Who Killed the Cobot Dream? The Truth About Collaborative Robots, Ep #35
40m 59s
This podcast episode investigates the gap between the initial hype around collaborative robots (cobots) and their actual adoption in machine shops. The hosts set the scene in 2015, when Universal Robots and others promised cobots would revolutionize manufacturing by being safe, easy to deploy, and capable of running alongside humans without cages. Forecasts predicted a $4 billion market by 2023, but reality fell short. The discussion identifies several "suspects" for this failure. First, marketing overpromised, suggesting cobots could replace workers immediately and required no programming—claims that were only true for simple tasks. Second, technical limitations emerged: cobots are intentionally slower for safety, have low payload capacities (often only 60% of advertised when reaching into machines), and suffer from reliability issues in dirty, humid shop environments. Third, misapplication occurred when shops used cobots for heavy-duty machining tasks better suited to industrial robots. Finally, integration complexity proved a major hurdle. While cobots worked in controlled demos, real-world job shops revealed payload constraints, speed mismatches with high-speed machining centers, and the need for skilled staff to troubleshoot issues. The hosts conclude that cobots are not a universal solution but can succeed in specific, appropriate applications—particularly for lights-out operations where consistent, slow work yields significant gains. The dream wasn't entirely killed, but it was tempered by practical realities.
I think with anything we discuss on this show, the best automation is the automation that actually works and runs in your shop and makes you money. Not the coolest flashiest tech, not the trendiest solution is the one that works for your parts in your specific situation. Talent is scarce, experience even more so, driving a cost in sleepless nights. You may even think to yourself, "Can I do this?" But this new era brings new technology and a different perspective on your operations. What you will hear will turn on a light bulb in your mind so you can turn out the lights in your shop. Get ready to sleep soundly while your shop never stops. Get ready for making chips, lights out with your hosts, Nick Golnar, Eric Neckich and David Buick. Gentlemen, there has been a murder on the shop floor. I do decree we must work as a team to find those responsible parties and bring them to the swiftest of justice. May the Lord join you on your journey today. Oh, good man. Who does? How many times did you practice that, Eric? All the entire ride flight home from New Jersey. Yeah, lovely. The people sitting next to him on the plane were like, "What the heck is wrong with this guy?" He said, "It's not that I always thought." How many times? The fact that cold open, welcome everybody to the Lights Out podcast. The only podcast that equips and inspires manufacturing leaders with advanced manufacturing technology to thrive. And in today's episode, I'm joined by my co-host Eric Neckich and the long-lost plane, Nicholas Golnar. Welcome back to the show, Nick. Yeah, thanks for having me back. Dude, I was here for the last one. I just missed the two before that. You've all made me off. You're right. Sorry. It's just, you know what, you're such a critical part of the show, Nick, that as soon as you're gone for one, it feels like you're gone forever. Uh-oh. Thanks guys. Absolutely. Yeah. Let's find out who done it. I'm excited for this murder mystery. Me? Yeah. In today's episode, we're going to talk about co-bots and really who killed the collaborative robot dream. So let's set the stage. Picture this. It's 2015. There's a Danish startup called Universal Robots. Selling collaborative robot arms worldwide. The industry is really buzzing about this. Experts are predicting that co-bots are going to capture a massive part of the robot market. And some forecast had that market hitting 4 billion by 2023. It was like an annual growth rate of about 50%. Fast forward to today. And that market is still growing, but just not nearly at that predicted rate that was being talked about way back then. So tonight, we are putting on our detective hats to solve this mystery. Who or what killed the co-bots dream for machine chops? That was really good. Bigger, bigger, bigger. I think that's the CIS thong. Yeah. Yeah. What is that show? We're going to have to edit this out. NCIS. NCIS. Yeah. Okay. You can't. Okay. Anyway. Is that the NCIS thong, Eric? Is that what that was? I believe so, yes. Oh, good job. It was so good. We have copyright problems now, thanks. All right. Let's get in. All right. So I think we should kick into this by starting with the scene of the crime. We'll talk about the golden age of co-bot hype. Guys, do you remember when co-bots first started showing up at trade shows and some of the kind of demonstrations you saw? You want to talk about that a little bit? Yeah. I mean, I can jump in right here. It steamingly overnight went from like, you are robots to then like everybody. Like, they were just everywhere. I remember going to the Automate Show in Chicago and I could, I was just taken aback by the amount of co-bots. Be brand I'd