Robotic, AI-Powered Warehouse Stacker IntroducedRobotic, AI-Powered Warehouse Stacker Introduced

AI-powered robotic sorting and stacking technology provider Ambi Robotics has launched a new robotic system that automates the stacking of items onto pallets or into containers with maximum density.

The company said its new AmbiStack robot features advanced AI technology, which allows it to stack items without prior knowledge of size, position or appearance. Using reinforcement learning, it can reason about the most efficient stacking of such items, adapting to real-world conditions on-the-fly to maximize the use of pallets and containers. 

The robot system uses an AI vision system based on foundation models and knowledge from over 200,000 operational hours of high-fidelity focused data, the company said. Using these tools, it analyzes, tracks and picks each item while performing quality control checks. The robot also has an AI planning system built on simulation-to-reality technology that does not need real world data collection, so can be rapidly put into production lines. 

The configurable system means warehouse operators could use the robot for multiple use cases, such as building inbound pallets from floor-loaded trucks, assembling outbound pallets for retail distribution or stacking walled containers for parcel transport, the company added. 

Related:Robots in Warehouses AI-Trained for Faster Sorting

For example, it could stack items in various containers such as pallets, the floor or walled containers and pick and place multiple cases simultaneously.

The system could address a critical demand for efficiency and scalability in supply chain operations by rapidly and reliably stacking high-density mixed SKUs, the company said. This could help reduce downstream transportation costs for warehouse operations across manufacturing, distribution and ecommerce sectors, it added. 

“Logistics companies are under continuous pressure to deliver items faster and for lower cost, which puts a lot of strain on operations to maintain reliability while rapidly adapting for the future,” said CEO Jim Liefer.

It could also alleviate the burden of repetitive, injury-prone motions for workers, enabling them to transition into higher-value roles as robot operators, assisting the robots when required, the company said. 

The AmbiStack robot adds stacking as a fundamental new robot skill into AmbiOS, the company’s AI operating system, which leverages proprietary simulation-to-reality technology called Sim2Real and the latest AI foundation models for high-speed robotic systems.

It’s an evolution of Ambi Robotics AI-powered robotic innovations, said company co-founder and chief technology officer Jeff Mahler, building on its technology in physical AI. 

“We’ve initiated the AI flywheel for stacking with Sim2Real reinforcement learning, and performance will only improve as robots collect data from production environments.”

He added that the company’s other robotics technology system, AmbiSort A-Series, has improved over 55% since launching in 2023. 

AmbiStack pre-sales have been strong and first deployments to select customers will begin mid-year and continue throughout the fall, the company said. It expects its manufacturing capability could fulfil growing demand from queued customers by early 2026.