Robotics

Today's Stories

1. 1

   NVIDIA launches robotics platform to bring AI robots into real-world jobs

   NVIDIA released Isaac GR00T, a platform for building humanoid robots, and Cosmos 3, a system that helps train robots in virtual environments before they work in the real world. The tools target healthcare automation, manufacturing, and autonomous vehicles. The company is shifting focus from pure data center AI to robots that can perform physical tasks.

   This could accelerate the arrival of AI robots in hospitals, factories, and warehouses, potentially changing how physical work gets done in these industries.

2. 2

   Figure AI ramps up robot production while facing real-world deployment challenges

   Figure AI is scaling robot production to 55 units per week, but commercial applications are still being developed. Meanwhile, Verobotics deployed window-cleaning robots at NVIDIA's campus but had to use a 60/40 split with human workers due to construction site dust interfering with the robots.

   Robot companies are discovering that scaling production is easier than solving real-world problems like dust, weather, and unpredictable environments.

3. 3

   ABB plans to sell robotics division to SoftBank while expanding AI data center work

   ABB is preparing to complete the sale of its robotics unit to SoftBank Group while deepening its partnership with NVIDIA for AI data center infrastructure. The move reshapes ABB's focus toward power systems and automation for AI computing facilities rather than manufacturing robots.

   This could change the competitive landscape for industrial robots and affect pricing and availability of factory automation equipment.

4. 4

   SK Telecom partners with NVIDIA to build massive AI cloud for Korean robotics companies

   South Korea's largest telecom company SK Telecom announced plans to build a gigawatt-scale AI cloud using NVIDIA's DSX platform, with the first facility launching in 2027. The infrastructure will support AI startups, robotics development, and industrial automation in Korea.

   Korean companies developing robots and AI applications will have access to powerful computing resources domestically, potentially accelerating innovation in these fields.

5. 5

   AGIBOT launches competition to test how AI performs on real robot tasks

   Chinese robotics company AGIBOT announced the World Challenge 2026, focusing on how AI models perform when controlling actual robots doing real tasks rather than just computer simulations. The industry is moving toward testing robots in real environments with real challenges.

   This shift from lab testing to real-world trials could reveal which AI systems actually work for practical robot applications in homes and workplaces.

6. 6

   Open-source companion robot project adds voice and emotional expressions

   An independent developer completed the 'expression engine' phase of Olaf, an open-source companion robot that can speak naturally, move its head and ears in sync with speech, and display a beating heart. The project demonstrates multilingual conversation abilities and is preparing for mobility features.

   Open-source robot projects like this could make companion robots more accessible and affordable for personal use, similar to how open-source software democratized computing.