never heard of, but they were just everywhere. Yeah. For me, I think you are was kind of like the only brand I really saw. That I was just like, oh, that's a co-bot. I know there was other stuff out there, but I just I know that like gray and baby blue. And I would see it at pretty much every show, every open house, everything. Rarely did I see it in actual shop, though. Well, I think the crazy thing at the time when you were seeing these demonstrations at trade shows was like that robot arm was so close to the people, you know, for the first time. No barriers, no cages. And I think that was a pretty big deal, you know, when they first came on the market to see that. And then to think about some of the possible applications and things that that could unlock in terms of accessible automation, people could just go out by a robot arm and start to tap into that. No matter how much you tempted it, it would never hurt you. It's like the guard in England that just can't move. No matter how much you taught them. And I would try. I would try, but I was so weird as visual and you taunting a co-bot. You're not a real robot. Let's see you try it. But I mean to be fair, I think there was a ton of legitimate reasons for the excitement, right? I mean, traditional industrial robots required expensive integrations, the safety guarding, programming traditional industrial robots seems very complex. And you know, you pair that with, you know, the labor shortage, which is, you know, as much a problem back then as it is today. And, you know, it was really exciting to think that, you know, you could just go out and buy one of these things and set it up next to your mail or next to your live and, you know, have these gains. Yeah, there was the sense that anybody could go buy one and you could just kind of like get your own automation going and there was all this cost savings that was going to be involved and I remember those kinds very well. Yeah, and you would see it too. Like when you would go to a demonstration at a trade-over and open house, like they would just like start from start to finish and set it up and teach it to load and unload a machine, for example. And so you'd be like, well, that looked really easy. I can do that too. Totally. Yeah. Yeah. So I think that's our victim, right? It's the co-bot dream, which kind of promise to democratize automation and solve, you know, manufacturing's labor crisis. And I think what we have to do next now is examine what happened with these things when they started to hit the shop floor, what were some of the pitfalls, and where are we now? Let's start with collecting the evidence. You know, what is the data show? I'm just ready for like the for a lot of home impersonation. It's not coming. It's interesting. It didn't work on that one. We got to get you on another flight. Exactly. Yeah, man, I honestly been pretty curious about this because I don't have all the answers for why the dream team to fizzle out didn't it? Thieve the heights that it promised everybody. So let's collect some of the evidence. What are your thoughts? I know we've talked about that on the floor that there was kind of this moment in the industry where like I know all of us travel professionally in the industry and are in a lot of shops and there was this moment where we are seeing more co-bots still in the package than actually loading part. So yeah, David set the scene. I've excited to get the cock line. Like what are we walking into here on the job board? I think the problem was you know there's this full promise of a collaborative robot that's easy to deploy and can replace human labor for a lot of these repetitive tasks that human beings didn't really like doing anyways. And you know maybe in practice what ended up happening was it's like okay the integration is actually more difficult than expected especially if you do that on a DIY basis and maybe we'll get into that a bit deeper in a little bit. And kind of turns out they're just slower industrial robots that at the time didn't need safety guarding or cages. And you know and let's talk about the elephant in the room which is speed. Every machine tending video I've seen online has the machine in the comments going why is it so slow? You've got these high speed machining centers you know I think of like the robot drills and the brother speedios and the Hoss DT for that matter. There's many others like them. You've got these really high velocity tool changing machines that move super fast and then you've got this like uber slow robot arm coming in to do the you know retrieve the part of course I'm doing this on video and most people are probably listening but it was painful to watch right? Like I was grandmother robot like it's like it was moving through cold molasses which was by design of course because co-bots prioritize safety over speed because they're working near human beings without guarding but I think that was one of the problems for sure. Hey if you're responsible for setting direction in your company you already know manufacturing isn't standing still from global trade and geopolitics to industrial AI and advanced automation. The forces reshaping our industry are accelerating and they're redefining competitiveness and risk in real time.
