Nearly Every Plant Runs MES. Fewer Than One in Four Has Connected It

For most of the last decade, the question on the plant floor was whether to buy a manufacturing execution system at all. That question is settled. Almost every manufacturer now has one running somewhere. The harder problem, and the one that decides whether the AI budget pays off, is that most of those systems still stop at the edge of a single site, or a single line, and do not talk to the ERP, quality and engineering systems sitting beside them.

Adoption Is Done. Scale Is the Problem

Rockwell Automation's "Scaling MES Across the Enterprise" report, released in July and based on 1,560 manufacturing and industrial operations decision makers across 17 countries, puts numbers on the gap. 93% of manufacturers have MES in place. Only 28% have deployed it enterprise-wide, and just 23% report full integration across enterprise resource planning, product lifecycle management, quality and operational technology systems. More than half of the sample (58%) work at organisations with over $1 billion in annual revenue, so this is not a story about small plants that lack the budget to finish.

"MES adoption is no longer the hurdle, but enterprise scale is," said Anthony Murphy, vice president of product management at Rockwell Automation, in the release. The vendor has an obvious commercial interest in that conclusion, but the survey respondents are pointing the same way. 44% rank integration as their top MES buying requirement, and 33% name MES as their single biggest data integration problem. The system bought to create a shared view of production has, in many companies, become another silo that has to be bridged by hand.

The pattern is familiar to anyone who has watched a multi-site rollout stall. The first plant gets a well-funded implementation and a champion. The second and third inherit a template that does not quite fit their equipment, their shift patterns or their product mix. By the time the programme reaches the long tail of sites, the budget has been spent and each facility is running its own version, configured differently and reporting against different definitions of the same metric.

Why the Gap Matters More Now

Disconnected execution systems were tolerable when their main job was tracking production. They are much harder to live with when the same data is supposed to feed forecasting, quality prediction and AI-supported decisions. The Rockwell research finds manufacturers expect 42% of their processes to be AI-supported within the next year and 54% by 2030, yet 43% acknowledge they are not effectively using the data they already collect. "With integration ranking as both the top buying requirement and the leading modernization challenge, organizations risk leaving significant value on the table if disconnected systems and underutilized data go unaddressed," said Lorenzo Veronesi, associate research director at IDC.

Other research published this year describes the same bottleneck from different angles. The Manufacturing Leadership Council's 2026 Smart Factories survey found that 37% of respondents now identify lack of data interoperability as their primary roadblock, up from 22% in 2025, even as more than 90% plan to maintain or increase smart factory investment. Among mid-sized manufacturers, Kaufman Rossin's research found roughly 45% still operate with siloed data, only 27% maintain a data warehouse or data lake, and 55% name the gap between legacy ERP and modern tools as their single biggest AI obstacle.

Put together, the numbers describe an industry that has bought the sensors, the execution software and increasingly the AI licences, but has not built the connective tissue between them. The spending is rising faster than the plumbing, which is why interoperability is climbing the list of complaints rather than falling off it.

What Connected Actually Buys

The payoff from finishing the job tends to show up in plain operating numbers rather than in AI demos. Manufacturing Dive's coverage of the report highlights Moldtech, a custom rubber manufacturer that implemented MES over six months and, in the first quarter of 2025, cut its scrap costs by 50%, improved efficiency by 20% and saved $25,000, according to the report. "We have live scrap reporting available to us now, allowing us to know exactly what impact our scrap will have on month-end bottom line," said Moldtech president Edward Halady.

That is a single vendor-supplied case study, so the figures should be read as illustrative rather than typical. The point they illustrate is sound, though: the value of an execution system comes from connecting the floor to the finance and quality functions that act on its data, not from the software being installed. A plant with MES and no integration has a better scoreboard. A plant with integrated MES has a decision system.

For operations leaders who already own the licences, the work ahead is less about new technology than about finishing what was started.

• Audit what "deployed" means at every site: list which plants run MES, which version, and which systems each instance actually exchanges data with, because the 93% figure hides very different levels of completion.
• Standardise metric definitions before integrating: connecting sites that calculate OEE, scrap or downtime differently only automates the argument.
• Make integration a buying requirement, not a phase two: 44% of manufacturers already rank it first, and contracts should name the ERP, quality and PLM connections explicitly.
• Tie AI pilots to connected data only: with 43% of manufacturers not using the data they collect, an AI project built on a single isolated line will not survive the move to a second site.
• Fund the long tail of sites up front: budget the rollout for the smallest and most awkward plant, not the flagship, or the programme will stall after the first wins.
• Treat security as part of the integration plan: 46% of manufacturers in the Rockwell survey experienced a cyber incident in the past year, and every new connection between IT and OT needs an owner and a review.