The tech is not the story (stupid!), says Siemens; it is a ppart of a solution, of course – just like IoT, and just like AI. Industrial enterprises have their own problems, and don’t buy the hype anyway; but sometimes private 5G helps – and so they call their crane supplier (etc), and not their telco.
In sum – what to know:
Higher uptake – Industry 4.0 interest in private 5G is growing, says Siemens – but it is slower and more use-case driven than vendors first imagined; the silver-bullet pitch is a tough sell.
Enabling tech – Enterprises don’t want to talk about tech, and don’t generally buy the big tech vision; Siemens says private 5G works best as an enabler, embedded in automation solutions
Ready channels – Siemens is selling 5G to existing enterprise and OEM customers (number unknown) via its existing sales teams – and not pitching a top-down panacea like a telco.
Maybe I am imagining it, but Siemens seems happier with its luck in the private 5G market – spring 2026, a meeting at MWC, versus a year ago, same place, with a vague memory of some uncertainty. “It is seriously picking up, but in a different way than we expected.” A couple of quick points for context: one, the private 5G market is still hard work, clearly, and much harder than it was supposed to be; two (and related), Nokia’s pending exit, as a sales operation at least, casts an even harder light. But despite this, the message from Siemens, an Industrie 4.0 aristocrat, is that there are reasons to be cheerful – that customers have problems, and 5G helps sometimes with solutions for them.
The quote about the pick-up is from Daniel Mai, in charge of industrial wireless at Siemens, and a familiar in these pages. He suggests RCR wanders over to the Qualcomm booth in Hall 3 at MWC to see an autonomous production demo, also with Palo Alto Networks. “It combines physical AI, generative AI, on-prem LLM, on-prem 5G; there are none of these dancing robots you see here. Instead, it shows how these technologies bring value in combination, based on existing products. 5G is not the main topic. It is just an enabler; a part of the solution. Which is what we said all along. But it enables really advanced technology solutions – which are practical, and available off the shelf.”
So if it is the same as Siemens always said – that 5G is the support act, not the hero – then what is the plot-twist? How has the narrative diverged, away from the cover story, as perceived when it picked up this 5G saga in 2022? “It is not about IT/OT and AI, just yet. It is still about basic use cases. Industries that have real connectivity problems and expensive critical assets are looking at this,” says Mai.
Which is, despite its surprise, how Siemens has pitched private 5G for 12 months, at least; curiously, it also describes how the vendor market has pivoted, even over a longer period, to present business solutions for business problems etc – rather than just shiny tech, whether 5G or AI or IoT.
The IoT industry has known this for ages; the 5G industry is just finding it out; the AI market will have to learn for itself. Seeing is believing, and all of that – especially for enterprises, more so for industrial ones. “Because they have different KPIs,” says Mai. “Which are not about adopting technology for the sake of it. They run critical operations, based on production and reliability targets. That is the yardstick, whatever it takes. New technology is always slow.” (The question for the AI industry – selling yet another “enabling tech”, albeit a richly layered one that will drive IoT and 5G besides – is how fast the penny drops for sales teams, or else how much the hype gets to enterprises.
Neither is the impetus in Industry 4.0 about the gnarly IT/OT integration to digitise operations – to automate or autonomize. The other grand irony, of course, is that ‘physical AI’ is suddenly a thing – as if any kind of connected semi-autonomous robot, a mainstay use-case for private 5G sales, won’t work without a line, long or short, into an Nvidia GPU somewhere. Clearly, this is not the case: as Mai says, the demo in Hall 3 is ‘physical AI’, familiar in Industry 4.0 case studies from 5G vendors, and perfectly viable with common parts; the difference is it engages an on-prem LLM so the robot arm (atop a robot AGV in the demo) can take conversational instruction. And it is whizzy.
Real industrial demand
But the slight revelation, between the way private 5G was sold and the way it is bought, is just that OT departments in industrial enterprises still need a hard-noised KPI-based rationale, which links to their domains – plus proof that it works. Mai says: “The way 5G was pitched 5G was as infrastructure to take care of pretty much everything – so you can add AI based on its latency and bandwidth. But there was no big-demand. Because these OT-types don’t see the big vision. Some do; and some bought 5G in the early days, and found it didn’t meet their expectations. Things have changed; the adoption rate is higher. But the demand needs to be created, based on use cases and applications.”
Funny, you’d think Siemens, experienced and pragmatic, would know better (than anyone) than to engage industrial firms in sales discussions about fluffy tech; probably, Mai is speaking more about general marketing of private 5G in Industry 4.0, more visible on the MWC showfloor in Barcelona last year. But it is interesting to hear Siemens say it. The firm has a well-circulated 5G reference case with Germany-based manufacturer Salzgitter, one of the largest steel producers in Europe (seven million tonnes per year). Salzgitter came to connect its AGVs (“like pretty much all customers start with”), to transport heavy steel slabs between storage locations and beam furnaces.
It is a hard job. “They don’t stack equally,” says Mai. They are red-hot on their return, from the fiery furnace to the rolling mill and the cutting line. “We use machine learning and machine vision so the system identifies the slabs, and stacks them and moves them. But their first question was: can we build redundancy? Because if it goes down, everything goes down. The financial impact if the process stops is far greater than buying a second system – whatever we do for redundancy. It is the key part to keep production moving. It cannot afford a single point of failure.” (I mean, ‘mission-critical’ is relative, right – telco vendors versus enterprise buyers; memo to Nokia et al.)