That's why MFG 2026 matters. March 10th through the 12th in Fort Lauderdale, Florida, industry leaders are coming together for high-level conversations, designed specifically for executives who need to think several moves ahead. If you're driving growth and shaping strategy, this is the room you want to be in. To learn more, go to amtonline.org, navigate to events, and you'll find MFG 2026 there. It's time to lead with confidence. I just want to say, like, kind of the counterpoint to the speed argument in favor of the co-optress. Yeah, they're not as fast, right? Of course. And when you've got an arm whipping around that you taunted, it will knock you out. So don't do this to an industrial robot. Don't have an existing constant because it will do the knock you out. But we do put boxing gloves on all the robots at Henning in eight or without reason. Case you're an idiot like me. No, but I just want to frame it to, okay, yeah, maybe it's moving flow as it loads and unloads apart. But I mean, if you're running too shift with no machining, it's still just tremendous gains, regardless of how fast the robot is moving. Now, you start to notice that speed thing when you're not just running one shift. You know, when you're getting the second and third shift and you're loading at percent more because the automation is moving faster. It's hurt the matter more. I just think people put too much emphasis on how quick the robot move from A to B. Surely you have something that's keeping this mental turning. Well, I think the other drawback aside from speed was payload. There were some high payload co-bots, but most of them were like in this three to five to as much as 10 kilogram payload capacity. Of course, the price point increases like crazy as the payload gets higher. But that also can be really limiting in CNC machining applications, especially when part of that payload includes the gripper that needs to be on the end of that arm as well. Well, and to that point, you know, we were always experimenting with our own automation. You kind of discover that like in reality, you're basically going to get 60% of the advertised payload to do like a consistent repeatable tending operation, right? I think that when they give you payload, correct me if I'm wrong, but that assumes in front of the robot, like right in front of the robot at a certain height, not three feet out inside a machine tool, you know, picking and placing part. Yeah, it's like, what can you bench press when the bar's right here and you're set up to like make that lift versus like, can you do the same thing with like one arm right way out there, right? 350 pound, but anyways, let's be fair. Some shops are seeing great success with collaborative robots. And you know, there's plenty of case studies out there showing that folks have experienced great ROI. And I think what it comes down to is that maybe co-bots aren't a universal solution. They're not good for everything. But in certain applications, they can be really good. There's just maybe some pitfalls to avoid in terms of how you're going to deploy that particular technology onto the shop floor. I think that the juicier piece of evidence in terms of speed versus payload is absolutely payload. And if we think about what kind of collaborative robot you're I think to another thing we have to talk about. And I believe it's probably better now, but early on, the ability and durability on a shop floor out in the open became an issue for a lot of people in curtain and doctoral climate. You bump into it. Get the pendant out. Sure. Okay, that's right up there with the payload as far as like what I've actually heard from real stop owners who have been frustrated with co-bots is that reliability, the dependability, the best ability is availability. And when one thing goes wrong, now you're all about just sitting there being like just annoying thing. You have to navigate around in front of your machine until somebody fixes it. Of course, this is both pretty easy to fix, but does everybody know how to like re-threat that that and get it working again? Right. Well, that's the thing, right? It's not set it and forget it. And do you have someone on your staff who's actually able to manage that and respond to those situations quickly and get that robot back up and running? Or if the owner is the only person that knows how to do that is that the best use of the owner's time, right? So I mean, that's that's also an important consideration there. So I mean, we've gathered some evidence here, co-bots totally work, but not for everything, necessarily, maybe not everything they were hyped to do initially. I think now would be a good time to interview our suspects and figure out who's really responsible for maybe that gap between the hype, the promise, and what the reality ends up being. I cannot bring forth. Bring forth stuff. You'll never take me alive, do it. All right. Well, let's kick it off with this. Nick, you are a sales and marketing whiz. Let's start with our first suspect, which would be marketing over promise. Do you think that co-bots were oversold in a way or tell us your thoughts on that? Yeah, it reminds me a bit of the early days of additive where it's like, we don't need machine tools anymore in 10 years now. So yeah, a little bit of that. And then they over-promise of like, you see all that labor, that guy who you were just complaining about who called them sick all the time? You don't need him. You can just get rid of him today because I got to hold on for you. Yeah, a little bit of that. Maybe a lot of that. Sure. Yeah. And to be fair, some of the marketing hype was textically true. So for some very simple applications, the no