Archives - 2026

JD.com founder Richard Liu said at the APEC China CEO Forum that robots will replace the company’s 700,000 delivery workers “sooner or later”, pairing the prediction with a retraining program called Nirvana that aims to move couriers into robot maintenance and AI training roles before automation arrives.

U.S. industrial robot installations grew 11% year-on-year to 38,000 units in 2025, according to the International Federation of Robotics, with food industry adoption surging 30% as the country climbs to eighth globally in robot density – still roughly ten times behind China’s annual installation pace.

NVIDIA has announced Halos for Robotics, the industry’s first full-stack safety architecture for physical AI systems, drawing on 18,600 engineering years of autonomous vehicle safety development, with Agility Robotics integrating it into Digit as the first commercial humanoid deployment.

Sanctuary AI has validated a 99.5% task success rate and 2.54-second cycle time on a wire-plugging task at a global Tier 1 automotive supplier’s live production line, as the company shifts strategy to deploy its physical AI on existing industrial robot platforms rather than waiting for humanoid commercialization.

Kawasaki Robotics has debuted the RL030N at Automate 2026 in Chicago, an 8-degree-of-freedom robot described as the industry’s first platform designed specifically for physical AI applications, alongside patented Pulseboard weld inspection technology capable of 10x faster inspection speeds.

CYBERDYNE and Pegasus Tech Ventures have launched a JPY 10 billion ($60 million) corporate venture capital fund to invest in startups developing robotics, physical AI, healthcare, and intelligent systems, with CYBERDYNE as the sole limited partner and Pegasus managing the fund as general partner.

Autonomique, a Physical AI software company spun out of SRI, has moved its robot autonomy platform from a paid pilot to live production deployment at F&P Mfg., a Canadian Tier-1 automotive supplier serving Honda, Toyota, and General Motors, with a global rollout across F.tech’s manufacturing network under discussion.

As Genesis AI’s Eno joins a growing list of robots that reject bipedal human appearance in favor of functional minimalism, the industry is quietly splitting into two camps with different answers to the same question: does a robot need to look human to be useful in a human world?

D-Link has publicly confirmed its entry into robotics for the first time, establishing a dedicated subsidiary focused on AI therapeutic robots with software and platform services rather than hardware, targeting initial shipments exceeding 100,000 units in Q2 2027.

Chinese automaker Seres has unveiled Xiaosai, its first humanoid robot, designed for visual recognition, autonomous greeting, and voice interaction, as the company expands its embodied AI presence beyond the Xiaosai 01 and 02 inspection robots already operating at its smart factory.

Swiss deep-tech company Verity has won the 2026 IERA Award for Innovation and Entrepreneurship in Robotics and Automation for its fully autonomous indoor drone system, deployed across approximately 200 warehouses and capturing 500,000 images per day for continuous inventory tracking.

Genesis AI has unveiled Eno, its first general-purpose robot, featuring a wheeled base, height-adjustable articulated tower, and proprietary dexterous hands matched to human hand dimensions, powered by the GENE foundation model and targeting industrial customer deployments by end of 2026.

Genesis AI and LG CNS have announced a long-term strategic partnership to evaluate and deploy Genesis’s Eno general-purpose robot and GENE AI foundation model across LG Group manufacturing and logistics operations in the U.S., with global expansion planned.

Alibaba has introduced the Qwen Robot Suite, its first family of AI models for physical robotics, comprising three interconnected systems for navigation, manipulation, and world modeling, already in pilot testing with Alibaba Cloud enterprise clients.

Boston Dynamics has published a technical breakdown of how Atlas learned to perform soccer drills – including the Ghost Rabona – using professional motion capture data, motion retargeting, and cloud-based reinforcement learning that condenses a year of physical trial and error into 24 hours of simulation.