A Russian Company Helped Prepare a Fighting Robot for an International Competition
Cloud provider Turbo Cloud supplied computing capacity to MindRobots to train a humanoid robot competing in China’s international URKL (Ultimate Robot Knockout Legend) fighting league.
In Shenzhen, China, the Ultimate Robot Knockout Legend (URKL) tournament for humanoid robots is underway. Organized by EngineAI, the competition combines new technological advances with a show format and professional engineering competition. These are actual fights between humanoid machines, with robots striking each other until one wins. The tournament is designed to showcase the capabilities of modern robots that can analyze information from cameras and sensors, perform fast, coordinated movements and respond to an opponent’s actions. Teams from around the world are competing in China using EngineAI T800 androids. Russian company MindRobots has advanced to the semifinals with its humanoid machine.
Dedicated Cloud Infrastructure for a Robot
EngineAI T800 robots are manufactured in China, and the teams competing in the tournament receive identical hardware. They are responsible for configuring the machines for specific tasks themselves. To prepare its robot for the competition, MindRobots used dedicated cloud infrastructure from full-service provider Turbo Cloud.
MindRobots said its own hardware was not sufficient, while public cloud services could not guarantee continuous access to the resources it needed. The team therefore set up dedicated cloud infrastructure that runs around the clock. Over the course of the project, it provided more than 1,200 GPU-hours of continuous operation.
According to MindRobots CEO Alexey Kaigorodov, using cloud infrastructure allowed the team to gradually increase the complexity of the robot’s tasks while focusing on improving the technology.
Lowering the Cost of Entry Into Robotics
Cloud computing was once primarily associated with data storage and software development, but the growth of robotics is creating a new demand. Cloud computing is becoming a tool for training autonomous devices that need substantial computing power to model behavior, process data and train neural networks.
The project demonstrates how Russian cloud platforms can be used in high-tech fields such as robotics and industrial AI. This lowers the barrier to entry for smaller robotics companies: instead of buying expensive computing clusters, developers can rent GPU resources.
Building specialized cloud services for AI and robotics is a promising new area. As the number of autonomous-systems projects grows, demand will increase for infrastructure used to train large models, build digital twins, run computer-vision systems and control robots. Over time, Russian cloud companies could compete not only domestically but also in international markets.
A New Market Segment
The joint project between MindRobots and Turbo Cloud demonstrates the emergence of a new segment of Russia’s IT market: infrastructure for training physical AI systems. The future development of humanoid robots will depend in part on access to computing power, the quality of AI algorithms and the ability to conduct large-scale testing in real-world conditions. Cloud infrastructure is becoming a key part of that chain, allowing smaller technology companies to work on complex robotic systems without major capital investments.
Humanoid robot competitions are a battle of technologies that can accelerate advances in robotics. Successful robotics solutions can ultimately be deployed in commercial sectors including manufacturing, transportation, medicine and service robotics.