programming needed claim. So there's like the guide through programming you could do. Definitely true to that. And it was like how applicable was it in reality on the shop floor? I think suspect number two would be the technical limitations. Again on the payload capacity we talked about already, maybe just not being heavy duty enough for a lot of the heavy components that get moved around in machine shops, especially if you're into heavier machining, die work and that kind of a stuff. I would also reiterate that tactical limitations, payload 100% a lot of shops, I think don't understand some of the constraints when you're just running from a pendant versus having like integrated software and how the differences in communicating with the machine tool. Yeah, absolutely. And I think there's some truth at the end of the day that by design, co-bots were slower and weaker because they worked alongside humans and what is really needed in heavy duty manufacturing applications are speed and strength, right, which is kind of more the realm of an typical industrial robot. I feel like the suspect may be working together. I'm going to make a bold accusation at the end of all the internet. I can't wait. What do you guys think about as a suspect, misapplication, which kind of ties a lot of this together? You're the very bane of my professional existence right there. Yeah, working as intended and not every again. I think like Nick said, right, like all these kind of suspects may work together, right? If you kind of drink the cool aid early on or maybe misunderstood and failed representation, technical representation can vary vastly region to region to region. So maybe you bought something assuming it was going to work in a certain way and that did not come to fruition. It happened in more than just co-bots. It happened with cutting tools and workholding and machine tools. Yeah, absolutely. Yeah, it's kind of like buying a sports car to like haul lumber. It's just, you know, it's not the right tool for the job necessarily. And you know, back to Nick's point, even if it is slow, if you want to call it the, the sloth of robot arms, it's the sloth that shows up every day. It doesn't take breaks, works through the night when you're torn. The tortoise in the hair. When you're when you're on watching Netflix. It's probably still moving. The lock machine tool operator is the most glorious visual in my brain right next to the top. We just get that on a t-shirt. Just die at the finger. Anyway, Nick, you were about to say something there and then you gave me the mic back. Yeah, no, I just, I'm now more convinced than ever that these suspects are working together. Because as you introduce this next suspect, by the way, I have you now. As you introduce the next suspect, I am just like, okay, they're in Kahoot. There's not one criminal here. This is a criminal organization that we have to take down. Yeah, bring me suspect for Mr. Vueck. Paul, tell me about your trip to Italy. I went there with the NTA and we got to tour SMW Auto Block. Two different facilities. Absolutely blew me away. World class, enormous places. I can't even count how many machines there were, but so much automation. One of the most incredible takeaways for me was the breadth of what they made. I did not know they made so many work-holding products and automation products. I loved watching a big chuck, machine a little chuck. They were machining chucks and huge lades. I thought that was super fun. You introduced them to me after you came back and they've been to my shop and the ideas they have is like, breathe new life and some of my machines and the tools and the jobs that we're running was just shocking. And then one of the coolest things is we're bringing this new automated
pallet delivery system, the Henig workflow to market and guess what the interface is. It's SMW zero point locators. It's really sweet. We're getting ready to get some of those. So how do people find SMW auto block to look through that massive catalog of products? It's SMW auto block BLOK.com. Yeah, get rid of that. SMW auto block BLOK.com. All right, suspect number four, integration complexity. Yeah. This is where I know suspect one, the marketing over promise was working with prospect number four because it's like, "Han, do it?" Yes. Then all the testing we did in a perfect environment, it worked. Then you go into an imperfect environment, what we like to call a job shop, and they uncover all of these issues. You see that all the time with all sorts of different technologies where they test it in a perfect environment. They're like, "Yes, it works. It works as we programmed it to." Then you bring it to where you get punched and everyone has a plan if you get punched in the space. In your job job. That's when it's going to be like, "Okay, well, maybe there's a payload limitation." Maybe we're using it for the wrong, "We're misapplying it." Maybe we thought it would deal a little bit faster than it actually is. We didn't really think through the entire integration and outfit into our entire profit. This whole, "Well, you can just do it yourself." It's really awesome until you have to do it yourself. You thought, "Oh, go ahead, David." I was just going to say the underlying theme there, I think, for everyone to think about is, if it's hard for a human to do it, it's going to be hard for a robot to do it. That's what you have to remember in sense check these marketing claims against. That one thing that keeps going around in my head because we all work in manufacturing. Are these the usual stuff back? It's just super unique. Are these suspects just completely unique to the great co-bot paper at this point in time? We see