Packaging robots are now cheaper than labor over the long run, PMMI says

Labor has become expensive enough that the long-run economics now favor investing in automation and robotics, Processing Magazine reports. Measured against what companies pay for workers over time, the technology comes out ahead on cost. That is the case Tom Egan, vice president of industry services at PMMI, The Association for Packaging and Processing Technologies, made in a Processing Magazine column in September 2025. He named food, baking and snack, beverage and other consumer packaged goods lines as where the shift is playing out.

It is a short claim with a long tail. For the plant manager who has carried the same automation quote from one budget cycle to the next, the comparison Egan describes moves a robot cell out of the discretionary column and into one that can be defended on cost alone.

Processing Magazine's beverage example is the concrete version: robots running high-speed packaging while cutting human error and holding hygiene standards. The labor line on a bottling hall's budget is the number being compared against.

Why the math only tipped recently

The hardware that can keep pace with a packaging line is a fairly new arrival, and that timing explains a lot about why Egan's argument lands now rather than a decade ago. Professor Mike Wilson, chief automation officer at the Manufacturing Technology Centre, laid out the timeline in a Packaging Innovations interview released in January 2025 and carried by WhatTheyThink. In the early 1980s the British Leyland plant at Cowley had two robots, and the Metro line at Longbridge had been fitted with about 15 the year before.

Cowley today runs more than 1,000 robots building the Mini, according to Wilson. Nobody was putting robots on packaging lines in the early 1980s because none of them were fast or accurate enough to keep up, and the heavier hydraulic units of the era were built for load, not speed. High-speed arms in a range of sizes, collaborative robots and vision systems that guide the arm to its target changed that.

Wilson's more useful point for a buyer is about where the next wave goes. Cobots need far less guarding and infrastructure than a caged industrial robot, install faster, and can be wheeled to a different station when the line changes, per the Packaging Innovations interview. He counts tens of thousands of UK small and mid-sized manufacturers that have adopted no robotic automation at all, and he expects that group to supply most of the growth in coming years.

There is a caveat he raises himself. Most cobots installed today still run independently rather than sharing a task with a person, and genuine human-robot collaboration is something he expects to see more of over time rather than something already common. For a co-packer whose line configuration changes with each customer run, the portability is the feature that matters now; the collaboration is a later benefit.

Wilson is also plainly skeptical of humanoid robots in manufacturing. His reasoning, as reported by Packaging Innovations, is that anyone designing a system for automation will design it so standard robots can do the work, which removes the need to pay for a humanoid's cost and complexity. A procurement team fielding humanoid pitches has a credible counterweight in that view.

On the workforce question, Wilson's framing is that robots take the dirty, dangerous and demanding tasks, freeing people for work where their skills add more. The ATM did not empty the bank branch, he notes, and he expects the packaging robot to follow a similar script.

The AI layer: ABB and Agilox on the floor

Packaging World's Matt Reynolds, writing in September 2025, went back to the International Federation of Robotics' start-of-year forecast for 2025 and judged it accurate: the IFR had put analytical, generative and physical AI at the top of its five robotics trends. In the IFR's description, analytical AI lets robots work through sensor data to handle variability in high-mix, low-volume production, while physical AI has robots train in simulated environments and operate from experience rather than line-by-line programming.

Reynolds' case for packaging is that AI-equipped robots and cobots have moved from experiments to working tools, particularly for brand owners dealing with SKU proliferation and labor volatility in high-mix, low-volume settings. Two mobile robots anchor his examples.

ABB's Flexley Mover P603 is a compact AMR rated to carry 1,500 kg, according to Packaging World. It maps its own surroundings using Visual SLAM, so it needs no fixed markers or floor guides, and its onboard AI recalculates load distribution and routing from center-of-gravity sensing to hit sub-centimeter positioning. ABB's AMR Studio 4.0 handles programming by drag-and-drop, and a Fleet Manager coordinates multiple units across a layout.

Agilox's OFL, short for Omnidirectional Free Lifter, lifts and moves pallets up to 800 kg, per Packaging World. The distinction is the X-SWARM software: instead of a central traffic controller, each vehicle talks to the others peer-to-peer in real time, sharing location, intent and task status, so a fleet reroutes and rebalances itself as conditions change. Reynolds frames that as a fit for CPG plants that swap packaging cell configurations or retool lines seasonally, since there is no fixed infrastructure to rip out.

The job these machines do in a packaging plant is unglamorous: moving corrugate, pallets, totes and kitted components between primary, secondary and tertiary packaging stations, as Packaging World describes it. For a plant whose layout changes more than once a year, the absence of floor markers and central controllers is the specification detail to weigh.

The supplier effect a line-level ROI misses

A robot cell's payback is usually modeled at the line: hours saved, scrap avoided, throughput gained. Research on heavy robot adopters suggests the effects spread further into the buying function. A June 2026 study in the China Journal of Accounting Research, using data on China's A-share listed firms from 2012 to 2022, found that extensive robot adoption significantly reduced supplier concentration, meaning firms spread purchases across more vendors.

The authors attribute that to stronger market power and wider product ranges, and they name a trade-off: better firm performance and supply chain resilience on one side, lower inventory efficiency and higher transaction costs on the other. For a procurement director, the sharper question it raises is whether the automation plan and the supplier consolidation plan are pulling in opposite directions.

On the quality side, a November 2024 review in Trends in Food Science & Technology concluded that robotic systems in food manufacturing cut labor and material costs while raising consistency, and that pairing them with AI and IoT improves real-time monitoring, traceability and regulatory compliance across supply chains. That is the part of Egan's cost case that rarely makes it into the spreadsheet, since a traceability gain is hard to price until an auditor asks for the record.

Where the claim gets tested next

The nearest place to check Egan's argument against actual equipment is PACK EXPO International, which Packaging World's show listing puts at 2,600 exhibitors showing packaging and processing solutions, along with more than 150 free educational sessions.

The signal that would confirm Wilson's forecast is simpler to spot than any payload spec. If the small and mid-sized manufacturers he says have never bought a robot start showing up as buyers rather than browsers, the cost comparison PMMI describes has reached the plants that needed it most.