The quiet revolution of robot warehouses

Amazon has 750,000 robots in its warehouses.

Seven hundred and fifty thousand. Working right now. While you read this.

They’re not humanoid. They don’t walk. They don’t talk. They’re orange platforms about the size of a coffee table that slide under shelving units, lift them, and carry them to human pickers. Originally made by Kiva Systems (which Amazon acquired in 2012 for $775 million), these robots are the most successful deployment of robotics in human history.

And nobody writes about them. Because they’re boring.

The invisible revolution

The flashy robotics stories are always about humanoid robots. Atlas doing parkour. Optimus walking on stage. Figure 01 taking its first steps. These make great videos. They generate excitement, discussion, Twitter threads.

The orange shelving robots generate packages. Millions of packages. Every day. They’re not exciting. They’re useful. And the gap between exciting and useful is where the real robotics revolution is happening.

Amazon’s warehouse robots have been operating for over a decade. In that time, they’ve reduced the “click to ship” time at fulfillment centers from over an hour to about 15 minutes. They’ve made it possible for warehouses to store 40% more inventory in the same space (the aisles can be narrower because humans don’t need to walk through them). They work 24/7 without breaks, without injuries, without complaints.

750,000 of them. The number is so large it becomes abstract. Let me make it concrete: if Amazon’s robots were a city, they’d be the third largest city in the United States, between Chicago and Houston.

The newer robots

Amazon has expanded beyond the orange platforms. Sparrow is a robotic arm that can identify, select, and handle individual products from a bin. It uses computer vision and machine learning to pick items of different shapes, sizes, and materials.

Locus Robotics makes autonomous mobile robots that work alongside human pickers, carrying bins from station to station. They’re deployed in non-Amazon warehouses across the logistics industry.

The humanoid robot startups might get the attention, but the logistics robots already have the market.

The pattern

I keep seeing the same pattern: the most impactful technology is the technology you stop noticing. The best robots aren’t the ones that look impressive in demos. They’re the ones that become invisible through reliability.

Elevators are robots. ATMs are robots. Dishwashers are robots. We don’t call them that because they’ve been around long enough that we’ve absorbed them into the background of daily life.

Amazon’s orange platforms are undergoing the same absorption. They were novel in 2012. By 2023, they’re infrastructure. Invisible. Boring. And running the backend of the world’s largest logistics network.

I think humanoid robots will follow this path. Not soon. The technology isn’t there yet. But eventually. The humanoid robot of 2040 won’t be a spectacle on a stage. It’ll be the thing that stocks the shelves at a grocery store after closing time, and nobody will think twice about it.

The quiet revolution. 750,000 strong and counting.

Related thinking: