Russia's Sberbank Unveils Green Humanoid, Bolsters Tech Sovereignty

Sovereignty Leap Unveiled

Sberbank, Russia's dominant tech powerhouse, debuted its first proprietary humanoid robot named Green at the AI Journey 2025 conference in Moscow, positioning it as a cornerstone of national technological independence. The bipedal machine, powered by Sber's GigaChat large language model, performed a dance routine before President Vladimir Putin, introducing itself as the inaugural Russian humanoid with embedded AI. This launch underscores Moscow's drive to cultivate domestic robotics amid Western sanctions limiting access to advanced chips, transforming Sber from banking giant to AI innovator. CEO German Gref hailed it as the dawn of "Physical AI," blending cognition with real-world action to redefine automation sovereignty.

GigaChat Brain Awakens

Green stands out through its seamless fusion of GigaChat's neural prowess with physical form, enabling real-time audiovisual processing, natural conversations, and voice-activated tasks. Unlike screen-bound AIs, this embodiment lets the robot navigate unknown spaces autonomously while sorting and packing items in chaotic settings. Sber emphasizes cognitive depth over mere movement, allowing Green to interpret commands intuitively and adapt on the fly. This integration promises to evolve with GigaChat upgrades, expanding from basic interactions to complex problem-solving in dynamic environments, marking a bold step in embodied intelligence.

Sensor Symphony Stabilizes

Engineers at Sber crafted Green with a robust sensor suite, including ten vision processors for environmental mapping, plus inertial, force, and dynamic sensors for precise equilibrium during motion. These advancements ensure safe, stable bipedal walking and object manipulation, even in unstructured warehouses. By prioritizing force feedback and collision awareness, the robot minimizes risks in human-shared spaces, a critical edge for practical deployment. This hardware-software harmony elevates Green beyond gimmicks, fostering reliable performance that supports Russia's push for self-reliant AI hardware.

Deployment Horizons Expand

Pilot programs loom for Green in logistics, manufacturing, retail, and food service, where it could streamline inventory handling or assist remote inspections. Sber envisions fleets tackling infrastructure checks and supply chain bottlenecks, leveraging its autonomy to cut downtime in sanction-hit industries. Early demos spotlight packing efficiency in messy storerooms, hinting at scalable transformations for Russian enterprises. As GigaChat matures, these applications will deepen, potentially spawning task-specific variants while retaining humanoid versatility for human-centric roles.

Dexterity Drivers Detailed

Green's bipedal mobility thrives on RGB cameras, depth sensing, LiDAR, IMU, force/torque detectors, gyroscope, accelerometer, and joint encoders, empowering fluid walking at 1.5-3 m/s and dexterous interactions in crowded areas. Visual SLAM and LiDAR mapping, paired with balance aids, enable seamless navigation through human workspaces, facilitating extended collaboration sessions backed by a 3-5 year battery endurance. Safety layers like force limits, collision alerts, emergency halts, and redundant checks ensure trustworthy partnerships, while its Linux-ROS 2-Python backbone supports custom skills for tool handling, sensor swaps, or industrial gear in manufacturing and remote ops.

Rivals Edge Examined