Automation creates a skills gap because the need for new skills often outpaces training availability. Workers need new competencies, such as operating, maintaining, and programming automated systems, which are not covered by traditional training. Digital e-learning is effective for delivering knowledge across shift-based workforces, though hands-on practice remains essential for practical competence. Successful upskilling requires integrating training planning directly into the automation project timeline. Manufacturers must treat workforce upskilling as an operational priority, budgeted alongside hardware, to realize the full productivity gains of automation.
Waymo is initiating the deployment of its fully autonomous ride-hailing service in Spain, starting with the Region of Madrid. The company plans to welcome its first riders in 2027, beginning with a phased, safety-first approach involving manual driving validation first. Waymo is collaborating with Spanish authorities and investing in local operations to align with regional and national safety goals. Data indicates that Waymo Driver has been involved in 95% fewer crashes involving serious injury or worse than human drivers. The autonomous technology aims to enhance urban mobility by reducing accidents caused by human error.
Twitter Library features curated content on AI concepts, including model architectures, RL approaches, and agentic memory, presented through infographics. Topics covered span embodied AI and robotics breakthroughs, AI agent frameworks, and advancements in LLM attention mechanisms. The content also reviews open-source tools for AI guardrails, RAG, and various AI model types. Turing Post, the source, tracks research in areas like LLMs, AI agents, and AI infrastructure. The platform aims to provide precise analysis for engineers, researchers, and founders interested in how machine learning and AI function.
Gecko Robotics secured an additional $300 million contract with the U.S. Navy and GSA. This increase quadruples the ceiling value of the contract to improve fleet readiness and reduce maintenance delays. The company's AI and robotic technology, including the Komodo system, identifies structural issues up to 50 times faster than traditional methods. This technology allows for damage assessment without stripping deck coatings, saving the Navy significant repair costs. The expanded contract supports the Navy's goal of achieving 80% fleet readiness by accelerating inspections.
Daimon Robotics is building the infrastructure layer connecting tactile sensing, data collection, simulation, and AI models for tactile intelligence. The company aims to provide tools for other developers to build more capable robotic systems rather than developing a single foundation model itself. They develop a pipeline to process tactile feedback at 30 Hz, linking real-world data with simulations to bridge the sim-to-real gap. This infrastructure supports manipulation tasks requiring robots to react to physical contact, such as assembly and packaging. Ultimately, Daimon seeks to create a tactile-grounded world model that allows robots to reason about physical interactions beyond visual perception.
Kodama raised $7 million to advance physical AI in the forestry industry, focusing on commercial deployments in the southeastern U.S. The company uses a hybrid approach combining supervised autonomy and teleoperation to enhance machine capabilities on job sites. Their technology has already enabled machines to move over 100,000 tons of timber in commercial logging operations. Kodama is integrating with existing Original Equipment Manufacturers (OEMs) to leverage existing machine knowledge and build a data advantage. This data collection aims to improve on-site productivity and enhance safety for forestry workers.
Agility Robotics launched Digit 5, the first humanoid robot designed for safe and economic work at scale in warehouses. Digit 5 is engineered to lift 23-kilogram bins and can operate safely near human coworkers using proximity sensors. The development of Digit 5 involved prioritizing customer needs and safety regulations over the original vision of a purely graceful design. This resulted in a larger, more functional robot that balances commercial viability with achieving the ultimate goal of human-like motion. Ultimately, Digit 5 serves as a practical foundation for pursuing the longer-term vision of fully agile and dexterous humanoid robots.
Toyota Material Handling is utilizing Nvidia simulation and digital twin technologies to accelerate the development and commissioning of autonomous forklifts. This simulation-first approach aims to reduce commissioning effort by up to 50 percent and shorten time-to-market by up to 40 percent. The process involves using Nvidia Omniverse and Isaac Sim to virtually test autonomous vehicle systems in simulated factory environments. Digital twins allow teams to test navigation, safety logic, and real-world interactions, enabling early issue resolution and reducing on-site adjustments. This method also incorporates machine vision using synthetic data to improve object detection for pallet handling systems.
Agility Robotics partnered with Nyobolt to upgrade its Digit 5 humanoid robot with a fast-charging battery system. This technology enables Digit 5 to work over 20 hours daily, improving uptime and operational economics. The new battery system allows for a 90-minute runtime with a 9-minute charge, achieving a 10:1 run-to-charge ratio. This innovation addresses power limitations that previously constrained autonomous robot fleet economics. The enhanced battery performance is designed to boost the value proposition of Digit 5 for warehouse and manufacturing operations.
