Seven hard-tech companies showed their achievements in tactile perception, autonomous driving navigation, etc., so demonstrating the landing potential of "AI + robotics" integrated innovation. Google Cloud shared the application breakthroughs of Gemini multimodal technology in the robotics field at the Shenzhen Convention and Exhibition Center on April 25, 2025.
Gemini Technology Breakthrough: Smarter Robots
Google Cloud experts demonstrated how the Gemini model enables robots to conduct real-time environmental reasoning through multimodal understanding (language, vision, 3D sensing).
Embodied Intelligence Startups: Focus on Scenarios
Most robots now, Kunlun Capital observed, are like "top students in only one subject." It urged businesses to follow a "seizing interim opportunities" approach, first implementing technologies in particular contexts like retail and warehouse. Seven companies-including 9Tian Chuangzhi (which provides a low-cost navigation system with a 5 cm error margin) and Paxini Perception-which creates bionic tactile sensors to overcome the challenge of functioning in dark environments-presented at the event. Covering fields including humanoids and autonomous cars, all of them are grounded on NVIDIA GPU/Omniverse technologies.
Ecosystem Empowerment: Big Players Support Startups
NVIDIA provides hardware discounts (RTX 5880 Ada at 35% off), Omniverse simulation tools (cutting R&D cycles by 50%) and international exposure (like a direct line to GTC 2026) for startups. Backed by this ecosystem, Paxini Perception and other companies have entered the Apple supply chain for testing and are expected to see significant revenue growth in 2025.
With the successful conclusion of the Shenzhen show, the next stop will be in Shanghai.