By Erik Sherman|

March 14, 2025

An Amazon warehouse that opened in 2024 in Shreveport, Louisiana, has produced ten times the amount of robotic equipment as previous facilities, the Financial Times reported. The company estimates that the increased automation led to a 25% cut to fulfillment costs.

This isn’t an overnight success, but a decade-long investment of resources into a vision of operations that, as analysts at Morgan Stanley have estimated, could generate $10 billion in annual savings.

The tech giant acquired robotics startup Kiva Systems in 2012, rebranding it Amazon Robotics. Since then, the group reportedly has put 750,000 mobile devices and tens of thousands robotic arms and autonomous systems into use.

In 2019, Amazon was building a $40 million robotics innovation hub. By 2020, new fulfillment centers with employees working alongside specialty robots were already under construction. Then, in 2022, Amazon said it planned to invest $1 billion into supply chain, fulfillment, and logistics innovation.

The company plans to invest $100 billion in capital expenditure this year. Robotics and AI investments are part of the spend.

The need for time and effort is something Amazon has found necessary. “As a builder, it’s hard to wait for these building blocks to be built versus just combining a bunch of components together to solve a specific problem,” CEO Andy Jassy wrote in the 2023 annual report. “The latter can be faster, but almost always slows you down in the future. We’ve seen this temptation in our robotics efforts in our fulfillment network.”

He noted dozens of processes they wanted to automate, with some of the biggest opportunities in storage automation, manipulation, sortation, mobility of large cages across long distances, and automatic identification of items.

The robots the FT mentioned were all specialty devices. This includes flat rectangular units that send packages down chutes for human employees to pack into trucks or the robotic arms that hand the packages to the transportation units. The descriptions sound increasingly like human workers act as exception processors, handling the non-routine situations that robots can’t easily recognize or efficiently respond to.

That is why people still pack trucks. Automation is too slow and ineffective when sorting through packages of many sizes and shapes and quickly fitting them all into the 3D space of a truck.

However, finding the proper amount of automation that matches the operational scale of a warehouse, adds to efficiency and not just to costs but will need a lot more time than has already passed.