Posted by warehouse_ops_manager_steve

We're planning a major facility expansion and I need to make a decision on material handling infrastructure. Our operations director is pushing hard for AMRs (mobile robots) because they're "flexible" and "the future", but our facilities engineer wants traditional fixed conveyor systems because they're "proven" and "reliable". The space is about 300,000 sq ft, we do ecommerce fulfillment with highly variable demand - peak season is 3x our base volume. From what I'm seeing in 2026 the AMR technology has matured a lot but I'm still worried about battery management, traffic coordination, and what happens when half your fleet needs charging during peak times. Anyone running large scale AMR deployments that can share real world pros and cons? Or am I overthinking this and conveyors are still the better choice for high throughput?

Reply by logistics_automation_jay

Steve we went through this exact decision in 2024 and ended up doing a hybrid approach which honestly has been the best of both worlds. Fixed conveyors for the main arteries where flow is predictable and consistent (inbound receiving to storage, picking stations to pack, pack to shipping), then AMRs for the variable stuff like replenishment, returns processing, and cross-docking. The conveyors can push 3000+ units per hour reliably while the AMRs handle the unpredictable flows that would require crazy amounts of conveyor infrastructure to cover every possible route. Our AMR fleet is 25 units (mix of Locus and MiR) and we scale them seasonally - only run 15 during slow months, all 25 during peak. Can't do that with conveyors, once you build it you're stuck with the capex whether you need it or not.

Reply by industrial_engineer_maria

The battery management concern is real but it's gotten way better in the past couple years. Most AMR vendors now have opportunity charging where robots top off for like 5 minutes whenever they pass a charging station rather than fully depleting and then charging for an hour. We're running Geek+ AMRs and during our peak Black Friday weekend we never had more than 2 robots charging at once out of a fleet of 40. The key is strategic placement of charging pads in low traffic areas. That said, conveyors don't have this issue at all - they just run 24/7 until something breaks. Our conveyor uptime is like 99.2% vs AMR fleet availability of 96.8%, though when a conveyor section goes down it takes out the whole line whereas one dead AMR barely affects throughput. Different failure modes to consider.

Reply by warehouse_ops_manager_steve

Jay the hybrid approach is interesting but doesn't that create integration complexity? How do you handle handoffs between conveyor and AMR zones? Maria those uptime numbers are helpful, 96.8% fleet availability sounds acceptable but I'm curious how much of that downtime is scheduled maintenance vs unexpected failures. Also what's your throughput per AMR? I've seen vendor claims of 50-100 picks per hour per robot but I'm skeptical those numbers are real world. We need to move about 8000 units per hour during peak and I'm trying to figure out if that's 80 AMRs or 160 AMRs lol.

Reply by supply_chain_consultant_chen

Steve those vendor throughput numbers are absolutely inflated, they're assuming perfect conditions with no congestion and optimal pick density. Real world you're looking at more like 30-40 picks per hour per AMR in an ecommerce environment once you factor in travel time, waiting for elevators, other robots in the way, occasional route recalculations, etc. So for 8000 units per hour you'd need roughly 200-250 AMRs if you went pure AMR, which frankly is probably not economical. Fixed conveyors can do that volume with way less complexity. Here's the rough math I use:


The real advantage of AMRs isn't raw throughput, it's flexibility for changing layouts and avoiding massive upfront capex. If your facility layout is stable conveyors win on cost per unit moved.

Reply by robotics_vendor_account_mgr_not_biased

Chen I appreciate the analysis but your AMR pricing is off - enterprise pricing in 2026 is more like $15k-$18k per unit for high volume orders, plus the TCO calculation needs to include conveyor maintenance which is way higher than AMR maintenance. Conveyors have mechanical wear, belt replacements, bearing failures, alignment issues. AMRs are basically wheels and computers, very little mechanical complexity. Also you're not factoring in the labor efficiency gains - with goods-to-person AMR systems you eliminate walking time for pickers. We're seeing 200-300 picks per hour per human with AMR assistance vs 80-100 with traditional pick carts. That labor productivity boost often justifies the robot investment alone without even considering the throughput.

