Welcome to this week's edition of Video Friday, where we curate the most compelling robot videos from across the globe. This week, we spotlight Google DeepMind's latest breakthrough—Gemini Robotics 2—alongside other standout clips featuring robotic shadow puppets and golden eagles engaging with a drone.
Featured Spotlight: Gemini Robotics 2
Google DeepMind has unveiled Gemini Robotics 2, a next-generation vision-language-action (VLA) model designed to bridge the gap between digital understanding and physical interaction. Building on the success of its predecessor, this 2026 iteration enhances embodied reasoning, enabling robots to perform complex manipulation tasks with greater precision and adaptability. The accompanying video showcases a humanoid robot and a robotic arm, both powered by Gemini Robotics 2, navigating a workshop environment with impressive dexterity—demonstrating everything from precise pick-and-place operations to adaptive tool use.
The model's improved perception and reasoning capabilities allow it to handle novel objects and dynamic scenarios, making it a significant step toward more general-purpose robots. This aligns with broader industry trends in 2026, where VLA models are becoming central to advancing autonomous systems.
More Robot Highlights
Beyond the DeepMind feature, this week's selection includes a playful demonstration of robot shadow puppets, blending traditional storytelling with modern robotics, and a striking aerial encounter where golden eagles inspect and interact with a drone—offering a glimpse into wildlife-technology interactions.
About the Author
Evan Ackerman is IEEE Spectrum's robotics editor, covering the latest innovations and breakthroughs in the field.