Best of LinkedIn: Smart Manufacturing CW 49/ 50

00:00:00: This episode is provided by Thomas Allgeier and Frannis based on the most relevant LinkedIn posts about smart manufacturing in calendar weeks, forty nine and fifty.

00:00:08: Frannis is a B to B market research company that supports enterprises across the smart manufacturing industry.

00:00:14: with the market, customer and competitive insights they need to navigate dynamic markets and drive customer centric product development.

00:00:22: Welcome to the deep dive.

00:00:24: So if you've been tracking the conversations happening on the factory floor and In the boardroom, you've probably noticed a pretty big shift.

00:00:32: The big abstract promises of tomorrow's tech are, well, they're being replaced by this laser focus on the fundamentals.

00:00:38: Yeah.

00:00:38: Things like clean data, disciplined execution.

00:00:41: Exactly.

00:00:41: And making sure the human operator is still at the center of all this digital change.

00:00:45: It's a noticeable pivot.

00:00:46: you know, toward pragmatism.

00:00:47: Right.

00:00:48: And this deep dive into the insights from the past few weeks really validates that trend.

00:00:52: We're going to unpack all of it.

00:00:54: Right.

00:00:54: Everything from the data foundations you need for AI to the rise of like physical robotics and even how these new manufacturing ecosystems are speeding up innovation.

00:01:04: So the real mission here is understanding how the industry is finally getting past just pilots.

00:01:10: Exactly.

00:01:10: How they're scaling value.

00:01:12: Okay.

00:01:12: Let's unpack this and jump right in.

00:01:14: The core tension seems to be.

00:01:17: everyone wants industrial AI, but we still seem to be stuck on the starting line.

00:01:22: Where does it actually begin?

00:01:24: Well, you'd think it's the algorithms, right?

00:01:25: That's the exciting part.

00:01:27: But the consensus is unanimous.

00:01:29: It starts with data.

00:01:31: Always the data.

00:01:31: Anupchandran's observations are crucial here.

00:01:34: He finds the big gap isn't a lack of interest in AI.

00:01:37: It's just insufficient data harvesting.

00:01:40: I mean, machines are disconnected.

00:01:42: So you have data silos everywhere.

00:01:44: Everywhere.

00:01:44: Critical data just isn't flowing into a unified model.

00:01:47: Yeah.

00:01:47: Clean, connected, standardized data.

00:01:50: That's the real starting point, not algorithms.

00:01:52: And if the data is dirty or incomplete, you're just automating chaos.

00:01:55: Precisely.

00:01:56: And that weak foundation connects right to execution quality.

00:02:00: Jeff Winter emphasizes this.

00:02:02: He says, before you can expect these huge transformational outcomes, you have to get the fundamentals right.

00:02:07: You mean organizing, digitizing, controlling the core operations.

00:02:10: Exactly that.

00:02:11: He argues that most industry four point oh failures come from these great ideas sitting on top of a really weak execution layer.

00:02:19: It's like putting a smart roof on a house with a collapsing foundation.

00:02:22: It just doesn't work.

00:02:24: And that structural weakness that leads directly to the scaling challenge.

00:02:28: Florian Poltrak highlighted a pretty sobering statistic.

00:02:31: he calls the eighty eight percent problem.

00:02:34: The eighty

00:02:34: eight percent problem.

00:02:35: That sounds.

00:02:36: Expensive.

00:02:37: What's behind that?

00:02:38: It's

00:02:38: the percentage of AI pilots that never make it from a successful proof of concept into full production.

00:02:44: So the tech works in a lab, but not in the real world.

00:02:46: The models work, yeah.

00:02:47: But the enterprise infrastructure to support them, the digital thread, it's missing.

00:02:52: That thread that connects design, engineering, manufacturing.

00:02:56: It's just not strong enough to let AI operate at scale.

00:02:59: So we need to shift from building a smart pilot to building an AI-ready enterprise?

00:03:03: That's the critical distinction.