that consistently. Think about it. We're talking about part presentation to the robot, six-turing vision systems, grippers, which is kind of like work holding. There's elements of a robot application or an automation application like this, which is also true of traditional CNC machining as well. There's a lot going on, a lot of overlap there in terms of building a process that's repeatable. Yeah. Eric, this is the usual aspect. This is the criminal organization that has committed many crimes. We're just packing the layers of the onion here and realizing that there's a lot more than just one crime. Yeah. Right. Three guys that are going to put them on trial, by least. And then kind of something that didn't really help either is all of a sudden the market got so saturated. So suspect number five would be like market competition, market saturation, all these little kind of low-cost competitor co-bots coming out of Asia, especially in the kind of run the 5 kilogram payload range, which all of a sudden just brought the price point down for some of the folks that had been in the market a little longer. And I think that kind of squeezed co-bots from both sides. I mean, traditional industrial robots also got better, cheaper and easier to program as well. I'm not sure that helped either. No, until your point, like we run into that with advanced manufacturing technology all the time. We run into it with monitoring. We run into it with automation is people try to go for this bargain. You know, hey, well, this one within my budget and it looks like I'll just say, oh, you are right. So they'll bring it in. They'll try it. It'll fail. And then they automatically, they're in the, we already tried that room. Right. And so it becomes bad for the technology as a whole, right? Where there's plenty of co-bots and co-bots companies that are outstanding product when used within the confines of, okay. We can summarize our suspects for us just to kind of sum up this segment. Always marketing guy. Yeah, yeah, yeah. It started with this will solve all of your problems. You don't need people anymore and you're going to run 24/7, the marketing overpronter. And then from there, it was like, all right, well, let's bring it into reality. What can it actually do with its payload and its speed limit, 30 seconds, and it's technical to. Perfect. Perfect. Perfect. So, stuff, spec three. Any years, Eric, just suspect is your, you know, marketing is to application. Yep. You want to equip this guy, but I mean, he needs, he needs to be put on trial. I don't know if you can add again. If you're going to go after Marty, you got to tas him. You can take him down with an elephant. Do they make tone? I think that's so mean if they do. You can't tas an elephant. That's illegal. Please email the answer to Nick's question to lights out at makingchips.com. And next week's episode, we'll learn about elephant tasers. Yeah, it is advanced technology. The spec four, this is where we started to realize, I think, that this is a group of suspects that this is a criminal organization, the cartel, automation, the integration reality versus the marketing claims, the whole, you can do it yourself. You don't need any special software. It doesn't need to be intuitive because you just teach it. And then it just knows exactly how you think and it continues on from there. Yes. And then we go five. Let's talk to our international hoe about the spec five. Yeah. Competition from traditional industrial robots that became more affordable and market saturation in the co-bots base with other manufacturers hitting the scene with just a ton of very similar looking five kilogram payload co-bots. And to Nick's point, you know, who actually killed the co-bot dream, which of these suspects, it was all of them. And here's the twist. The dream isn't actually dead. Time to reveal what really happened. Spoiler alert, this is foreshadowing. This might not be a murder mystery after all. It might be a story about unrealistic expectations meeting reality. Oh, there. Which takes us to know the suspect dead or we don't know if the victim is dead. Is that the point here? We might just we might just find that they've just been hiding and we need to bring them back to life. I don't know. But let's get into the verdict, the truth about co-bots in 20s. It's a commercial break right there, built the suspense, okay, then we come back. Eric, you gotta do your song again. Boom, boom, boom, boom, boom, boom, boom, boom, boom, boom, boom, boom, boom, boom, boom, boom, yeah, that one. Yeah, that one. Okay. And we're back. Wow. That was really the timing. There was perfect guys. I love that. So let's be clear. The market is still projected to grow. Seeing like a 20% compound annual growth rate versus like the hyped up initial 30 to 50% annual growth rate. This success with co-bots really requires the right application with the right expectations and the proper integration. So co-bots didn't die. The hype died. It's probably a good way of saying that the market is still growing. And it's a market segment that's maturing. What's really interesting is that quite recently the industry actually killed the term co-bots. So this isn't pretty important, especially considering we spent the better half of the last hour saying the term co-bots like 100 times. It's technically not a word anymore. So ISO and ANSI updated the US safety standards in 2025 to eliminate the term collaborative robot or co-bot entirely. Instead, the acceptable term is collaborative application. Here you correctly even co-botship pronouns. No. Sorry. Don't tell me. I don't identify it then anymore. That's great. I think there's some things that kept co-bots