Open-source toolkits, shared datasets and large robot contests are changing what students can build and study. See what is new, then equip your classroom or lab through The Humanoid Group.
What moved in education robotics this year.
17,802 competitors from 26 countries in the World Robot Conference's contests (Beijing Municipal Government, 2026)
720+ hours of two-arm demonstrations in Ai2's open MolmoAct 2 dataset (Allen Institute for AI, 2026)
6 simulation benchmarks built into LeRobot 0.6's evaluation command (Let's Data Science, 2026)
World Robot Conference contests draw 11,418 teams from 26 countries in Beijing
· Milestone
Beijing's 2026 World Robot Conference, held August 19 to 23, reported 11,418 teams and 17,802 competitors from 26 countries in its on-site competitions. The event also drew 557,000 visitors and showed more than 3,000 robot-related products, according to the city government's closing report.
Why it matters: Large contests show how widely robot building has spread. Teachers entering teams should check each event's rules, kit limits and travel costs early, and plan practice robots well before the season.
Survey maps how far robot foundation models are from running on a specific robot
· Research
Researchers surveyed 21 groups of robot-learning methods, including RT-X, Octo, OpenVLA and π0, and mapped how much each shares across robot bodies against how much work remains to run it on one robot. Three researchers coded the methods independently, and none was placed in the safety and recovery row.
Why it matters: Success rates alone hide how much new data or tuning a model needed. When a lab or class adopts an open model, budget time to adapt it to the exact robot and ask what was collected to reach published results.
China publishes a 2026 standard system for humanoid robots and embodied AI
· Policy
A preview of Beijing's 2026 World Robot Conference reports that China has named embodied intelligence a core direction for future industries and that its industry ministry has released the Humanoid Robot and Embodied Intelligence Standard System (2026 Edition). It describes the sector moving from technical testing toward large-scale factory deployment.
Why it matters: Standards shape how humanoids are tested, connected and certified. A university building a humanoid lab should ask each maker which standards its platform follows, and keep the answers with the purchase record.
Data protection in schools— Department for Education (GOV.UK). Guidance on personal data in schools, including CCTV and photos, EdTech procurement and generative AI.