00:03:04: Okay,

00:03:04: but here's where it gets complicated.

00:03:06: Even as we're struggling to get.

00:03:07: four point oh, right?

00:03:08: The conversation is already shifting to industry.

00:03:11: five point oh.

00:03:12: Oh, my G tackled this.

00:03:14: Right.

00:03:14: He argues that revolutions don't wait for the previous one to be perfectly adopted.

00:03:18: They're driven by a new game changing capability.

00:03:21: And

00:03:21: that new capability is

00:03:23: industrially gentic AI.

00:03:25: He provides a really helpful contrast.

00:03:27: So four point X systems are just automation.

00:03:29: They follow predefined static logic.

00:03:31: You tell them what to do.

00:03:33: Right.

00:03:33: Five point O systems using agentic AI introduce autonomy.

00:03:37: They can set and adapt their own logic to hit a goal adjusting on the fly.

00:03:41: So it's less about controlling the machine and more about giving the machine a purpose and letting it figure out how to get there.

00:03:47: Exactly.

00:03:47: And this solves problems that four point oh couldn't even touch, like reducing the cognitive load on your staff.

00:03:53: A genetic AI can shoulder that load and it can even absorb the knowledge of retiring experts.

00:03:59: That's how it delivers ROI so much faster.

00:04:01: It seems like the technology is finally catching up to the need for actual human empowerment.

00:04:06: Davey DeMeier really wraps this up well.

00:04:08: He said the flaw of industry four point zero was the mindset.

00:04:12: Thinking of it as a fixed target.

00:04:14: Yeah, a finish line.

00:04:15: He says instead we have to embrace continuous acceleration.

00:04:19: It's this industrial accelerator act mentality.

00:04:22: Always be improving.

00:04:24: And to do that, you need tools that let you iterate quickly and you know, fail safely.

00:04:31: Which is why something like digital twins are finally moving past the hype.

00:04:34: Speaking of digital twins, they always felt like something only for the biggest players.

00:04:39: But Arthur Sariman shared something really encouraging on this.

00:04:41: Yeah, that smaller manufacturers can get ready for a DT in just three to six months.

00:04:46: That really demystifies it.

00:04:48: It does.

00:04:48: She says the foundations are just process, data, and people.

00:04:52: But crucially, a digital twin has to be a decision discipline.

00:04:57: Meaning if it doesn't help you make a better decision, it's useless.

00:05:00: It's just a glossy visualization.

00:05:02: Expensive shelfware.

00:05:03: You have to start small with the minimum data you need for one key use case.

00:05:07: And as those use cases get more mature, we're seeing new variables come into play.

00:05:12: Mark Stani-Jerez argues that digital twins have to integrate energy as a core variable.

00:05:16: It has to be a decision variable, not just a passive metric you look at later.

00:05:21: I mean, you can't optimize for cost without factoring in power consumption.

00:05:24: It's non-negotiable now.

00:05:26: And what's fascinating is how these digital tools are redefining the physical production methods, especially in hybrid manufacturing.

00:05:32: Oh,

00:05:32: absolutely.

00:05:33: Amelia Dallas detailed the relationship between additive or AM and traditional CNC.

00:05:40: For years, they were seen as rivals, right?

00:05:42: Distractive versus additive.

00:05:43: But now they're teammates.

00:05:45: AM gives you that incredible design freedom, complex internal shapes.

00:05:50: And then CNC comes in for the micron level accuracy and surface finishing.

00:05:54: So Chris Jost's conclusion makes perfect sense.

00:05:56: He said, metal AM's real value comes when it's used with traditional methods.

00:06:00: In conjunction, yeah.

00:06:02: And the only thing that can manage that complex sequence is the digital glue.

00:06:06: IoT, digital twins, MES.

00:06:08: That's what orchestrates it all.

00:06:09: And we're seeing this at industrial scale now.

00:06:11: Carson Hoyzer showed off the EOSM-IV ONYX for metal AM.

