AMRs turned warehouse automation from a capital bet into a capacity decision. The distributors who win will be the ones who measure it like one.
Warehouse robotics used to be a capital project only the largest distributors could fund. Autonomous mobile robots, sold in small fleets and often by subscription, have turned it into a capacity decision a $50 million distributor can test in one zone of one building. That moves the hard part of the decision from finding the money to measuring the result, and most distributors haven’t built that muscle yet.
The Machine Most of Us Couldn’t Buy
When I started at Grainger, and through the early ’90s, the company ran an automated storage and retrieval system that was state of the art. It was massive. It was impressive. And it made Grainger more productive than most distributors of that era, which everybody in the industry understood.
Everybody also understood the other half of the equation. The capital it took to build that system, and the cost to keep it running, sat far outside what most distributors could ever justify. If you ran a regional house with a handful of branches and one DC, you competed on relationships, local knowledge, and hustle, because automation at that scale wasn’t on the table. The AS/RS was a moat made of steel and money.
Fixed automation also came with a second cost that rarely made the board deck: you sized it to a forecast, built it into a layout, and paid for it up front. If the business changed shape, the machine didn’t. That equation has shifted, and it has shifted faster than most distribution leaders have updated how they think about robots.
The Price of Entry Moved
AMRs changed the unit of purchase. Instead of a system you build into the building, you deploy a fleet that works alongside your pickers on the floor you already have. Locus Robotics, the vendor in both cases below, sells on subscription: a per-robot annual fee billed in advance, plus a one-time setup fee for each new site, with software, maintenance and support included. Moving from capital to expense matters for more than the balance sheet. It lets you start small, prove the result, and add robots when the volume shows up.
DHL Supply Chain ran that exact play at a life sciences fulfillment site in Tennessee. According to the Locus case study published November 2, DHL began with six robots in an 18,000-square-foot section, integrated them with its existing WMS, and then expanded to 56 robots across 99,000 square feet. Locus reports the full deployment took less than 90 days.
Saddle Creek Logistics Services did something similar for a Texas-based client shipping medical apparel and supplies, whose order volume had climbed to as much as three times prior levels. Locus reports that operation ramped in about two months, and Saddle Creek is now preparing to grow its robot count by 85%. Both companies made the call in quarters, on a fleet that grows or shrinks with the work. That’s the shift.
The Two Cases, by the Numbers
| Metric | Reported result | Operator |
| Worker productivity | +200% | DHL |
| Order cycle time | 50% shorter | DHL |
| Training time | 80% shorter | DHL |
| Throughput | 150+ units/hour; 20,000+ units/day | DHL |
| Productivity | More than 2x | Saddle Creek |
| Demand absorbed above plan | 30 to 40% | Saddle Creek |
| New-hire time to full productivity | Cut in half | Saddle Creek |
All figures are vendor-published (Locus Robotics customer case studies, November 2 and 3, 2026). Neither case discloses robot cost, the pre-automation baseline, or an independently audited facility return.
The Saddle Creek figure I’d pay closest attention to is the new-hire ramp. Peak season in distribution is a training problem as much as a hiring problem. If a seasonal associate reaches standard in half the time, you can absorb a spike with people you hired two weeks ago instead of turning away orders or burning out your veterans on overtime. Saddle Creek’s senior director of operations described it in exactly those terms: getting a new associate to full productivity standard takes half as long.
The Little Guy’s Real Opening
Peerless Research Group’s 2026 Intralogistics Robotics Survey, run for Modern Materials Handling and its sister publications, reached 166 warehouse, distribution and manufacturing operations in March and April. Fifty-two percent use at least one type of robot, up from 48% a year earlier. Half of the respondents (46%) report less than $50 million in annual revenue, and wholesale trade made up 14% of the sample. Robotics isn’t a big-company club anymore.
Two findings from the same study keep this honest. First, 74% of users said robotics met their business goals, which means about one in five (21%) said it fell short. Second, pure subscription is still the minority path among current users: 53% buy the hardware and subscribe to the software, 37% go full capex, and only 11% use a robots-as-a-service model. The entry ticket dropped across all three, because a distributor can now buy proof in a single zone before committing the building.
Dwight Klappich of DK Advascent, who analyzed the survey, made a point that every distributor near a metro market should take seriously. Respondents ranked physical space as their top constraint, ahead of labor cost. Labor is a problem you can work on with pay, hiring, and location. Well-placed square footage is much harder to add. Getting more throughput out of the building you already have is often cheaper and faster than opening the next one.
That’s where the smaller distributor gets to look bigger and compete. A fleet you can flex lets you absorb a 30% spike without hiring 30 temps, push a same-day cutoff later in the afternoon, and get seasonal help productive in days. None of those moves required a Grainger-sized balance sheet. They require knowing your own numbers.
Five Numbers to Own Before You Sign
- Cost per correct unit shipped. Fully loaded labor, plus the robot subscription or depreciation, plus setup amortized over the term, divided by units that left the building right. Productivity without accuracy is just faster mistakes.
- WMS integration readiness. DHL and Saddle Creek both integrated with an existing WMS before they scaled. If your slotting, location accuracy and pick logic live in an aging ERP module or in the heads of three veterans, fix that first. Robots will automate whatever confusion you hand them.
- Intervention frequency. Count how often a person must step in for a blocked aisle, an empty location, or an exception. Ask the vendor for interventions per 1,000 picks at a reference site, then measure your own during the pilot.
- New-hire time to standard. Measure it before and after. That number decides whether you can staff a peak, and it’s the one most operators never write down.
- Fit to your order profile. AMRs earn their keep on high-line-count, each-pick work. Long, heavy, or bulky products (pipe, wire reels, sheet goods) or a narrow set of fast movers can tip the math back toward manual. Check congestion at peak and replenishment too: robots that pick faster empty forward locations faster.
What Changes Monday Morning
Pull one quarter of pick data from your highest-volume building: lines per order, units per line, weight and cube, and travel distance if your WMS captures it. Calculate your current cost per correct unit, fully loaded. That’s your baseline, and without it no case study, good or bad, tells you anything about your operation. Then ask two vendors to model a single-zone pilot against that baseline, with intervention rates and new-hire ramp written into the success criteria before anyone signs.
Grainger’s AS/RS was an advantage because almost nobody else could afford one. That barrier is gone. The advantage now goes to the distributor who knows its own building well enough to tell a real result from a good case study.
If you’ve piloted AMRs, which number surprised you most once you measured your own building?
If you’re working through automation and AI decisions like this one, the Applied AI for Distribution Conference in Chicago is where distribution leaders compare notes on what’s working. Details at appliedaifordistributors.com.
Share this article:
