AGV vs AMR — independent comparison of when each wins.

AGVs (Automated Guided Vehicles) are fixed-path robots that follow markers — magnetic tape, embedded wires, optical lines, or fixed-position QR codes. They're deterministic, reliable, and proven at scale in high-volume repeatable tasks.

AMRs (Autonomous Mobile Robots) use onboard sensors (LiDAR, depth cameras, IMU) to build a map of the environment and navigate dynamically — they replan around obstacles, share dynamic paths with peers, and adapt to layout changes without re-tape-laying.

AGVs win on deterministic throughput in stable layouts; AMRs win on flexibility and faster deployment in changing environments.

AGV: needs floor preparation (tape, wires, QR codes). Layout changes require re-laying the path infrastructure. Strong in environments where the layout is fixed and throughput is predictable.

AMR: needs a clean-floor scan during commissioning, plus charging stations. Layout changes handled by re-mapping (minutes to hours, not weeks). Strong in environments where workflows or layouts shift over time.

AGV: capital cost per unit can be higher for tugger / lifting vehicles, but lifetime TCO is well-understood and predictable; high-throughput applications (kitting, line replenishment, finished-goods buffer) are AGV's traditional strong suit.

AMR: faster deployment, lower per-unit capex on lighter-duty vehicles, but software, integration, fleet management and traffic-control engineering add cost. Payback is usually faster on AMR for variable workflows; AGV wins on steady high-volume routes.

AGV: JBT, Dematic, Daifuku, KION (Linde / Dematic / STILL), Toyota Material Handling, Egemin. Fork-truck-shaped AGVs from forklift OEMs are mainstream.

AMR: Mobile Industrial Robots (MIR, KION-owned), Geek+, OTTO Motors (Rockwell), Locus Robotics, Fetch (Zebra), Vecna Robotics, 6 River Systems (Shopify). The market consolidated rapidly post-2020 with major industrial groups acquiring AMR vendors.

VDA 5050 is the interface standard between mixed-fleet AGVs/AMRs and a master control system, published by the German Association of the Automotive Industry. It's now broadly adopted across European and growing North American deployments.

With VDA 5050, you can run a heterogeneous fleet — AGVs from one vendor, AMRs from another — under a single orchestration layer. This dramatically de-risks vendor commitment and enables phased rollout. See /solutions/fleet-orchestration.

Mature deployments often layer both. AGV for high-volume repeatable trunk routes (line-side replenishment, finished-goods buffer to dispatch).

AMR for variable goods-to-person picking, ad-hoc moves, and routes that change with seasonality or product mix. VDA 5050 binds the fleet under one master control. We pick the right vehicle for the right route in stage 1 — see /solutions/fleet-orchestration.

AGVs follow fixed physical or virtual guidepaths; AMRs localise with onboard sensing (often SLAM) and replan around obstacles. VDA 5050 (including v3.x work) standardises fleet-manager to robot messaging (typically MQTT/JSON) so mixed fleets can share a master controller. It does not replace safety cases, traffic logic, cybersecurity or WMS integration.

TRACIO advises on locating and fleet orchestration overlays; we do not sell robots. Score route stability, layout change frequency, brownfield versus greenfield, throughput predictability, and how completely each vendor implements VDA 5050.

• Stable, high-volume loops expected to last years, where simplicity beats autonomy.
• Greenfield sites where guidepath infrastructure can be designed in cheaply.
• Lower per-vehicle cost and deterministic behaviour outweigh flexibility.
• Lanes are segregated enough that stopping on blockage is operationally acceptable.

• Brownfield warehouses with frequent layout change — software remaps beat ripping tape or reflectors.
• Shared aisles with people and dynamic obstacles need local replanning.
• Multi-vendor growth path via VDA 5050-capable fleet managers matters strategically.
• Time-to-first-value must avoid long infrastructure install outages.

Trap: comparing only robot sticker price (AMRs cost more sensors and compute; AGVs hide install and re-route cost). Trap: ignoring AMR map maintenance labour in dynamic plants. Trap: assuming VDA 5050 “supported” means full feature parity across vendors. Trap: buying fleets without a locating and traffic plan for mixed human–robot zones. Trap: under-scoping WMS/MES tasking interfaces. Trap: forgetting charger strategy and peak shift concurrency.