So here is the use case to swing the business case – and the way it now rolls in private 5G (in Industry 4.0), says Siemens. Salzgitter has since expanded its 5G system to accommodate a second major “critical” use case: to automate its cranes. “Again, it is one of its most critical assets,” explains Mai. “It wants to fully-automate for various reasons: mainly, because it’s difficult work, and people don’t want to do it. It is an emotional topic.” The setup is similar: camera vision sensors, private 5G network, upgraded crane electronics; sorting steel slabs in the yard to put onto a conveyor, to hand over to the AGV. “More or less, [both together] represent its entire production chain.”
Cranes and control rooms
But hang on; haven’t we moved beyond use-case deployments in factory cubby-holes? RCR wrote some weeks ago that US-headquartered process manufacturing giant Cargill, now with 50 private 5G networks, thinks such an approach is wrongheaded. Forget use cases, it said; because you light up a corner of a factory, and it never works for anything else – and the momentum falters. The new Cargill model says to trust the tech, and the need for it; put the network in, and watch the use cases multiply. “Yes, that is right; Cargill is right, too. But the point is to serve the customer, and solve the problem,” responds Mai.
“It depends who’s driving it – which is not IT, because IT sticks to what it knows. Automotive, say, doesn’t have IT and OT; it has a joint department with clear direction to support OT use cases. So IT is not pushing OT to meet its own needs and standards. It sees things from both angles, focused on OT. Which might also be the Cargill approach – to identify the [broader] value the infrastructure brings. But the other way is that [OT] has a problem, with one use case: crane automation, say. And so you focus on that, and take baby steps between. And then it says, ‘well, we have a warehouse, as well, and the Wi-Fi is shitty.” Hence the point at the start about OT KPIs, as the conversation starter.
So same as Cargill then; just the vendor view, going into different enterprises in different industries – with different problems, agendas, cultures, budgets. More interestingly, perhaps, is the Siemens setup; the firm’s crane division is selling remote-controlled crane automation (SIMOCRANE RCOS) with its 5G system – to crane OEMs and operators. “You basically remove the cabin on top and put it somewhere more convenient – and control a fleet of them with a couple of joysticks from a comfy chair. But not every crane has a fiber optic cable to it, the fiber breaks because the crane moves, and Wi-Fi is not made for those distances. And 5G ticks all the boxes,” says Mai.
“Deterministic comms for video and CCTV, safety traffic; industrial protocols like PROFINET and PROFIsafe – all running in parallel with time-critical control. These ship-to-shore cranes [on ports] are €15 million each. So a 5G system – licensed spectrum, better coverage, traffic prioritization – is not much [by comparison].” What happens when the cranes are sold into ports with Nokia or Ericsson systems? “Yes, that is a challenge. But some customers have struggled with time-critical traffic, and an end-to-end system is best for reliability. So you probably put in a second system – which goes against the 5G principle, one infrastructure for everything, but it is just too critical.
Churn below the surface
Plus, of course, the ports sector was a first adopter of private 5G (when it was still private LTE), so a considerable 4G base (mostly Nokia’s) will be up for renewal anyway. But how many cranes are being sold and upgraded? How many cranes are there? This does not sound like a mass-market bump. “No, but digitization is important for these guys – for safety and productivity; to attract staff. With this, a single operator can be in charge of multiple cranes – without going between. And the electrification [componentry] of a crane is replaced every 10 or 15 years. Not many ports are being built, but there is lots of refurbishment, and the crane industry is way bigger than you think.”
Of course, this is the whole difference with Siemens. It is not an enterprise customer like Cargill or Salzgitter, clearly, with a single type of business problem and a fragmented multiplicity of possible tech solutions. Except this is simplistic, of course: Siemens is a producer as well, with a bunch of factories (including in Karlsruhe, where it put 5G through its paces), and Cargill and Salzgitter face different challenges in different plants in different jurisdictions. But equally, Siemens is not a telco vendor like Nokia or Ericsson in the industrial 5G market: 5G is an add-on for the German firm, to its ‘digital industries’ play – to extend its solutions and deepen its sales.
Nokia is getting out of private (campus) 5G because Industry 4.0 is too hard; Siemens is getting into it because private 5G makes Industry 4.0 easier. It doesn’t have to build (or justify) new sales teams; it already has them, and they already know their customers’ businesses. Where it works (which is not everywhere), private 5G is a proven fix for industrial automation challenges in harsh work environments. As it stands, the company’s strategy with private 5G is not really about driving new business, but about extending and deepening its existing ties. Something like that? “Salzgitter had the vision, proved the vision, and went to the next evolutionary step,” responds Mai.
“But most enterprises are old fashioned, or don’t see the vision, or don’t see beyond their own tasks and operations. They don’t adopt technology for the sake of it; they have KPIs, and so 5G is an enabling technology for a crane. But there is a lot of movement out there – old systems, new requirements; the whole AI story. Enterprises need to collect data in the field. They need connectivity. Whether it is wireless LAN or private 5G just depends on the case. But the need is there – to digitalize and innovate. 5G is just an enabler to make it work.”
Robots, robots, robots
RCR wanders over to the Qualcomm stand in Hall 3, and instructs an on-prem LLM on a private 5G channel to pass a message (which might be in any language) to an AMR in a production cell to build a classic robot toy in red and green. The AMR considers its pile of appendages, and a message comes back via the LLM interface (“on-prem, not in the cloud somewhere; another tick box” – we could have a-whole-nother discussion) that it is out of red arms, and wants to know if blue will do; and so the text instruction is returned, two robot arms pick the parts, and transfer them to a tray on a mobile AGV, which wheels about the place like a model intralogistic system while two more robots arms on top of it assemble a pint-sized automation – in blue and green, with yellow boots.
