Talk to any manufacturer across the UK, France, or Germany and you’ll hear the same appetite for AI.
Governments are backing it too, with re-industrialisation high on the agenda in London, Paris and Berlin alike. But ambition is rarely what fails. In manufacturing today, the thing quietly undermining it is the fragmented network that carries and connects the new technology. Stitched together over years into something no single team can fully see or control, digital infrastructure is the invisible thorn in the side of global manufacturers.
Expereo’s 2026 Enterprise Horizons research puts a number on the problem. Only 2% of manufacturers across these three markets believe their network is fully ready for the next wave of AI, cloud and digital demand.
Every manufacturer has adopted AI. Almost none have scaled it
The manufacturers that we surveyed weren’t suffering from AI reluctance. Every one of our respondents is using AI in some form. The problem ambitious organisations face is what happens after adoption, when a pilot in one factory needs to become standard practice across twenty. More than 82% of manufacturers said AI remains confined to sections of the business rather than running across operations, against a European average of 75% across all sectors. Only 8% call their adoption extensive, and fewer than one in fifty say AI has genuinely transformed how they operate.
That failure to scale shows up directly in results. 38% of manufacturers say their AI projects have only partially met expectations, in line with the European picture overall. Notably, at the top end, manufacturing performs just as well as anywhere else on the continent, with 13% reporting results beyond expectations against a European average of 12%. The technology clearly works when conditions allow it to. The question is why so few manufacturers create those conditions.
Fragmentation is the real ceiling
When AI underdelivers, IT buyers instinctively look at the model, because it is the most visible part of the technology. Rarely does anyone question the network carrying it, even though it’s usually the actual constraint. On AI-enabled factory floors, predictive maintenance needs a constant, accurate stream of machine data. Vision systems checking for defects need consistent, high-quality feeds to catch faults before they leave the line. None of that is possible on a network stitched together from multiple ISPs, inconsistent service levels, scattered points of contact, and no single view of how any site is performing. We’ve seen this pattern directly when connecting a major European manufacturer running hundreds of factories and offices worldwide. Each site added fractured connectivity in the same broken cycle until managing the network became a job in itself.
That kind of sprawl is invisible when a network only carries email and back-office traffic. For global manufacturers it becomes the whole story once AI-driven production depends on every site performing the same way, every time.
What manufacturers are asking for tells you where the fix lies
When asked what their network needs to properly support AI, 61% of manufacturers ranked resilience high on the list, against a European average of 53%. Flexibility to connect resources and keep pace with demand followed closely, alongside optimised multi-cloud connectivity. More than a quarter said they now need an external partner to help implement or manage their network. That is a sign manufacturers are diagnosing the problem correctly, as structural rather than a matter of buying better equipment and concluding it is not one to solve alone.
The pressure to solve this is only growing. Nearly two thirds of manufacturers expect supply chain disruption to be a major challenge over the coming year, and 16% now cite tariffs and trade disputes as a significant threat to their strategy. Reshoring and supplier diversification add more sites to connect, more border regimes to navigate and more partners to serve, all of it landing on a network that is already stretched.
Consolidation, not addition, is what fixes this
The manufacturers we work with don’t solve their network problems by adding more bandwidth or more providers. They solve them by removing complexity and fragmentation. Sites move onto dedicated access under a single agreement, with diverse circuits where the site warrants them, and one platform gives IT teams site-level visibility worldwide and finance a single view of what it costs. The outcome is a network that can scale and flex on demand, adding new sites and new technology without adding complexity.
At Expereo, we give our manufacturers one partner accountable for a network that performs the same way whether they’re operating in one market or five. The ambition for AI-driven growth in European manufacturing is not in doubt. What determines the outcome is the network underneath it.
Author: Jean-Philippe Avelange, Chief Information Officer at Expereo
For more articles like this, visit our Industrial Data & AI channel.
