🤖 About the project

Project "Robot." It aims to create a unique ecosystem of intelligent robots capable of performing a wide range of production and home tasks. We're planning. Purchase several prototype robots in different countriesTrain them with advanced artificial intelligence technologies and adapt them for real-world applications.

📦 What robots are we planning to buy

🇪🇺 NEURA 4NE1

Full-size cognitive humanoid robot developed in collaboration with Studio F.A. Porsche. The cost €98,000 per unit (up to 20 units).
Features: growth of 180 cm, lift to 110 kg, speed of 5 km / h, NVIDIA Thor T5000 processor. It is expected to go on sale in late 2026.

🇨🇳 Agibot A2/G2 (China)

At MWC 2026, Agibot presented a full range of robots for various tasks.
Available models:

• A2 Series Full-size humanoids for meeting guests and navigation
• G2 Series Wheeled robots for industrial tasks (precise assembly)
• D1 Four-legged robots for logistics and patrol

Conditions: Possible purchase or lease from €900/day through the RaaS platform.

🇪🇺 NEURA 4NE1 Mini

A compact version of a cognitive robot created as a “Western response” to affordable Chinese models. The cost €19,999 - €29,999 depending on the version.
Features: height 132 cm, weight 36 kg, 25 degrees of freedom, payload 3 kg, working time ~2.5 hours.
Version Pro It is equipped with 12-speed hands and access to a digital twin. It is expected to be on sale from April 2026.

🇨🌍0 UWORLD by UBTECH (China)

The world’s first full-size hyper-bionic humanoid robot focused on emotional interaction.
Features:

• Male version: height 183 cm, weight 42 kg
• Female version: height 168 cm, weight 35.2 kg
• 88 degrees of freedom, 2-4 hours of work

Price: Pre-order with a deposit of 3,000 yuan (~30,000 ). Delivery from September 2026.

🌍1 How we will train robots with AI

Robot training is a key part of our project. We will use the most advanced methods:

• Imitation learning (Imitation Learning): Robots will learn by observing human actions. This is the fastest way to teach a robot new skills. For example, an operator can once show the robot how to assemble a particular part, and the robot can repeat the action itself.
• Reinforcement Learning (Reinforcement Learning): Artificial intelligence will learn by trial and error, receiving a “reward” for correctly performed actions. This technology already allows groups of robots to work together.
• Tactile skills (Tactile Skills): We will use a breakthrough technique described in the journal Nature Machine Intelligence that allows robots to perform complex tactile tasks with almost 100% success, even when conditions change.
• Skill Generalization (Skill Generalization): A robot trained in a single task can adapt to similar tasks without reprogramming. For example, by learning to capture one type of detail, it can work with other shapes and sizes.

🌍3 Market Potential and Payback

The humanoid robot market is shifting from concepts to mass production. Our advantages:

• Affordable: There are models at the price of a smartphone (from 110,000 ый), which opens the mass market.
• Robot-as-a-service (RaaS) model: The possibility of renting from €900 per day allows you to test solutions without major investments.
• Payback: Reducing the time for adjustment from weeks to hours – saving hundreds of thousands of rubles at each workplace.

🌍4 Who We Need