How Robotic Technologies Tackle Material Handling Challenges in Warehouses

     

    Tune in to our fireside chat with Plus One Robotics CEO & Co-founder, Erik Nieves were we dug into the current demand for robotics in warehouses - and some of the obstacles Plus One is working to overcome when it comes to the complex process of picking and packing packages.


    Webinar Transcript:

    Jeremie Capron:

    My name is Jeremie Capron. I'm the director of research here at ROBO Global. And today we are going to focus on one of the most exciting areas of robotics and AI, something that our research team is very passionate about. That is logistics and warehouse automation.

    And we believe that this is a very important topic for investors, not only because it has already started delivering impressive investment returns, but also because it touches on so many important aspects of the robotics revolution. And we are going to be talking about some of the enabling technologies that make logistics automation possible, like the sensing, the computing, the AI. And also about what's happening in terms of businesses adopting this technology across a large spectrum of industries.

    And to do that, I'm really thrilled to be joined today by a very special guest. Erik Nieves is a founder and the CEO of Plus One Robotics, which is one of the most advanced robotics and computer vision companies in the field of logistics today. And before founding Plus One, a few years ago, Erik was technology director at Yaskawa, which is one of the most successful industrial robotics companies in the world, where he worked for, I believe nearly 20 years.

     

    Erik Nieves:

    25.

     

    Jeremie Capron:

    I think Erik is a thought leader in robotics and he has a very deep technical expertise, but also I think Erik is a remarkable communicator. And over the 10 years or so that we've known each other, I've learned a tremendous amount from Erik as we researched and invest in the world of robotics. So I hope today you can also learn something, and pretty sure you will. So for that, I'm very grateful. So Erik, thank you and welcome. And where does this webcast find you today?

     

    Erik Nieves:

    Well, good morning. You are beamed into San Antonio, Texas, which is where Plus One is headquartered. And I've been looking forward to this webinar for some time. I appreciate the opportunity to speak to the community on just sort of where we are with warehouse automation and where this all ends. So looking forward to it this morning. Thanks, Jeremie.

     

    Jeremie Capron:

    Great. So before we start drilling into it, I wanted to show participants how we see logistics as a piece of the investment thesis around the robotics revolution. And as some of you know, our goal at ROBO Global is really to provide exposure to the best in class companies from around the world across the entire value chain.

    So we're not only looking at robot manufacturers or providers of turnkey solution for the manufacturing applications and so on, but really we're looking at a broad range of very attractive application domains of which logistics is a very significant one. And you can see on the left hand side here, we're also providing exposure to the enablers, the technologies that make robots and autonomous systems possible, like the sensing, like the actuation, the computing and so on. And overall, logistics and warehouse automation represents about 13% of the ROBO index portfolio today.

    And so that is one of the most important sectors in terms of exposure in the ROBO portfolio. And if you look at historical performance, here's a chart that my colleague Brad, put together about a week ago that displays the performance of the ROBO index in blue. And you can see since the inception, almost 10 years ago now, the returns that reached almost 300% at the highs of last year and then major pullback so far this year. But you can see an orange, the logistics and warehouse automation subsector has been really outperforming quite significantly and also consistently.

    And of course, during the pandemic with so significant enthusiasm around the booming e-commerce and how those companies could help with supply chain issues and the tremendous volumes going through supply chains. You can also see the significant pullback, which we view as a major opportunity. The sector has pulled back more than the 35% from the highest reach last year. So pardon me for this side note here. I want to go straight into Plus One and ask Erik about the genesis of Plus One. After your long career at Yaskawa, why did you decide to start a new company?

     

    Erik Nieves:

    Well, I was at Yaskawa for about 25 years. I joined them in 1990. And the robotics industry was very indexed to the automotive sector, and we'll actually see that in some of the data. But around 2010, really 2008 when the crisis was as bad as it was, the robot industry was really struggling because it was so tightly coupled to the automotive sector.

    And I was tasked along with my colleagues on the marketing side, "Please go find something else for robots to do that wasn't tied to a car part of some sort." We kissed a lot of frogs. I mean, we looked at aerospace. If you can automate the building of cars, why not automate the building of airframes. And you can, it's just not that many robots. The long life cycle of the product kind of keeps you from that. We looked at clinical lab automation, nobody wanted to be handling blood samples anymore.

    But all of these were sort of niche. And in the end there were only two sectors that really mattered that were going to adopt robotics at scale. And that was electronics assembly and supply chain. And you could argue that electronics assembly is a bigger opportunity and it's more akin to what robot people are used to dealing with. It's high volume, it's repeatability, it's precision, it's engineers talking to engineers.

    But from our perspective it didn't much matter because it was going to stay in Southeast Asia no matter what you did. And I still argue that that's true. So logistics though is by definition always local. And so I spent some time even within Yaskawa trying to pivot a hundred year old Japanese industrial to building products that were going to be relevant in the warehouse.

    In the end, we all learned that the warehouse is not really a robot problem. It's a perception and grasping problem. It's vision and it's actuation at the end of factor. The arm is the arm is the arm. So I decided to separate from Yaskawa in 2015 and established Plus One as a result.

     

    Jeremie Capron:

    I go back to this chart I showed earlier that really kind of highlight the boom in the logistics automation business, the companies that are represented in that subsector of the ROBO Global portfolio are the likes of Zebra Technologies and Daifuku in Japan, and more recently GXO Logistics. And some much smaller less known companies that are in Switzerland like Kardex, the automated storage retrieval systems or-

     

    Erik Nieves:

    AutoStore.

     

    Jeremie Capron:

    ... AutoStore that went public last year. So what is it that really triggered that inflection in terms of the adoption of automation technology by the logistics industry?

     

    Erik Nieves:

    The main inflection point obviously is our continued push of more of our commerce going online. In a sense, you can think of, we were the robots doing the order fulfillment ourselves. When we would go to the big-box store to the grocery or what have you, we did our own picking, we did our own sortation, and we did our own delivery. And e-com changed that. We just started click to ship, and that meant somebody else was having to pick our orders and pack them and deliver them to our doorstep.

    So that was already happening. COVID just accelerated the inevitable, where before grandma would still go to the grocery store because she didn't want to have to mess with a computer to get her delivery. She figured it out. And COVID is, we could say, now come and gone. Grandma's not going back to the store to buy a 40 pound sack of dog food and have to lug it into her trunk. That's never going to happen.

    So we have brought on a whole new class of online consumers and there's just no stepping back from that. The person that's going to be lugging that 40 pound sack of dog food from now and forever, is the FedEx or UPS delivery person that's showing up on your porch. And so all of that conspired to seeing this real spike in volumes, at the same time as we were so constricted in labor. And the growth of the market, outstripped labor's availability and the result is burgeoning market for automation in the warehouse.

     

    Jeremie Capron:

    And what about the technical side and the technology? Because I go back 10 years ago and visiting the trade shows like the IMTS in Chicago where I'm at today, and seeing very impressive high speed, high accuracy industrial robots that were impressive by their such performance. But at the end of the day were very dumb robots and unable to do anything out of the ordinary of the industrial or the production plan. Meaning, that if there was a change in the part that was supposed to be picked or even just location by a few millimeters, the robot was stuck. And today, we've seen some tremendous advances to the point where the holy grail of robotics from 10 years ago, which was a random object picking capability, we now have that. So what's happened that's enabled that over the last few years?

     

    Erik Nieves:

    I would say this is maybe instructive here. So this sort of speaks to the dichotomy that you're referring to. We said that robotics was tied to the automotive sector and you can see that in this data where the dark blue is automotive and light is non-automotive. Starting 2018, the automotive industry was effectively flat. But the non-automotive orders continued to grow.

    Now, the automotive orders are the ones that you're thinking about, where the robot is just strong and dumb. You don't need computer vision, you don't need AI to spot weld the car body. And we've been successfully doing that now for 40 years. The good thing is, spot welding demands a lot of robots. It's 200, 300 robots in a facility to do that type of work. But it was also possible because the robot didn't really need to know much more than its own repeatability and endurance. It could lug that 200 pound gun all day every day to within a half a millimeter repeatability and that's what the process required. So there was no sensing.

    But get outside of automotive, you can't afford to structure the world this way. The value to volume curve means you're going to have to deal with things as they are, not as you ought. The automotive sector I would argue is one of the very few that has enough value and volume to make it a structured environment. You're going to do the same thing the same way for the next four to eight years.

    But in non-automotive, and that light blue line, yes, that's robots at Flippy doing burgers and fries. It's some construction robots, but a lot of it is in that 13% you talked about which is in logistics automation. And in that space, the repeatability is not the rule. Variability is the rule. I don't know what's going to come down the line next, and I sure don't know what's going to come down the line next week. So it's predicated on variability. That's part of the reason it has taken such a long time for warehouse automation to really light up, because to do it you needed sensing.

