Robot, starting to take K12

Article | Embodied Business Outpost

Recently, the field of robotics has begun to focus on K12 education.

This reminds me of the hot search for Xiaotiantian phone and watch a few days ago. The reason is that the Little Genius Phone Watch has formed a relatively closed social circle for children: all the classmates around them are using it, and even if parents think it is expensive, it is difficult not to buy it.

No matter which angle you look at it, buying watches, tablets, and learning machines is a necessity for primary school students.

A Beijing parent told Embodied Business Outpost: “A little genius costs RMB 3,000 to 3,000, and a learning machine costs RMB 3,000 to 4,000. If a robot can help children learn something, I would be willing to buy one and give it a try within 10,000 RMB.”

On October 30, Digua Robots joined forces with Yushu Technology to launch an embodied intelligent education solution for primary and secondary schools; on the same day, GoQing Dynamics released an overall solution for embodied intelligent education at the China Educational Equipment Exhibition; last night, Qiyuan Robot announced the first OTA of T1 and Q1, with upgraded content including Hi Echo oral training, AI enlightenment courses and robot skill development.

Please enter a caption

All three news items point to the same thing: robots are fighting for children’s desks.

In the past, robotics education mainly served universities and research institutions. Wang Xingxing said in his early years that most of the buyers of Yushu products were universities and research institutes, which were used for papers, research and teaching. Now, the robot starts walking toward the younger children.

The three companies have different positions in entering K12: Digua focuses on the underlying computing power, Gaoqing solves how to open schools, and Qiyuan brings robots into families.

My judgment is straightforward: after phones, watches, learning machines and tablets, robots are vying for the position of the next generation of educational terminals.

Before looking at the robot, you must first look at the core

If you want robots to enter K12 classrooms, the first step is to enable students to get started.

It’s fun for children to watch robots running, jumping, and somersaulting. But an embodied intelligence class ultimately requires them to figure out: how the robot sees the environment, how to make judgments, and how to turn judgments into actions.

This link used to be complicated. Traditional educational robots often put visual recognition and motor drive on different boards. One is responsible for seeing and the other is responsible for moving. In the middle, communication, delay and debugging must be handled. However, a class in primary and secondary schools only lasts for more than 40 minutes, and complex software and hardware failures will directly occupy teaching time.

So when I looked at this collaboration, the first thing I looked at was the Rising Sun S100.

It integrates CPU, BPU and MCU in one SoC. The CPU is responsible for general computing, the BPU is responsible for AI reasoning, and the MCU is responsible for real-time motion control. Perception, decision-making and action are put into the same computing platform. The robot can form a closed loop on the end side from “seeing the stairs” to “deciding how to go” to controlling the joints to complete the action.

Put into the classroom, this change is very real. With fewer boards, communication links are shorter and deployment and debugging are simpler. Students can spend time on algorithms, tasks and actions, and it is easier for teachers to put a real machine development into a complete lesson.

Di Gua found Yu Shu again. I think this combination is very smart. According to Wang Xingxing’s previous disclosure, the global market share of Yushu’s quadruped robots exceeds 60%.

Go2 quadruped robot and R1 humanoid robot provide real bodies, and Rising Sun S100 provides end-side computing power and integrated computing and control support. Digua put his advantage on Duandian’s calculations and handed his body to Yushu.

The learning path starts from desktop robots equipped with RDK X3 and RDK X5, and then progresses to the comprehensive application of quadruped and humanoid robots. The entire program has designed 15 theme courses and 180 lessons, covering primary schools, junior high schools and high schools; there are also 6 course integration courses, including dual robot relay, Go2 stable transportation and R1 robot referee.

Hu Chunxu, Vice President of Digua Robot Ecology, said: “Ushu’s R&D capabilities in core components, motion control and complete machine products are well complementary to the Digua Robot intelligent computing platform. The high-performance four-legged, humanoid body will bring richer development scenarios to the Digua Computing Platform, and education is the starting point for long-term collaboration between the two parties.”

This sentence also points out the position of Diguo: it stands at the bottom of the teaching system, supporting the ontology, tools and courses on the same computing power base.

These 180 lessons all focus on one thing: making the code come out of the screen and turn into robot actions. This is the most intuitive difference from ordinary programming classes.

A lot of companies are showing off what robots can do. What Diguo does is to enable teachers to teach and students to make corrections, and to enable a set of real-machine experiments to run in forty minutes. The chip hidden in the robot body is also helping embodied intelligent education truly enter the classroom.

Gao Qing turns robots into a continuous course

Digua solved “how to teach”, and Gao Qing continued to solve “how to start classes”.

