLAM AMR portfolio with new autonomous forklift UMA Unveils Its Vision for the Next Generation of Humanoid Robots Robbyant Unveils LingBot-Depth 2.0 and LingBot-Vision to Redefine Robotic Spatial Perception Articles Unleash AI Innovation: The Power of NVIDIA RTX PRO 6000 Blackwell Workstation Edition Fueled by PNY-Supplied GPUs Automate 2026 Q&A with DESTACO Automate 2026 Q&A with Roboteon Advances in Robots to See & Interpret within Warehouse Environments Building Resilient Fulfillment Networks with Robotics and Real-Time Logistics Data Today, drones, ships, and satellites often operate largely independently, limiting how quickly warfighters can see and respond to threats. HANGTIME will utilize Palladyne AI's patented SwarmOS™ software platform—the defense variant of the Palladyne™ Pilot embodied AI software—as the baseline technology to bridge that gap, connecting disparate systems so they can share intelligence, adapt to changing conditions, and act in sync across domains, including space, air, maritime, and land. By integrating satellites for the first time, this project also extends Palladyne AI's technology from the ground to orbit, enabling faster, more informed decision-making and coordinated mission execution, turning tactical commanders into strategic commanders by giving them more cross-domain intelligence, surveillance, and reconnaissance (ISR) capabilities than ever before. "Our collaboration with AFRL showcases what's next for autonomous operations," said Ben Wolff, President and CEO, Palladyne AI. "This isn't about replacing humans—it's about giving them sharper, faster insight. By connecting satellite, aerial, and ground systems using the patented SwarmOS embodied AI platform as a foundational technology, we're helping the warfighter make better decisions in real time and stay one step ahead on the battlefield." "The HANGTIME project is a breakthrough that unites high-altitude assets and situational unmanned systems into one coordinated sensor network—delivering a major advantage for the defense industry," said Dr. Denis Garagic, Chief Technology Officer, Palladyne AI. "For the first time, a single AI framework can coordinate assets across multiple domains, including satellites. That means these systems can now think and act together as a team, sharing what they see and learning as conditions change." "The HANGTIME effort represents a critical step in multi-domain autonomy for coordinated execution in challenging environments," said Caleb Williams, Program Manager, AFRL/RIEA. For more information on Palladyne AI and its patented collaborative autonomy software, including SwarmOS, please visit www.palladyneai.com. For more information about AFRL, please visit www.afrl.af.mil. About Palladyne AI Palladyne AI is a U.S.-based technology company developing patented embodied artificial intelligence, collaborative autonomy solutions, advanced avionics, autonomous systems, advanced UAV engineering services, and precision-manufactured components for defense and industrial markets. Palladyne AI delivers secure, American-developed and operated platforms designed to meet the stringent requirements of U.S. government and public-sector customers, including data sovereignty, security, and compliance. Palladyne AI's embodied AI is designed to operate in complex, contested, and high-risk environments, enabling distributed tasking, human-on-the-loop decision-making, degraded-communications resilience, and multi-domain coordination. Its platform-agnostic autonomy stack combines real-time sensor fusion, adaptive AI models, and edge-native orchestration—without vendor lock-in—to support autonomous and collaborative systems across air, ground, maritime, and industrial domains w
Liquid AI's smallest model yet LFM2.5-230M beats models 4X its size at data extraction, can run 'anywhere'
Liquid AI, a company founded by former MIT computer scientists, has unveiled its latest AI language model, LFM2.5-230M, which is designed for efficient data extraction and local deployment on devices such as smartphones and laptops. Released today, this 230-million-parameter model is noted for its ability to run on various hardware platforms, outperforming larger models like Alibaba's Qwen3.5 and Google's Gemma 3 in specific benchmarks. Targeting developers and engineers, LFM2.5-230M operates under a dual-use commercial license, allowing free access for individuals and companies with annual revenues below $10 million, while larger enterprises must secure a paid agreement. The model distinguishes itself by utilizing the LFM2 architecture, enabling high inference speeds with a minimal memory footprint, making it suitable for edge computing. Liquid AI's launch reflects a broader industry shift towards architectural efficiency rather than sheer parameter counts, as major AI firms focus on models with hundreds of billions of parameters. The LFM2.5-230M is specifically tailored for lightweight data extraction tasks, allowing businesses to automate processes without relying on costly cloud services. In practical applications, the model has been successfully deployed in a humanoid robot, demonstrating its capability to process complex commands efficiently. Available immediately on platforms like Hugging Face, LFM2.5-230M aims to revolutionize how enterprises manage data extraction, moving away from traditional, rigid systems to more adaptable AI-driven solutions.
Technology
UMA Exits Stealth: The European Robotics ''Supergroup'' Arrives in Paris
Rémi Cadene, a former researcher at Tesla and Hugging Face, has unveiled UMA, a new venture that boasts a formidable founding team and aims to develop general-purpose robots by 2026. The announcement highlights the company's ambitious vision to revolutionize the robotics industry, leveraging advanced technologies to create versatile machines capable of performing a wide range of tasks. With a strong emphasis on innovation and practicality, UMA seeks to address the growing demand for automation across various sectors. The initiative is set to position itself at the forefront of the robotics field, promising significant advancements in functionality and efficiency.
Uma EuropeRobotics needs a service framework.
RSF defines a common language for robot service capability, lifecycle operations, certification pathways, and service-provider networks.