alive. They're still much more affordable than a big massive investment in industrial application. I don't know if we could call them industrial robots either. We got to see if they're name and change. But it should be much easier to apply a co-bot in most cases financially. Yeah. Well, the thing that's point earlier, I've seen plenty of the acceptable co-bots applications. Were they running them exactly how they wanted them to run? They're running on night. They're running on weekend. Yeah. That's right. The big takeaway here is look, with the old standard, you could just go by a co-bot and assume it was safe. The new standard says safety is determined at the application level. You need to conduct an application-specific risk assessment. You need to document how the humans and the robots are going to interact. There's a layer of complexity with that. For good reason, we want to keep our human team members safe when they're working in close proximity to co-bots. And frankly, I think in terms of de-risking the investment in automation, conducting an application risk assessment is also forcing you to look at the application that much more closely to help you kind of determine if it's the right application for that specific technology. Are you going to see the corresponding ROI?
plugged back to kind of our previous series that Eric and I did on calculating that ROI. So the standards finally caught up with what we kind of learned the hard way if that makes sense. Just because a robot can work near humans doesn't mean it's automatically gonna be safe to do so. - And I was talking with Caleb who used to host this with us and sometimes comes out of the guest's hosties, the only one about through that, the robotics applications engineer. And he was telling me like, you can switch mode now. So you can have it in industrial mode or collaborative mode. And I was just picturing myself like taunting the robot. Like you can touch me, you can touch me. It's in the wrong mode and I'm just knocked out on the stop-war 'cause it's in industrial robot mode where it can to knock you out. So that's where you have to do the application-specific assessment and also probably just don't do what I do. - Yeah, don't taunt an animate object. Literally the plot of Terminator. So, anyways. - Exactly. - So guys, I mean, when do you think a collaborative robot or a collaborative application with robotics, I guess is the right way to say it now, when do you think that makes sense in the context of a machine-shot now in 2026? - Eric, go ahead now. I'll do it. You say what I'm looking at there. - That's a great question. When does it make sense? You know, and this might be an uncharacetment, but I think the first thought that come to my mind are gonna be budget and fourth base, some of the industrial robots because of the way as six acts of robot have to move because of some of the guarding. I mean, they require maybe a little more floor space. You know, we know some of the co-bought on a cart systems, even though I know we're not supposed to say co-bought, but I think everybody listening will understand what I'm talking about. Those tend to be more budget-friendly, right? And those systems get really re-entered. Though those are the two scenarios that come to my mind. Attending, tending is usually right where my brain tends to go, right, I've got like a rat. - A rat? - A rat? - Or a. - Oh! - I've got a rat or a table or some sort of magazine full of a rust stock that I'm just going in and out of the machine, lightweight, really repetitive. The vice manages the repeat abilities, just have to get close. Yeah, that's similar to what I was going to say. What you said at the end there, so like the more repeatable and mundane and repetitive the more it's probably a good fit for the co-bought and then the more it needs some intuition and some thinking. Some of those, what are we supposed to call them? Now, I don't want to miss main them anymore. Collaborative applications, is that what it is? - Some of the more collaborative applications where it has to make maybe some decisions on what to load and unload next or some decisions on basically how to keep the process running. That's where adding some software and maybe a different style of automation around the kind of industrial fried with, to me, that's where it would make more sense than the DIY co-bought that you see oftentimes just like in a box next to a machine because they couldn't figure it out. - Yeah, totally agree. I mean, part of some of the qualifying factors with that have to do with part weight, load and unload kind of time requirements as well if that's a goal for the automation project. Flexibility of changeovers, if you're running smaller batches versus longer production runs, all of that plays into whether or not this is the right fit. And I think what it highlights is working with an automation partner that can help you walk through the application and really determine what is the best technology solution for this particular project or for what your goals are in your shop when trying to achieve some degree of automation. - Leverage the experts people? - Yeah, so Eric, you mentioned just in some really strong fix up cases with collaborative applications or co-bought. So can you give us like a story of maybe one of them and why it feels successful and why you think it was the perfect bit? - Yeah, 100%. So medical manufacturer, I won't name them, but David and I talked and I think we're going to try to get her on the show because it's a remarkable story. But medical manufacturer in Indiana doing plastics. So, you know, a material that is prone to deformation, you kind of got to be careful with it when you're