Intelligence.
In motion.

The research behind our homepage image starts with a remarkable possibility: living neural networks could help inform faster, lower-power computing.

A Chinese Academy of Sciences team studied cultured neuronal networks using a microelectrode array. According to its May 2026 report, predictable electrical stimulation increased neural communication speed by 1.79 times without increasing energy consumption.

The work explores a route toward biohybrid processors and brain-inspired systems. It is an experimental result in biological networks, rather than a demonstrated replacement for a data center or a commercial AI chip.

Unitree G1 · Product platform and demonstrations

Learning to move.
Learning to handle.

Unitree’s G1 brings a humanoid form, depth sensing, 3D LiDAR, and articulated joints into a compact robotics platform. Its manufacturer describes imitation learning and reinforcement learning as part of the technology’s development.

The G1 EDU configurations support further development and optional dexterous hands. These platforms give researchers hardware on which to study balance, manipulation, and the connection between AI decisions and physical movement.

Unitree notes that some demonstrated functions remain under development. A movement demonstration is evidence of a specific capability; reliable operation across unfamiliar workplaces is a separate test.