Regardless of AGV or AMR, large fleets need traffic management, tasking from WMS/MES, charger strategy and often a locating overlay for shared human–robot spaces. VDA 5050 helps mixed OEM fleets talk to one fleet manager, but safety PLCs, zone rules and cybersecurity remain your programme. TRACIO’s role is advisory on locating and orchestration interfaces — we do not sell vehicles.

Ask vendors to demonstrate obstacle behaviour, recovery from blocked paths, map update workflows (AMR), guidepath change cost (AGV), and how location events flow into your digital twin or YMS. Hybrid fleets — AGVs on frozen highways, AMRs in variable pick faces — are a legitimate 2026 pattern when economics and flexibility both matter.

• What VDA 5050 version and feature coverage is certified in production?
• Who owns the master fleet controller licence and multi-site expansion rights?
• What is the measured throughput on a path that matches your aisle width and dwell?
• How are emergency stops, human exclusion zones and mustering linked to locating?
• What is five-year TCO including map labour or guidepath rework?

Choose AGVs for frozen high-volume paths with lower vehicle complexity; choose AMRs for brownfield flexibility and dynamic aisles; choose hybrids when both exist on one site. Insist on VDA 5050 clarity, WMS tasking and a locating/traffic plan. TRACIO advises on orchestration and locating only — robot CapEx stays with the OEM and integrator you select on merit.

AGVs follow fixed physical or virtual guidepaths; AMRs localise with onboard sensing (often SLAM) and replan around obstacles. VDA 5050 (including v3.x work) standardises fleet-manager to robot messaging (typically MQTT/JSON) so mixed fleets can share a master controller. It does not replace safety cases, traffic logic, cybersecurity or WMS integration.

TRACIO advises on locating and fleet orchestration overlays; we do not sell robots. Score route stability, layout change frequency, brownfield versus greenfield, throughput predictability, and how completely each vendor implements VDA 5050.

• Stable, high-volume loops expected to last years, where simplicity beats autonomy.
• Greenfield sites where guidepath infrastructure can be designed in cheaply.
• Lower per-vehicle cost and deterministic behaviour outweigh flexibility.
• Lanes are segregated enough that stopping on blockage is operationally acceptable.

• Brownfield warehouses with frequent layout change — software remaps beat ripping tape or reflectors.
• Shared aisles with people and dynamic obstacles need local replanning.
• Multi-vendor growth path via VDA 5050-capable fleet managers matters strategically.
• Time-to-first-value must avoid long infrastructure install outages.

Trap: comparing only robot sticker price (AMRs cost more sensors and compute; AGVs hide install and re-route cost). Trap: ignoring AMR map maintenance labour in dynamic plants. Trap: assuming VDA 5050 “supported” means full feature parity across vendors. Trap: buying fleets without a locating and traffic plan for mixed human–robot zones. Trap: under-scoping WMS/MES tasking interfaces. Trap: forgetting charger strategy and peak shift concurrency.

Regardless of AGV or AMR, large fleets need traffic management, tasking from WMS/MES, charger strategy and often a locating overlay for shared human–robot spaces. VDA 5050 helps mixed OEM fleets talk to one fleet manager, but safety PLCs, zone rules and cybersecurity remain your programme. TRACIO’s role is advisory on locating and orchestration interfaces — we do not sell vehicles.

Ask vendors to demonstrate obstacle behaviour, recovery from blocked paths, map update workflows (AMR), guidepath change cost (AGV), and how location events flow into your digital twin or YMS. Hybrid fleets — AGVs on frozen highways, AMRs in variable pick faces — are a legitimate 2026 pattern when economics and flexibility both matter.

• What VDA 5050 version and feature coverage is certified in production?
• Who owns the master fleet controller licence and multi-site expansion rights?
• What is the measured throughput on a path that matches your aisle width and dwell?
• How are emergency stops, human exclusion zones and mustering linked to locating?
• What is five-year TCO including map labour or guidepath rework?

Choose AGVs for frozen high-volume paths with lower vehicle complexity; choose AMRs for brownfield flexibility and dynamic aisles; choose hybrids when both exist on one site. Insist on VDA 5050 clarity, WMS tasking and a locating/traffic plan. TRACIO advises on orchestration and locating only — robot CapEx stays with the OEM and integrator you select on merit.