    In manufacturing the use of sensors is limited, the use of data is even more limited. But in warehouse automation, you need a lot of sophisticated sensors, these are now 3D sensors, and you need the data to drive the AI learning. So you have to generate that in a pretty much on every cycle. And then the last thing that I say about the constraints that have historically been on warehouse automation, is it's been an in a legacy market that hasn't had a lot of automation tools to bear.

    What does that mean? They don't have any technicians or engineers on staff to run the type of equipment that manufacturing has been using for a long time. So those manufacturing and warehouse automation are not the same thing. And the differences are these. The main one is variability, which means you have to have sensing, and it's a legacy market that just hasn't had the right expertise. But the reason it's now 13% of the ROBO portfolio is because the acceleration that we're seeing.

    Still today, only about 10% of operations are automated at all. And that does not even full automation, that might be semi-automated. Given the growth, there's a million jobs left hanging in three years time. There is not a warehouse in America that doesn't have a "help wanted" sign outside. All of them are short-staffed and their churn is the highest out there. It's fast food churn.

    The attrition rates at Amazon last year were nearly 160%, which means if you have a task that needs doing, you're going to have to find bodies for that task. Not for one shift, not for two shifts, but probably for six shifts because that job's going to churn that many times. So where does that leave you? It leaves you looking at automation as your path forward.

     

    Jeremie Capron:

    Can I interrupt for a second here?

     

    Erik Nieves:

    Yeah.

     

    Jeremie Capron:

    I think the staffing is a huge issue for sure. There's the attrition problem, but there's also the difficulty of the job itself. And it's something that I really like about Plus One is your motto, which I believe is, robots work and people rule.

     

    Erik Nieves:

    That's right.

     

    Jeremie Capron:

    So I'm curious, if you could just touch on that for a minute. When you automate a warehouse, what exactly happens to the structure of employment within the business of a warehouse?

     

    Erik Nieves:

    Yeah. You're right that given what we just talked about, that they're all short-staffed, this is not a zero-sum game. It's not robot in, person out. No robot ever deployed in a warehouse resulted in a pink slip. We don't have enough people showing up at the front door anyway. So really what automation is, is a leverage play so that the employees you do have, those FTEs can be even more productive and valuable.

     

    Erik Nieves:

    Think of it like this. And we'll see some of an example of our work at FedEx. If my job is to move packages from right to left, 25 to 30 times a minute for the next six hours, that's not a desirable task. But it's one that it's a job that needs doing. So if instead we can say to one of those operators, "Hey, look, where before you were responsible for 1,500 picks an hour, I'm going to make you responsible for 6,000 picks an hour."

    At first their hair would go on fire because they know they can't do that. But when you tell them, "And the way we're going to do that is I'm making you responsible for this line of robots now that's going to be picking at 1,500 an hour." Warehouse automation is a leverage play, not a labor substitution play. So everybody that showed up still gets to stay, they just become more productive in the tasks that they were responsible for by leveraging the tools, and they move up scale in value to the organization. They usually get a pay raise out of the thing. So it's not a bone of contention when the robots show up in any of these facilities.

     

    Jeremie Capron:

    Got it.

     

    Erik Nieves:

    Yeah. And that's the reason we say robots work, people rule. So you're going to find these robots, they're either going to be doing picking or they're going to be palletizing of some kind. But of course the biggest one and a substantial portion of the value so far in the space is these mobile collaborative robots, the AMRs.

    And there's two classes of AMRs. You have these, which Locus is a clear leader. And then you have Shopify acquiring 6 River. And this one, which is MiR that was acquired by Teradyne. And Fetch, which was acquired by Zebra. In the interest of full disclosure, you mentioned Zebra earlier. They're an investor in Plus One ClearPath. All of these robots are effectively looking at the mobility problem in the warehouse. And trying to either eliminate the number of folks that have to be responsible for driving forklift or hauling carts or something, and then letting them do more value added tasks.

    On the left, that was 22,000 AMRs sold, or deployed rather, in the US last year. That didn't count a single one of these robots on the right, because these weren't sold. These are Amazon's proprietary AMRs. And there were more of these deployed than the entirety of the ones on the left. And that's because Amazon just adopts automation at a different pace than anyone else. And that kind of speaks to this market.

    We talk about the market is top heavy in warehouse automation. There's really three different levels. You've got a level here, which is sort of where the DHL and FedEx exist. They're adopting robotics in their operations, they're scaling them out, et cetera. They deploy robots by the robot arms 20 and some odd at a time. DHL deploys Locus robots a thousand at a time. Then you have sort of down market. Think about the 3PLS, direct to consumer brands, et cetera. These folks are buying robots by each or the pair. So you can say that the market is top heavy. But then you have Amazon that's above both of those that's deploying robots by the hundreds, and AMRs by the thousands themselves. So it really is sort of a three-tiered market where Amazon is sui generis, they're their own thing. Any thoughts, questions on that?