00:06:16: And Siemens is providing a complete digital thread for it.

00:06:19: which is enabling these incredibly complex applications like chip cooling that you just couldn't do before with that kind of precision.

00:06:26: Right.

00:06:27: And that need for precision and throughput leads us right to the physical layer.

00:06:31: Automation, robotics, and the people.

00:06:34: When it comes to practical automation, Chris Sturgeau offered a great guiding principle.

00:06:38: Okay.

00:06:39: He says to focus on automating the function, not the convenience.

00:06:43: What does that mean exactly?

00:06:44: It means automation should be a force amplifier for people.

00:06:48: Let the machines handle the heavy lifting, the precision, the repetition.

00:06:52: Let the people handle perception, dexterity, and critical thinking.

00:06:56: So it's about making your people more effective, not just replacing them.

00:06:59: Exactly.

00:07:00: And that's critical as we look at the next frontier.

00:07:03: Humanoid robots.

00:07:05: Matthias Heinecke reports that market is just expected to explode.

00:07:09: Which introduces a whole new level of complexity.

00:07:12: humanoids have to work in the same space as people.

00:07:15: Right.

00:07:15: How do you integrate that safely?

00:07:16: Yeah.

00:07:17: How do you make sure it doesn't mess up your cycle times?

00:07:19: That's why tools like Siemens Process Simulate are so important.

00:07:23: you model it all virtually first.

00:07:25: You're answering those critical questions in the digital world before you spend a dime on hardware.

00:07:29: It's the only way to scale it safely.

00:07:32: And the money is pouring in.

00:07:34: Robert Little pointed out that Jeff Bezos' venture, Project Prometheus, has raised over six billion dollars.

00:07:40: Six billion.

00:07:41: Just to focus on physical AI and autonomous robotics for manufacturing.

00:07:45: That's a massive vote of confidence.

00:07:48: But no matter how much money is involved, the conversation always swings back to the workforce.

00:07:53: Günter Beidinger from Siemens highlighted their approach.

00:07:56: It's still fiercely human-centric.

00:07:59: It's about empowerment and a learning culture, not replacement.

00:08:02: And they're backing it up.

00:08:04: Del Costi announced a huge goal.

00:08:06: to train two hundred thousand electricians and manufacturing experts by twenty thirty.

00:08:10: Wow.

00:08:11: And they're specifically emphasizing AI enhanced skills.

00:08:14: They know the factory of the future needs people with a dramatically different skill set to manage these complex systems.

00:08:21: That's how you tackle the talent shortage head on.

00:08:23: Okay, so let's shift from the physical to the digital backbone that makes it all possible.

00:08:27: Platforms,

00:08:27: connectivity, secure operations.

00:08:29: We're seeing a big evolution beyond the traditional MES, the manufacturing execution system.

00:08:34: Absolutely.

00:08:35: Alessandro Sarasito introduced Ops Center X, which uses the cloud for scalability, for real-time data, for digital twin capabilities.

00:08:44: It's a much more holistic, interactive view of the plan.

00:08:47: So it's moving from a system of record, what happened, to a system of engagement, what should happen next.

00:08:52: Right.

00:08:53: But to scale that, you need sound architecture.

00:08:55: It all connects.

00:08:56: Muhammad Bukri gave three keys for scaling AI successfully.

00:09:00: Let me guess.

00:09:01: One is data architecture.

00:09:03: Number one, the pipes have to be strong.

00:09:06: Second, clear protocols for the edge in the cloud.

00:09:09: And third, and this is critical, the human machine interface has to be a priority so operators can actually trust the system.

00:09:15: Trust is everything.

00:09:16: And speaking of the edge, Jacob Abel had a really interesting technical takeaway from the SPS- Twenty-Twenty-Five conference.

00:09:22: Yes, about portainer.io's new industrial app portal.

00:09:26: He was impressed because it adopts the ISA- ninety-five equipment hierarchy model.

