The teaser showed ET1 executing a series of tennis strokes, demonstrating high-speed whole-body coordination and dynamic movement. The demonstration builds on LATENT, an athletic learning framework co-developed by Galbot researchers earlier this year.
LATENT, short for Learning Athletic Humanoid Tennis Skills from Imperfect Human Motion Data, is designed to teach humanoids complex athletic movements using fragmented human motion-capture data rather than complete demonstrations. The framework was previously validated on Unitree’s G1, achieving a 96% forehand success rate in simulation and enabling multi-shot rallies with human players in real-world testing.
ET1 represents a departure from Galbot’s commercial G1 platform, which uses a wheeled base and dual arms for structured tasks across logistics and retail environments.
Galbot shipped approximately 900 commercial robots during the first half of 2026, placing it among China’s larger humanoid robotics vendors by unit shipments.
The company is also reportedly exploring a Hong Kong IPO at a valuation of up to $4 billion. Galbot has raised capital from investors including CATL as it continues to expand its robotics portfolio and commercial deployments.
