Autonomous mobile robots in the warehouse: running the economics
AMRs are no longer a future concept - they are a working tool. I look at when they pay off and when buying a robot turns out to be an expensive mistake.
Autonomous mobile robots for warehouses - AMRs - have stopped being a curiosity in 2019. Dozens of companies make them, prices have come down, and rental and robotics-as-a-service models have appeared. The question has shifted from "does this exist?" to "when does it make sense for us?"
I occasionally look at these projects alongside operations directors and owners, and I keep seeing the same mistake: the decision is made on the basis of an impressive demo rather than the economics of the specific warehouse. I want to work through how to think about this properly.
What AMRs can and cannot do
AMRs solve one specific task well: moving a load from point A to point B inside a space, dynamically avoiding obstacles. This is not the same category as fixed conveyors or robotic arms. AMRs are mobile, require no infrastructure changes (no special tracks or floor markings), and work alongside people.
What they cannot do on their own: pick an item from a shelf and place it. That is still done by a person or a separate manipulator. An AMR is a transport device, not a full replacement for a picker.
A typical scenario: a picker stands at a shelving unit, the robot arrives with a cart, the person transfers the item, the robot takes it to packing or dispatch. The person does not walk around the warehouse - the robot does.
When the economics work
The main variable for calculation is how much of the employee's time goes to walking versus productive work at the shelf.
In warehouses with large distances and a high picking pace, walking can account for 50-70% of working time. Here an AMR eliminates that component, and the productivity gain is significant. If you have 20 pickers and each one spends half a shift walking, that is the equivalent of 10 people who are not picking anything.
The second factor is flow consistency. AMRs work well under steady load. In warehouses with strong peaks - seasonal spikes, promotions - robots do not scale as flexibly as temporary staff. This changes the calculation.
The third factor is the cost of floor space. In cities with high commercial real estate prices, a compact warehouse with AMRs can be more cost-effective than a large warehouse with a large workforce.
When the economics do not work
In small warehouses with short routes, the gain from AMRs is minimal and the cost of the fleet and software is significant.
In warehouses with a chaotic SKU structure and constantly changing topology, robots need to be reconfigured frequently, which is expensive in time.
With non-standard loads - unusual containers, fragile items, varied weights - AMRs need special equipment or the whole logic breaks down.
Questions before the decision
Before talking to a vendor about a pilot, it is worth having answers to a few concrete questions:
- What is the actual share of walking in a picker's working time at your warehouse? Has this been measured?
- What is the current real cost of a labour hour, including all costs - taxes, recruitment, training, turnover?
- What does your load profile look like across the year - are there pronounced peaks?
- How stable is the warehouse topology and SKU structure?
- What payback period is acceptable for your business?
Good AMR vendors are willing to build a joint ROI calculation using your data. If a vendor avoids that conversation and suggests buying first and calculating later, that is a bad sign.