An AGV navigates by following predefined physical or virtual guides (magnetic tape, wire, QR markers, or wall-mounted laser targets), while an AMR uses onboard LiDAR, 3D cameras, and SLAM to build a map and plan its own path in real time [S1][S2][S4].
Both belong to the same "driverless industrial truck" family under ISO 3691-4, and fleets can mix them since VDA 5050 v3.0 was released in March 2026 [S3]. For specifiers, the decision is not which label wins but which point on the navigation-and-autonomy spectrum matches the payload, takt time, and route stability of the work.
Navigation Stack: External Guidance vs Onboard SLAM
AGVs depend on external guidance such as magnetic strips, inductive wires, colored tape, or laser targets mounted on walls; route changes require physical re-installation of the guide medium, with corresponding recommissioning cost [S1][S2].
AMRs use 3D sensor fusion (LiDAR, vision, 3D perception) and SLAM to localize and perceive the world without modifying the facility, then reassess multiple candidate routes and pick the most efficient one on every dispatch [S1][S4]. The Vecna Robotics comparison frames this as AMR = path-planning robot versus AGV = path-following robot, a distinction that has held since at least 2021 industry literature [S6].
Obstacle Response: Stop-and-Wait vs Reroute
AGVs typically detect an obstruction, stop, and wait for it to clear; in many installations a safety stop requires operator intervention or a manual reset, which the operator guidance from S1 explicitly flags as a hidden operating cost [S1][S4].
AMRs detect the same obstacle, plot a safe bypass, and keep moving, which is why they are preferred for mixed-traffic warehouse aisles, picking/putaway, replenishment, and sortation workflows where human and pallet traffic shift minute by minute [S1][S2]. S2 states that "they can detect obstacles and safely maneuver around them, choosing the best alternative route," which preserves material flow that an AGV would otherwise stall [S2].
Payload and Docking Tolerance
Invio AGVs are documented at payloads up to 70,000 lb for heavy assembly-line work, where takt-critical, repeatable moves are the requirement [S3]. AGVs dock to a fixed coordinate, so the payload staging tolerance is tight; out-of-spec staging typically triggers a fault and operator intervention [S1].
AMRs use onboard vision to identify the right pallet or cart and dynamically adjust pickup even when the payload is misaligned, which is the practical reason most picking-to-tote and bin-handling cells now default to AMRs rather than AGVs [S1]. Heavy manufacturing cells (multi-ton fabrication, large body-in-white, engine machining) still tend to stay on AGVs because the cost of a misalignment is a crashed fixture, not a missed pick.
Decision Matrix: AGV vs AMR Across 4 Spec Criteria
Use this matrix when writing the spec:
(1) Route stability. Stable, unchanging routes favor AGV; layouts that change monthly or seasonally favor AMR [S5]. S5 puts it bluntly: "AGVs deliver predictability and throughput on stable routes; AMRs deliver flexibility in dynamic environments."
(2) Payload and precision. Multi-ton loads, repeatable docking under 10 mm, and takt-critical assembly favor AGV; sub-ton totes and bins with loose staging tolerance favor AMR [S1][S3].
(3) Infrastructure cost. AGVs need magnetic tape, wire, or reflectors, which adds recommissioning cost on every layout change; AMRs need a one-time map and a Wi-Fi or 5G link, with no floor modifications [S1][S2].
(4) Mixed-fleet future-proofing. VDA 5050 v3.0 (March 2026) lets one fleet manager orchestrate AGVs and AMRs side by side, so either choice is no longer a 10-year lock-in [S3].
Safety and Standards: One Family, One Rulebook
ISO 3691-4 covers both AGVs and AMRs as "driverless industrial trucks"; in North America the parallel references are ANSI/ITSDF B56.5 and ANSI/RIA R15.08, and MHI now groups its members under the neutral Mobile Automation Group label [S3]. Picking the right standard is now a question of where the vehicle runs, not which acronym is on the nameplate.
For an in-depth look at the sensor and LiDAR layer behind AMR navigation, the AGV vs AMR navigation primer walks through SLAM, fiducial markers, and safety-scanner zoning in more detail, and the broader mobile robot landscape maps how AMR fleets hand off to conveyors and WMS/WCS.
Cost, Lifecycle, and When the Acronym Stops Mattering
AGVs usually have a lower per-unit sticker price, but the tape, wire, reflectors, recommissioning labor, and "rescue FTEs" for safety stops shift the total-cost-of-ownership curve; the Invio Automation writeup notes that even infrastructure-free AGVs commonly need multiple dedicated staff to manage resets [S1][S3]. AMRs trade higher unit cost for lower install cost and faster layout changeovers, which is why AMR unit volume is climbing fastest in lighter, high-variability applications while AGVs continue to anchor heavy, high-precision work [S3].
For battery sizing on a two- or three-shift AGV/AMR fleet, the AGV battery sizing walkthrough covers duty cycle, opportunity-charge headroom, and end-of-shift derating; for the energy side, opportunity charging vs battery swap is the natural follow-on decision once the navigation choice is locked.
Selection Recommendation
Specify an AGV when the route is fixed for the life of the cell, the payload exceeds roughly 2,000 kg, and docking repeatability must hold under 10 mm; specify an AMR when routes change weekly, payloads stay under 1,500 kg, and the workspace shares lanes with people, forklifts, or other robots [S1][S3][S5].
Watch for two trackable signals over the next two quarters: VDA 5050 v3.0-compliant fleet-manager releases from at least two of the major WMS/WCS vendors, and a published case study of a mixed AGV-plus-AMR cell governed by a single fleet manager on a VDA 5050 v3.0 interface. Either would confirm that the AGV-versus-AMR label is becoming a navigation-style choice inside one fleet, rather than a fleet-versus-fleet decision.
For the relevant spec sheets and selection criteria, see fixed gas detector.