machining on it, but lightweight, right? And then one of our build on the rock partners, the Faxus has this product called the LightVite, which is essentially designed to be loaded with a co-bot. And then Faxus also sell like its own Shelpin Rack kit, right? And an ematic zero point that goes in the machine. And these are all built around the idea that you kind of buy the pieces and then you have people on the payroll that know how to like bring it all together and make it into this system. But I think the key component there is this particular manufacturer has like two or three people that are on the payroll, they're automation people, that's what they do. So it's not like they've taken someone from the tool room and said, here's a co-bot, here's a vice, figure out how to load it. So they have the right people bringing in components that are designed to do what they want to do. And then the vice is awesome because it really gives a, we always say the robot doesn't care what's in the vice, right? As long as the payload's right, it's just always loading the same vice in and out of the machine and the work piece is whatever the work piece is. But because their work pieces are so light, it gives them a high level of flexibility. Yeah, I think definitely perfect storm in that regard. As you were talking, I was just thinking about, so kudos to them for having dedicated automation people that their job is to just keep the machines running when humans aren't in front of them. Also, the humans that are in front of machines, if you want to be bulletproof for the future, become that as well. I think the machineness of the future has an automation hat that they wear and they need to understand how to keep the machine running when they're not in front of it. In addition to all the things that make them a great traditional machine, give me a shop full of people with that mixed skill that and after all your world class. Yeah, absolutely. Okay, so case closed, who killed the robot dream? The answer is no one really. The dream didn't die. It just kind of grew up. In fact, the industry just made it official. They eliminated the term co-bought entirely. It's now collaborative application, which is important to remember, which I think a lot of industrial automation integrators probably saw coming for a number of years. Safety has always been a really important topic around any kind of industrial automation application. So it's a more mature, realistic approach, but these robots are finding their place in modern manufacturing. It's perhaps a more specific place than the hype initially suggested, but they certainly have application all the same. Yeah, I think you said it well. They didn't die. They just grew up and got a reality check. I mean, when you're 18 years old, you think you're gonna conquer the world before you're 25 and then you realize, okay, maybe it's gonna be a little bit harder. So that's kind of a nice parallel here. There's still a place. I think the keyword is application. Just making sure you really understand the application and not, you know, that one's back in the process. Like we talk about for frequently, you understand the entire flow through, through your shop and when does it apply and think another point to be made is like, maybe it's a mix, maybe some time at the collaborative application, maybe sometimes it's more of a traditional industrial robot application. We've got people like Caleb who can help you figure out exactly when to apply each one, but there's rarely a panathy area in this industry where like, okay, additive, now we don't need machine tools, co-bots, now we don't need people and I think the next one that we're gonna end up talking about is the humanoid robot that we're hearing about that, you know, optimist is coming and this person, that this robotic person will just be able to walk you work for one day and then you'll never have to work again. I'm pretty, the specialists are better. I've access to, it was gonna kill three access, right? I mean, it just kind of always happens and I love the excitement and that's what keeps me going in this industry and I'm sure it's the same for you guys and a bunch of our listeners, right? - Well, I think with anything we discuss on this show, the best automation is the automation that actually works and runs in your shop and makes you money. Not the coolest flashyest tech, not the trendiest solution is the one that works for your parts in your specific situation. - That was beautiful. - I think we can end right there, David. - Okay, like I'm gonna start crying, it was so beautiful. (laughing) - Your part is your piece. Let me get through a virtual tissue, Nick. Cool, well, let's close this out then. I think that's our show for today. Mystery solved. If you found this useful, subscribe and share it with your fellow shop owners, fellow, he members and drop a comment with your own co-bought experience. If you have one, we'd love to hear about it and also any real automation stories that you wanna bring to our attention. That's always good to work into our show. All right, thanks for listening. Get back to running your world class shop. Thank you for listening to Lights Out. Let's dim the lights on this episode. (dramatic music) - In manufacturing, growth is an accidental. It's engineered. But a lot of suppliers are still building their pipeline on outdated lists, inconsistent follow-up and hope-based marketing. Hoping their website will somehow bring in leads. Gabe Draper, the founder of Factor, knows that struggle. Years ago, he ran a shop that went bankrupt, not because of quality, but because he didn't have a steady way to generate new opportunities. that experience.
