AMR vs AGV: How to Decide What Your Facility Actually Needs
Search “AMR vs AGV” and you will find the same article a hundred times: two columns, a list of features, and a verdict that lands somewhere between “it depends” and “AMRs are the future”. None of that helps you decide. What you actually need is a way to look at your building, your flow and your budget, and come out with an answer. That is what this page does.
The difference is navigation — everything else follows
An AGV — automated guided vehicle — follows a path that is defined outside the vehicle. An AMR — autonomous mobile robot — carries its own map and works out the route as it goes. Every other difference you will read about is a downstream consequence of that one distinction.
AGV: the path is part of your building
A conventional AGV navigates by following infrastructure: magnetic tape laid in the floor, reflective markers, QR codes on the ground, or a wire buried under the surface. The vehicle detects the guide and follows it. This is reliable and extremely repeatable, and it is why AGVs have been running in factories for decades.
The trade-off is that the path is physical. To change a route, you change the building. To add a station, you lay more tape. And if a pallet is left in the aisle, the AGV stops and waits — its instruction is to follow the line, and the line is blocked.
AMR: the path is in the software
An AMR builds a map with onboard sensors — LiDAR, usually with cameras as well — and localises itself within that map. It plans a route, and re-plans when something is in the way. Blocked aisle: it goes around, or waits and re-routes, depending on how the job is configured. Changing a route is a configuration change, not a construction project.
Side by side
| AGV | AMR | |
|---|---|---|
| Navigation | Follows fixed infrastructure (tape, markers, wire, QR) | Builds and uses its own map; plans routes dynamically |
| Infrastructure required | Yes — installed in the floor or along the route | None fixed |
| Obstacle handling | Typically stops and waits | Routes around, or re-plans |
| Changing a route | Physical work on site | Software change |
| Working around people | Limited; usually needs segregated areas | Designed for shared environments |
| Typical best fit | Fixed, repetitive, high-volume flows | Changing layouts and mixed traffic |
The unit price is not the comparison
An AGV usually costs less per unit than an AMR. That gap is real, and for the right application the cheaper choice is also the correct one. But the purchase price is not the number you are comparing.
An AGV deployment carries a second cost that never appears on the machine quotation: the infrastructure, plus the downtime and disruption of installing it. That cost is paid once, and the constraint it creates is permanent — every future layout change carries a bill you have already paid one version of.
An AMR deployment front-loads cost into the vehicle and the mapping, then keeps change cheap. If your layout is stable for the next decade, that flexibility is worth little to you, and you should not pay for it. If you reconfigure racking, add lines or move pick faces every couple of years, it is worth a great deal.
When an AGV is still the right answer
- The flow is genuinely fixed — same route, same stops, same volume, year after year.
- Throughput and repeatability matter more than flexibility.
- The environment can be kept clear of people and unexpected obstacles.
- The layout is not going to change, and everyone involved knows it.
In a purpose-built facility with a stable process, a fixed-path AGV can be the most economical and the most predictable choice available. Buying flexibility you will never use is not a virtue.
When an AMR wins
- The layout changes, or is expected to.
- Robots and people share the same floor.
- Obstacles on the route are normal rather than exceptional — pallets, carts, shrink-wrapped stock.
- You want to start with one unit and add more as volume grows, without re-engineering the site.
- Several zones are served from one fleet, with routes that overlap and shift.
These are the conditions most distribution centres actually operate under — which is why the AMR category has grown so quickly in buildings that were never designed around automation in the first place.
Four questions that settle it
Answer these in order. The first “no” usually decides.
- Will this route still be the same in three years? Yes leans AGV. No leans AMR.
- Can the area be kept clear of people and unexpected obstacles? Yes leans AGV. No means AMR.
- What does it cost you to disrupt operations and install infrastructure? If the answer is “a great deal”, the AGV’s lower unit price is worth less than it appears.
- Are you starting with one route and planning to grow? Growth without re-engineering favours AMRs.
If two or more answers point the same way, you have your decision. If they split, the practical route is a mixed deployment: AGVs on a fixed high-volume backbone, AMRs on the flexible work around it — mixed routes, changing destinations, and shared floor space.
Frequently asked questions
Is an AMR just an AGV without tape?
No. Removing the tape is a consequence of the navigation method, not the definition of it. The meaningful difference is that an AMR maintains its own understanding of position and route, which is what allows it to re-plan when the situation changes.
Do AMRs need Wi-Fi to work?
A fleet normally uses a network so units can be coordinated and monitored, but an individual robot’s navigation does not depend on continuous connectivity to find its way. Network requirements are worth confirming early in any deployment plan.
Are AMRs safe to run around people?
AMRs built for shared environments carry sensor and safety systems designed for exactly that — obstacle detection, speed control and protective stops. The specific standard that applies depends on your market and application, and it belongs in the assessment rather than in an assumption.
Which is cheaper to run over five years?
It depends on how often your layout changes and how much disruption costs you. Unit price is only one input. A fixed process in a stable building can favour AGV economics; a changing layout in a shared space usually favours AMR.
Next step
If you are weighing the two for a specific building, send us the route, the payload and the shift pattern. We will tell you which approach fits — and whether a mixed deployment makes more sense than either one alone.
Talk to us about your application, or see the AMR range.