The logistics innovation of Coupang and Amazon is beginning with fulfillment logistics automation technology. A fulfillment center is a facility where online sellers store products sold on a platform and process orders, and it has become established as a logistics hub. As online distribution led by Coupang, Amazon, and other companies around the world has grown rapidly, the efficiency of fulfillment centers, which handle the entire process from inventory management to picking, packaging, and shipping, has become a key factor determining corporate competitiveness. However, within complex logistics processes, “irregular boxes” and “irregular inventory” still remain challenges for automation.
[Fig] Cases of Fulfillment Automation Technology Introduced by Amazon and Coupang
※Source: Amazon and Coupang Websites
Existing fulfillment automation systems were optimized for standardized workflows. Various automation technologies, including conveyor belts, ASRSs (Automated Storage and Retrieval Systems), and robotic pickers (robots that pick up objects), have been introduced, but most automation technologies operate based on standardized specifications and fixed layouts. However, at actual sites, boxes of various sizes and shapes are randomly stacked or tilted, and barcodes are frequently obscured or labels are damaged. As a result, processes such as inbound and outbound handling, rearrangement, and palletizing still depend on workers and act as bottlenecks that reduce the efficiency of automation.
[Fig] Stacking Problems of Existing Fulfillment Automation Technologies
※Source: Unsplash
The technology that has recently attracted attention as a way to overcome this bottleneck is the “Random Box Stack robot (RBS).” An RBS (Random Box Stack) system combines an economical 3D camera with an AI vision algorithm to recognize boxes of various sizes and shapes in real time and allow a robot to automatically stack them through palletizing or unload them through depalletizing. In such fulfillment environments, the robot’s performance in handling objects, making decisions, and responding to loads becomes the most important source of competitiveness and differentiation.
RBS technology is becoming a core growth strategy for global logistics robot companies, particularly those based in the United States. Ambi Robotics of the United States developed AmbiStack, an AI vision-based automated box-stacking solution. The system is designed to allow robots to densely stack logistics boxes of various shapes, and Sim2Real-based reinforcement learning technology is applied to secure real-time adaptability and accuracy. Ambi Robotics is currently collaborating with major North American distributors and expanding its commercialization track record across various fulfillment environments. Dexterity AI of the United States is also attracting attention for developing RBS robot technology specialized for trailer loading. The DexR system uses an AI-based dual robotic-arm structure to rapidly recognize boxes and automatically load them inside trucks, and its field applicability and commercialization capabilities have been verified through cooperation with FedEx. FedEx is conducting the world’s first demonstration by applying automation technology to rear-end logistics operations that were previously centered on manual work.
[Fig] RBS Systems of Ambi Robotics and Dexterity AI
※Source: Ambi Robotics and Dexterity AI Websites
In Korea, Euler Robotics Co., Ltd. (CEO: Bong-seong Kim) is also challenging global competition through RBS technology. Through its independently developed RBS system, Euler Robotics developed a robotic solution capable of making decisions within 0.5 seconds, stacking up to a height of 1.8 meters, and handling loads of up to 20 kilograms. The system can perform both palletizing and depalletizing, giving it high versatility, and it can be efficiently applied to small and medium-sized fulfillment centers through API integration and a modular design. Euler Robotics recently entered into a strategic partnership with TDS in the United States, establishing a bridgehead for entry into the North American market. It also successfully secured Seed investment, with the amount undisclosed, and is preparing to advance its technology and enter overseas markets.
[Fig] Overview and Configuration of Euler Robotics Co., Ltd.’s RBS System
※Source: Euler Robotics Co., Ltd. Website
When developing RBS technology, a strategic approach is required to protect intellectual property rights. First, a strong defensive barrier should be established by subdividing core technologies and filing divisional and continuation applications. Patent applications should be filed in detail for each technological element, including AI vision algorithms, real-time decision-making logic, load-distribution control, and robotic-arm design. After securing an initial foundational patent, a strategy is needed to vertically expand the portfolio through continuation and divisional applications. In addition, when designing the scope of patent rights, it is necessary to go beyond simple device technology and use a “configuration-operation-effect” structure to reflect performance-based effects, such as the ability to respond to irregular environments or improvements in work efficiency, in the claims. AI processing flows, such as recognizing exceptional situations and generating alternative routes, may also be subject to protection.
Meanwhile, the published patents of competitors such as Ambi Robotics and Dexterity AI should be thoroughly analyzed to secure FTO (Freedom to Operate), while a blocking patent strategy should be established to prevent design-around attempts. In addition, where the technology has a SaaS-based structure capable of integration with external ERP and WMS systems, it is important to pursue software-based patents and copyright registration for the relevant module-integration technologies.
Patent design that considers future market expandability is also important. Considering that RBS technology may later expand into port unloading, factory material stacking, construction logistics, and other fields, general-purpose patents assuming platformization and multifunctionality should be designed from the initial stage in order to gain an advantage in market entry.
The BLT Research Center stated, “RBS-related technologies are expected to create new opportunities for companies.” It added, “Securing rights to various RBS-related technologies is expected to be an effective strategy.”
As of 2025, BLT Patent & Law Firm has been selected by more than 2,000 innovative startups as a partner, supporting corporate growth and success through IP acquisition, response strategy development, investment attraction, and IP-based business support such as technology-specialized listings.