Gaoqing robot has a very intuitive advantage: Mini Pi plus is small in size, flexible in movement, and looks like a familiar cartoon friend of children. This feeling of intimacy first opened the door for the course.

But cuteness can only attract children, and lessons can retain interest.

Gaoqing bundles lightweight high-power hardware, open software platform, course teaching system and event practice export. Younger students can drag and drop graphics modules to make the robot stand, walk, and perform; older students can write their own programs, connect the robot, download tasks, and watch the code drive the real machine. Teachers have lesson plans, project task packages, teaching platforms and training resources in their hands.

For the school, this is the most realistic step: if the robot is bought back, the teacher can really start the class; the same equipment can continue to be used in clubs and competitions.

Outside of the classroom, robots can also be used in campus performances, club activities and football games. The actions and codes learned in the classroom are finally put to the test in real tasks.

According to Gao Qing, relevant courses have been carried out on a regular basis in schools such as Renmin University High School and Dongzhimen Middle School, and have provided technical support for many schools to participate in robotics competitions.

For children who actually sit in front of the robot, the feedback is very direct. At the 2025 Beijing Economic Development Zone Primary and Secondary School Students Science and Technology Festival, Chen Yutu, a student from Beijing Yizhuang Experimental Middle School, controlled Xiaopai to participate in the robot football match. He said after the game: “The sense of technology and control are very real! My teammates also said it was like controlling Mario.”

I like the phrase “like controlling Mario.” For children, there is no need to start with technical terms, interest must arise first.

The most distinctive part of Gaoqing’s solution is here: the hardware attracts students, the courses help teachers start classes, the competitions give children goals, and the open platform allows for more in-depth development. From this, a robot moves from the classroom to clubs and competition fields, accompanying children in their continuous learning.

Qiyuan wants a robot to accompany his child home

Robots in schools mainly allow children to understand and develop robots. Qiyuan Robot’s approach is more direct: let robots participate in children’s daily learning.

Parents will understand Qiyuan’s approach as soon as they hear it: let the robot practice English with their children, and then let the children learn embodied intelligent programming.

Speaking requires frequent practice and timely feedback, and also relies on a patient partner; programming allows children to turn their thoughts into robot movements, expressions and language. The first round of OTA in Q1 and T1 is to strengthen these two things.

The Hi Echo equipped with Qiyuan supports natural multiple rounds of English conversations. AI can continuously respond and actively ask questions based on the context. The robot screen displays bilingual subtitles in real time. Children can also interrupt and pause at any time, and switch between slow, standard and fast speech speeds.

The learning software on the mobile phone has to wait for the child to open it on his own initiative, but the robot can come to the child and initiate exercises at a fixed time. Oral language learning can also extend from fixed topics on the screen to rooms, objects and daily actions.

After the exercise, PrimeBot APP will generate a complete report. Parents and children can see ratings, pronunciation and grammar issues. A chat thus becomes a learning record that can be reviewed.

Hi Echo’s children’s mode is aimed at K12 students. It uses gamified scenarios and age-appropriate topics to guide children to take the initiative to speak and complete vocabulary accumulation and basic oral training.

I personally prefer T1. It has integrated Hi Echo oral training in this OTA, and also added a new AI enlightenment course.

Let’s look at Q1, which can be taken away with a backpack. It further opens up action choreography and skill development. Children can program themselves, add new abilities to the robot, or let the robot practice speaking with them.

To put it bluntly, Youdao provides a professional Hi Echo learning system, and Qiyuan gives this system a body that can move, perceive, and interact. The robot has thus changed from a teaching object to a learning partner, and the buyers have also extended from schools to ordinary families.

Among the three companies, Qiyuan is closest to the logic of consumer products and most similar to an educational robot that parents are willing to take home.

結語

Looking back now, Digua, Gao Qing and Qiyuan just form the three levels for robots to enter K12: the underlying technology can enter the classroom, schools can have classes, and families have daily use scenarios.

This will also change the robotics business. In addition to purchasing the robot itself, schools also need courses, teacher training and event services; after families purchase robots, they also need learning content, skill packages and OTA upgrades. A hardware sale has the opportunity to be extended to ongoing content and services.

Changes are also more intuitive for children. In the past, when learning programming, the code mostly stayed on the screen; now, a few lines of programs can allow robots to recognize people, avoid chairs, and kick footballs. Mathematics, physics, mechanics, programming and AI meet on the same moving carrier.

When this road truly opens, there will be a very real sign: the price of robots will enter the learning machine range, and courses, competitions and OTAs can be continuously updated.