Amazon Blue Jay multi-arm robotics system handles 75% of item types while collapsing three workstations into one

Front-line employees at Amazon's fulfillment facilities were performing physically demanding, repetitive tasks — reaching and lifting across three separate robotic stations dedicated to picking, stowing, and consolidating items. These three stations took up more floor space than a single workspace would. Developing earlier robotics systems had historically required three or more years per system.

Amazon developed Blue Jay, a next-generation multi-arm robotics system that performs picking, stowing, and consolidation simultaneously within a single workspace. Its accelerated development relied on advances in AI and digital twin simulation — physics-based virtual environments that allowed engineers to iterate on dozens of prototypes. The team combined this with the AI, data, and learned experiences of Amazon's current robot fleet, and Blue Jay builds on recent advances like Vulcan and DeepFleet. This simulation-first approach compressed what had historically been a 3+ year development cycle into just over one year. Blue Jay was then tested in production at a fulfillment facility in South Carolina, and was planned to become a core technology at Amazon's Same-Day sites over time.

Blue Jay was tested in production at a facility in South Carolina:

• 75% item type coverage: Blue Jay can handle approximately three-quarters of all item types stored across Amazon sites
• 3-to-1 workstation consolidation: Three discrete robotic stations (pick, stow, consolidate) collapsed into a single streamlined workspace, directly reducing floor footprint
• Development cycle: A process formerly requiring 3+ years completed in just over one year through simulation-driven prototyping

Amazon says Blue Jay helps keep employees working in their ergonomic power zone, reducing repetitive reaching and lifting. As of February 2026, Amazon is no longer using Blue Jay in operations, but the underlying technology used to develop the program will continue to support employees across its network.

• Digital twin simulation can compress years of trial-and-error into months of development, letting engineers iterate on prototypes virtually.
• Consolidating multiple robotic functions into a single workspace can deliver greater efficiency in less physical space.
• Building on the AI, data, and learned experiences of an existing robot fleet can help new systems be developed smarter and more quickly.
• Even when a specific system is retired, the underlying technology used to develop it can continue to support employees across the network.