Physical AI Powering the New Age of Industrial Operations 2025

2.2 Spotlight on the pioneers – transformation journeys of early adopters As robotics reshapes the industrial operations landscape, a pioneering group of companies is charting bold new territory and redefining what is possible. By integrating intelligent robotics across previously isolated functions, they are closing the last remaining gaps in automation with physical AI. These leaders are not merely adopting tools – they are reinventing workflows, unlocking new levels of flexibility and precision, and transforming how work is done on the factory floor and in large warehouses. Their transformation journeys offer critical insights into how frontier capabilities in physical AI can be scaled effectively, demonstrating both the organizational foundation and the technical enablers required to sustain innovation at scale. The following case studies spotlight how early adopters are navigating this transformation, offering tangible proof points for what the future of industrial operations looks like in action. CASE STUDY 1 E-commerce fulfilment Reinventing fulfilment through the world’s largest robotics operation Amazon operates more than 1 million robots across its operations network, making it the largest user of robotics globally. These robots are deployed across 300 fulfilment centres, working alongside employees to perform repetitive tasks such as sorting, lifting and transporting packages. The company’s progress in robotics reflects an ongoing journey of experimentation and innovation, focused on continuously improving both employees’ working conditions and the customer experience. Mission Over the past decade, Amazon has introduced a series of successive “unlocks” that applied physical AI across its fulfilment centres: –Mobile goods-to-person robots that deliver inventory directly to employees –Computer vision-based sortation systems to smooth inventory flow –Mechatronic packing lines engineered to minimize use of packing materials, supporting its sustainability goals –Robotic manipulators capable of grasping the majority of catalogue items While these solutions improved safety and productivity, they operated in isolation. The main challenge was integrating them to enable true end-to-end transformation. Innovation in actionIn response, Amazon redesigned its fulfilment centre operating system around predictive AI planning and system interoperability. This redesign is anchored by three cornerstone technologies connecting the entire fulfilment process – from inbound receiving to outbound loading – into a single, end-to-end flow: –Sequoia is an automated storage-and-retrieval system. –Sparrow, an articulated manipulator, uses advanced vision and generative AI-guided motion planning to identify, pick and place roughly 60% of the items in the company’s inventory, learning continuously from the industry-scale data generated every day. –Proteus, a collaborative autonomous mobile robot, maps open spaces in real time, interprets social cues and charts efficient paths alongside employees, moving pallets that once required fenced zones. Sequoia, Sparrow and Proteus (from left to right) at work alongside Amazon employees Physical AI: Powering the New Age of Industrial Operations 13