is what led him to build Factor. Factor helps suppliers create consistent sales opportunities by combining targeted outreach with marketing that puts you in front of the right buyers. Their team finds the right companies, starts real conversations, and sets qualified meetings. So if your pipeline swings with referrals or seasonality, Factor brings a system that puts you back in control. And for listeners of this show, Factor is offering a free custom report that highlights the best opportunities in your industry, plus a list of real companies that match what you do. Request yours at FactorMFG.com/chips. That's FACTRMFG.com/chips. [MUSIC PLAYING] [BLANK_AUDIO]
Podcast Summary
Key Points:
The collaborative robot (cobot) dream promised to democratize automation and solve manufacturing labor shortages, but fell short of hype.
Cobots were marketed as easy-to-deploy, safe alternatives to industrial robots, but faced real-world limitations like slow speed, low payload, and integration complexity.
Technical drawbacks include slower cycle times (prioritizing safety), reduced effective payload (especially at reach), and reliability issues in harsh shop floor environments.
Misapplication and overpromising marketing led to cobots being used for unsuitable tasks, while integration challenges emerged when moving from perfect demo conditions to imperfect job shops.
Despite these issues, cobots can still provide ROI in appropriate applications, particularly for lights-out operations where consistent, slow work outperforms human labor.
Summary:
This podcast episode investigates the gap between the initial hype around collaborative robots (cobots) and their actual adoption in machine shops. The hosts set the scene in 2015, when Universal Robots and others promised cobots would revolutionize manufacturing by being safe, easy to deploy, and capable of running alongside humans without cages. Forecasts predicted a $4 billion market by 2023, but reality fell short.
The discussion identifies several "suspects" for this failure. First, marketing overpromised, suggesting cobots could replace workers immediately and required no programming—claims that were only true for simple tasks. Second, technical limitations emerged: cobots are intentionally slower for safety, have low payload capacities (often only 60% of advertised when reaching into machines), and suffer from reliability issues in dirty, humid shop environments.
Third, misapplication occurred when shops used cobots for heavy-duty machining tasks better suited to industrial robots. Finally, integration complexity proved a major hurdle. While cobots worked in controlled demos, real-world job shops revealed payload constraints, speed mismatches with high-speed machining centers, and the need for skilled staff to troubleshoot issues.
The hosts conclude that cobots are not a universal solution but can succeed in specific, appropriate applications—particularly for lights-out operations where consistent, slow work yields significant gains. The dream wasn't entirely killed, but it was tempered by practical realities.
FAQs
The best automation is one that actually works in your shop and makes you money, not necessarily the flashiest or trendiest solution.
Co-bots were promised to democratize automation, be easy to deploy, and solve the labor crisis by working alongside humans without safety guarding.
Co-bots were slower and weaker by design for safety, had payload limitations, and integration was often more complex than expected, leading to slower adoption.
Co-bots are slow compared to industrial robots, have limited payload capacity (often only 60% of advertised in real applications), and can be unreliable on the shop floor.
Marketing overpromise, technical limitations like payload and speed, misapplication in unsuitable tasks, and integration complexity in real job shop environments all played a role.
Yes, co-bots can provide great ROI in specific applications where their slow speed and lower payload are acceptable, such as light-duty tasks running multiple shifts.
Chat with AI
Loading...
Pro features
Go deeper with this episode
Unlock creator-grade tools that turn any transcript into show notes and subtitle files.