#Fulfillment #RBS #3DVisionRobot #RandomBoxStack #LogisticsAutomation #Palletizing #IrregularPalletizing #EulerRobotics #AmbiRobotics #DexterityAI #Coupang #Amazon
The logistics innovation of Coupang and Amazon is beginning with fulfillment logistics automation technology. A fulfillment center is a facility where online sellers store products sold on a platform and process orders, and it has become established as a logistics hub. As online distribution led by Coupang, Amazon, and other companies around the world has grown rapidly, the efficiency of fulfillment centers, which handle the entire process from inventory management to picking, packaging, and shipping, has become a key factor determining corporate competitiveness. However, within complex logistics processes, “irregular boxes” and “irregular inventory” still remain challenges for automation.
[Fig] Cases of Fulfillment Automation Technology Introduced by Amazon and Coupang
※Source: Amazon and Coupang Websites
Existing fulfillment automation systems were optimized for standardized workflows. Various automation technologies, including conveyor belts, ASRSs (Automated Storage and Retrieval Systems), and robotic pickers (robots that pick up objects), have been introduced, but most automation technologies operate based on standardized specifications and fixed layouts. However, at actual sites, boxes of various sizes and shapes are randomly stacked or tilted, and barcodes are frequently obscured or labels are damaged. As a result, processes such as inbound and outbound handling, rearrangement, and palletizing still depend on workers and act as bottlenecks that reduce the efficiency of automation.
[Fig] Stacking Problems of Existing Fulfillment Automation Technologies
※Source: Unsplash
The technology that has recently attracted attention as a way to overcome this bottleneck is the “Random Box Stack robot (RBS).” An RBS (Random Box Stack) system combines an economical 3D camera with an AI vision algorithm to recognize boxes of various sizes and shapes in real time and allow a robot to automatically stack them through palletizing or unload them through depalletizing. In such fulfillment environments, the robot’s performance in handling objects, making decisions, and responding to loads becomes the most important source of competitiveness and differentiation.
RBS technology is becoming a core growth strategy for global logistics robot companies, particularly those based in the United States. Ambi Robotics of the United States developed AmbiStack, an AI vision-based automated box-stacking solution. The system is designed to allow robots to densely stack logistics boxes of various shapes, and Sim2Real-based reinforcement learning technology is applied to secure real-time adaptability and accuracy. Ambi Robotics is currently collaborating with major North American distributors and expanding its commercialization track record across various fulfillment environments. Dexterity AI of the United States is also attracting attention for developing RBS robot technology specialized for trailer loading. The DexR system uses an AI-based dual robotic-arm structure to rapidly recognize boxes and automatically load them inside trucks, and its field applicability and commercialization capabilities have been verified through cooperation with FedEx. FedEx is conducting the world’s first demonstration by applying automation technology to rear-end logistics operations that were previously centered on manual work.
[Fig] RBS Systems of Ambi Robotics and Dexterity AI
※Source: Ambi Robotics and Dexterity AI Websites
In Korea, Euler Robotics Co., Ltd. (CEO: Bong-seong Kim) is also challenging global competition through RBS technology. Through its independently developed RBS system, Euler Robotics developed a robotic solution capable of making decisions within 0.5 seconds, stacking up to a height of 1.8 meters, and handling loads of up to 20 kilograms. The system can perform both palletizing and depalletizing, giving it high versatility, and it can be efficiently applied to small and medium-sized fulfillment centers through API integration and a modular design. Euler Robotics recently entered into a strategic partnership with TDS in the United States, establishing a bridgehead for entry into the North American market. It also successfully secured Seed investment, with the amount undisclosed, and is preparing to advance its technology and enter overseas markets.
[Fig] Overview and Configuration of Euler Robotics Co., Ltd.’s RBS System
※Source: Euler Robotics Co., Ltd. Website
When developing RBS technology, a strategic approach is required to protect intellectual property rights. First, a strong defensive barrier should be established by subdividing core technologies and filing divisional and continuation applications. Patent applications should be filed in detail for each technological element, including AI vision algorithms, real-time decision-making logic, load-distribution control, and robotic-arm design. After securing an initial foundational patent, a strategy is needed to vertically expand the portfolio through continuation and divisional applications. In addition, when designing the scope of patent rights, it is necessary to go beyond simple device technology and use a “configuration-operation-effect” structure to reflect performance-based effects, such as the ability to respond to irregular environments or improvements in work efficiency, in the claims. AI processing flows, such as recognizing exceptional situations and generating alternative routes, may also be subject to protection.
Meanwhile, the published patents of competitors such as Ambi Robotics and Dexterity AI should be thoroughly analyzed to secure FTO (Freedom to Operate), while a blocking patent strategy should be established to prevent design-around attempts. In addition, where the technology has a SaaS-based structure capable of integration with external ERP and WMS systems, it is important to pursue software-based patents and copyright registration for the relevant module-integration technologies.
Patent design that considers future market expandability is also important. Considering that RBS technology may later expand into port unloading, factory material stacking, construction logistics, and other fields, general-purpose patents assuming platformization and multifunctionality should be designed from the initial stage in order to gain an advantage in market entry.
The BLT Research Center stated, “RBS-related technologies are expected to create new opportunities for companies.” It added, “Securing rights to various RBS-related technologies is expected to be an effective strategy.”
As of 2025, BLT Patent & Law Firm has been selected by more than 2,000 innovative startups as a partner, supporting corporate growth and success through IP acquisition, response strategy development, investment attraction, and IP-based business support such as technology-specialized listings.
#Fulfillment #RBS #3DVisionRobot #RandomBoxStack #LogisticsAutomation #Palletizing #IrregularPalletizing #EulerRobotics #AmbiRobotics #DexterityAI #Coupang #Amazon