ABB Robotics has introduced two new compact, high-performance robots, the IRB 1700 and IRB 1700ID. These robots offer up to 20 percent shorter cycle times than comparable models, improving productivity in precision manufacturing. They feature payloads from 6 kg to 12 kg, high precision repeatability of 0.03 mm, and joint speeds up to 700 degrees per second. The new robots also reduce operating costs with up to 15 percent lower power consumption and longer service intervals. These systems are suitable for demanding tasks like arc welding and new electronics applications.
Compact robotic systems face complex thermal management challenges due to the proximity of multiple heat-generating electronic components with restricted airflow. Effective thermal design requires evaluating system-level performance based on actual operating conditions, including peak loads and duty cycles, rather than just component ratings. Heat sink selection should be based on a calculated thermal budget, considering the thermal interface and mounting surface quality. Forced-air cooling effectiveness is limited by internal restrictions, and the mechanical mounting interface significantly impacts heat transfer. Ultimately, reliable thermal management demands early collaboration between thermal, mechanical, and manufacturing teams to ensure practical and reliable solutions.
Tesollo, a South Korean robotics company, anticipates exceeding 1,000 cumulative sales of its robotic hands and grippers by the end of 2026. This growth is driven by increasing demand in humanoid robotics, physical AI, and industrial automation for flexible end-effectors. The company has expanded its production facilities and manufacturing infrastructure to meet this growing global demand. Tesollo currently offers a portfolio of articulated hands ranging from one to five fingers, serving both research and industrial applications. The company plans to further strengthen its manufacturing capabilities and pursue a potential public listing in South Korea.
ASTM International is developing a standard to measure manipulator and mobile manipulator task performance under controlled optical-signal degradation. This proposed method will quantify how degradation in optical information impacts a robot's ability to complete tasks in real-world conditions. Degradation from factors like changing light or interference can cause downstream machine-learning models to receive poorer information, potentially leading to task failure or safety risks. The standard aims to provide a repeatable way to test performance under stress, helping developers identify system limitations before deployment. This research supports responsible real-world deployment and UN Sustainable Development Goal #8.
Primepick is expanding its NSF-certified, robotic hot-meal kiosk, MealMate, to new locations. The kiosk serves hot dishes in 5-7 minutes and is being deployed in hospitals, nursing homes, and new sites in Huntsville, Chicago, and Dallas. MealMate provides access to hot meals outside traditional foodservice hours without requiring on-site kitchen staff. The system utilizes robotic arms, specialized storage, and proprietary software for automated management and inventory. This marketplace model connects customers with meals from local restaurants and dark kitchens.
Laplace Intelligence builds automated red-teaming systems to stress-test physical AI for robotics and self-driving cars before real-world deployment. The company uses high-fidelity simulations, grounded in real-world data, with AI agents actively probing for failure points. Their method involves using AI to design corner cases by analyzing real-world data like news and police reports to challenge autonomous systems. This approach aims to create a shared foundation of evidence for safety and build trust among customers, insurers, and regulators. Laplace seeks to establish a "dynamic interactive testing process" as a new paradigm for physical AI safety.
Schaeffler launched its Taicang Humanoid Lighthouse on October 20, 2026, deploying six humanoid robots on a joint module assembly line. The 7,500 square meter facility integrates robot component production, humanoid training, and industrial deployment. The robots perform tasks such as material handling, sorting, fitting, inspection, and packaging within the site. The factory offers six training scenarios and uses a digital twin system to link training with production validation. Schaeffler intends for the facility to serve as a collaboration platform for Chinese robotics companies.
SPD robot training utilizes simulated demonstrations to develop a dexterous bimanual policy for real tasks like Jenga. This method combines extensive virtual teleoperation with brief real teleoperated demonstrations to gather training data for two arms and two hands. Five individuals collected 75 hours of demonstrations in VR, involving a 56-DOF simulated system across various scenes and objects. The collected data was used to pretrain a 222 million parameter diffusion transformer policy before adapting it to real data. Simulation pretraining outperformed behavior cloning for these tasks, offering a practical signal for evaluating performance.