Is physical world AI the future of autonomous machines? - The Robot Report

The article discusses the emerging role of physical world AI—cloud-based systems integrated with AI models that create ultra high-precision, spatially aware representations of the physical environment—in advancing autonomous machines such as cars, drones, and tractors. While companies like Waymo have developed sophisticated onboard AI and navigation hardware, the article argues that relying solely on onboard compute is insufficient for widespread autonomous machine deployment. Instead, leveraging cloud-based spatial intelligence can enhance route optimization and hazard detection by providing machines with detailed, real-time environmental context beyond their immediate sensor inputs. Currently, most AI in autonomous machines operates locally on the edge, lacking awareness of the broader physical landscape. However, abundant data from satellites, drones, and other sources can feed cloud systems that process complex spatial information—such as vectors representing terrain and obstacles—making AI models more capable of understanding and navigating the environment. This spatial intelligence cloud approach, pursued by companies like Wherobots, can improve autonomous vehicle performance in challenging scenarios like rural deliveries or complex urban settings