Reply by warehouse_ops_manager_steve

OK so the AMR vendor showed up lol. No offense but I need to hear from actual operators not salespeople. That said the labor productivity point is valid, we're paying $22/hr for warehouse associates and turnover is brutal. If AMRs can actually improve pick rates that substantially it changes the ROI calculation. Has anyone measured this in a real operation? Like before and after AMR deployment with actual time studies not vendor projections?

Reply by dc_operations_director_amanda

Steve I can give you real numbers. We deployed 60 Locus AMRs in Q2 2025 in our 250k sq ft facility, previously was all push carts and RF guns. Before AMRs our average picker did 85 units per hour, after deployment we're at 165 units per hour sustained average (some of our best pickers hit 220+). The improvement comes from eliminating travel between pick locations and having the robots queue up the next tote while you're finishing the current one. We cut our picking labor by about 40% which was 18 FTEs. At $22/hr fully loaded that's roughly $800k per year in labor savings. The AMR fleet cost us $1.1M so payback in less than 18 months even with conservative assumptions. However, we still have fixed conveyors for inbound/outbound because moving pallets and gaylords on AMRs would be insane. The right answer really is hybrid for most operations.

Reply by facilities_engineer_robert

What nobody's talking about is the facility infrastructure requirements. Fixed conveyors need structural support, power distribution every 50 feet, dedicated control rooms, and they're basically permanent once installed. AMRs need flat smooth floors (if your concrete is cracked or uneven you'll have problems), wifi coverage everywhere with proper handoff between access points, and charging infrastructure on separate circuits. We spent an extra $200k on floor remediation and network upgrades before our AMR deployment that wasn't in the initial budget. Also consider ceiling height - conveyors can be elevated to free up floor space but then you need overhead clearance. AMRs are floor-based only so they compete with forklifts and people for the same pathways. Traffic management becomes critical when you've got 50 robots and 20 forklifts sharing the same aisles.

Reply by warehouse_ops_manager_steve

Amanda those labor savings numbers are compelling, definitely helps justify the investment. Robert the infrastructure requirements are a good point, our floors are probably 15 years old with some wear so we'd likely need some work there. Thinking through the hybrid approach more - what percentage of your material flow is on conveyors vs AMRs? Trying to figure out what the right split is. Also has anyone dealt with integrating AMRs with a WMS that wasn't originally designed for them? We're on Manhattan WMS and I'm worried about the integration complexity and whether we need to upgrade/replace.

Reply by logistics_automation_jay

Our split is roughly 70% conveyor for volume, 30% AMR for flexibility. The high volume predictable flows (80% of units) go on conveyors, the long tail variability (20% of units but 80% of complexity) goes to AMRs. For WMS integration, Manhattan has good AMR APIs now, we integrated with our Locus fleet in about 6 weeks including testing. The AMR vendor provides middleware that translates between their robot orchestration system and your WMS. It's not plug-and-play but it's definitely doable without replacing your WMS. Main thing is making sure your WMS can provide real-time inventory location data and accept task completion callbacks from the robot system. If your Manhattan version is older than 2023 you might need an upgrade to get the APIs you need.

Reply by supply_chain_analyst_priya

One factor I haven't seen mentioned - what's your plan for the next 10 years? Conveyor systems last 15-20 years typically, so you're committing to a layout for a long time. AMRs have a useful life of maybe 5-7 years before they're technologically obsolete, but you can change your operation much more easily. If you think your product mix, order profiles, or building layout might change significantly, AMRs give you options. If you're doing the same operation at the same volumes for the foreseeable future, conveyors are probably more cost effective. Also consider resale value - you can sell used AMRs, but nobody wants your used conveyor system, it's scrap metal. We're seeing more companies do 3-5 year lease arrangements for AMRs now to avoid the obsolescence risk.