00:09:30: Okay, so for those of us not dealing with massive scale every day Why is adopting ISA- ninety-five on an edge platform such a big deal?

00:09:37: It brings structure to chaos.

00:09:39: instead of a flat list of thousands of devices You can organize them logically by site then area then production line.

00:09:46: Ah, so it makes managing everything Sane.

00:09:50: it simplifies everything troubleshooting patching deploying apps.

00:09:54: It's how IT and OT finally starts speaking the same language.

00:09:57: And as everything gets connected and organized, security can't be an afterthought.

00:10:01: Brian Carroll II emphasized that secure digital operations, or SDO, is actually a growth engine.

00:10:07: That's such a powerful shift in mindset.

00:10:09: You stop bolting security on at the end.

00:10:12: You build it in from the start.

00:10:13: That resilience is what lets you scale confidently.

00:10:17: Security becomes the trust layer for growth.

00:10:19: And that resilience isn't just built inside the factory walls, is it?

00:10:22: It's also about these powerful external ecosystems.

00:10:25: Right.

00:10:25: Our final theme, partnerships and clusters.

00:10:28: Marcus Romeli underscored the huge advantage of manufacturing clusters.

00:10:32: Think Stuttgart or Greenville Spartanburg in South Carolina.

00:10:35: So why does that physical proximity still matter so much in the digital

00:10:39: world?

00:10:39: Proximity drives speed.

00:10:42: When your suppliers, your machine builders, your talent pool are all right there, it creates these incredibly resilient supply chains and you can scale innovation so much faster.

00:10:51: Manufacturing excellence doesn't happen in a vacuum.

00:10:54: Never.

00:10:54: And we're seeing vendor collaborations speed things up too.

00:10:57: Christian Hash shared that MAT and Janice Engineering created a center of excellence by combining their capabilities.

00:11:03: What's the benefit for the customer there?

00:11:06: They're aligning PLM.

00:11:07: NXAM and ERP.

00:11:09: It streamlines the whole digital process from design to execution.

00:11:13: It gets rid of those frustrating gaps between siloed systems.

00:11:16: And then, of course, you have the huge platform announcements.

00:11:19: Osgur Tohumku highlighted Amazon Nova Forge from awsre.invent.

00:11:24: That's a big one.

00:11:25: It lets enterprises build their own frontier AI models, specialized for industrial problems, physical robotics, material science.

00:11:32: So it's moving beyond the generalized LLMs.

00:11:35: Exactly.

00:11:36: deep domain specialization is finally coming to AI development.

00:11:39: That feels like a perfect place to wrap things up.

00:11:42: If you look across all these themes, the AI architecture, the hybrid production, the bets on physical AI, the message is always about discipline and visibility.

00:11:54: It is.

00:11:54: The United Nationality showed how just identifying the single line causing the most breakdowns can change everything.

00:12:00: that simple clarity empowers teens to act faster.

00:12:03: So what's the one thought to leave our listeners with, the one idea that ties all of this together?

00:12:08: Focus on process consistency.

00:12:10: Muserat Hussein offered this powerful idea that the goal of great operations is to make the core process boringly predictable.

00:12:18: Boringly

00:12:18: predictable.

00:12:19: That sounds like the opposite of an innovative factory.

00:12:21: Well, it only sounds boring to you realize that monotony is just a scary word for standardization.

00:12:26: When the process is predictable, all the chaos, the firefighting, the breakdowns, the defects, it all just disappears.

00:12:31: And that stability frees up brain power for other things.

00:12:34: For the high value activities, optimization, creativity, critical thinking.

00:12:38: The whole point of the smart factory is to make the machinery predictable so that people can be truly revolutionary.

00:12:44: That is a profound distinction.

00:12:46: If you enjoyed this deep dive, new episodes drop every two weeks.

00:12:49: Also check out our other editions on digital construction and digital power tools.

00:12:54: Thank you for diving deep with us and be sure to subscribe.