     

    Jeremie Capron:

    No, I think it's really interesting that Amazon itself essentially builds and ships and deploys more robots that the entire independent AMR industry. And yet the AMR business would believe is a fantastic one now, because the pie is growing so fast. And of course Amazon has a lot of competition, and it raises the bar for everybody else to essentially mimic or reach the same type of productivity and efficacy levels in their warehouses and supply chains. So at the end of the day, I think the AMR business should be a pretty good one for the years to come and would you apply on that?

     

    Erik Nieves:

    I agree. If you are in a warehouse, the tasks to be done, the labor in the building, can be thought of as three different types of labor. First is the mobility labor. These are folks that are either driving a forklift, or they're pushing a dolly, or they're hauling a cart or something associated with a wheel. So that's one class of labor. The second class of labor is people that are walking, the ambulatory class of labor. And these may be the ones walking up and down shelf stacks or what have you. And then the third class is static labor. These are people that are at a station of some kind, maybe they're at a packing station, maybe they're at an induction lane, et cetera.

    The mobile class of labor is the smallest, but it's also the one that is accelerated the furthest in its adoption of robotics. And there's a couple of reasons for that. One of them is the physics of the problem is sort of easier. A mobile robot has two and a half degrees of freedom, forward, back, left, right, turn. And it's always wheels on a flat floor. You kind of understand the morphology of the robot. Now you've got agility and some folks trying to put legs on robots. We'll see where that goes.

    But the preponderance are going to be wheels. The AMR problem technically is really just the sort of coordination among the fleet of these robots, sort of deconflicting them, and charging. You have to do opportunity charging and that type of thing. But AMRs lend themselves to the warehouse right now because one, it's a two-way door. I can put AMRs in and if they work great, my efficiency goes up. If they don't work, I'll park them for a minute and I'll just have people go back to doing what they were doing. So the sort of catastrophe gap is minimized on AMRs.

    Second, AMRs, almost from the very beginning, lent themselves to an OpEx model. You don't buy a Locus robot, you hire it. It's on some sort of performance OpEx model and that made the ROI easy for these folks to deploy. So I do believe that AMRs will continue to lead the pace in adoption in the market. But it is true that the manipulation class, people whether they're walking or staying in a station that have stuff in their hands, is by far the bigger class of labor in the warehouse.

    But it's a harder problem. It's not as easily a two-way door and it's more CapEx intensive. There's more stuff to deliver. And so it hasn't classically lent itself to OpEx models as readily as AMRs have. We will see that change over time. But today that is a truism, and AMRs are going to lead the pace certainly for a while and we applaud that.

     

    Jeremie Capron:

    So I think we're going to move on to the manipulation side and what Plus One is solving. But I'd like to highlight that, you look back a decade ago which is when Amazon acquired Kiva, which was the foundation of their AMR solution for their distributions centers. It's taken the enabling technologies to reach the level where you could make these type of autonomy possible on the ground floor.

    It took a while, and when you think about what those technologies are, the sensing, the motion control, and the motion planning, the computing, and as you said, the fleet management. All of these essentially converged about 10 years ago. Is it fair to say that now the enabling technologies for manipulation are at this point where they've reached the level of performance capabilities and a low enough price point that now we can move on to automating and manipulation side?

     

    Erik Nieves:

    Yes, we can. But before we do that, because you mentioned the Kiva acquisition, I should tell you, we looked at these markets and the way that they're structured. The warehouse automation market is very leery of [that happening] again. So what that means is no one major user is going to be single-threaded. They will have multiple suppliers of the technologies that you see, whether it's AMRs or manipulation. They're not going to be single-threaded because they don't want to be fired someday if their supplier of choice gets gobbled up or rolled up. And that plays out.

    I mean you see now, these big players will require escrow agreements or something to sort of satisfy themselves that they have put themselves in a position to continue to be successful in the face of [that kind of] acquisition again. So here we are 10 years later, Jeremie. And that transaction is still having ripple effects in this industry, right?

     

    Jeremie Capron:

    Right.

     

    Erik Nieves:

    But as for manipulation, yes. We said it was a perception and grasping problem. If I pick up my phone off of my desk, I need my eyes, my arm, and my hand in that order. And of the three, the arm is the only one that's an engineered commodity you can just go buy. I came from that world. That's what Yaskawa, and Fanuc, and ABB, and KUKA, and Universal. That's their world. So what was needed was low cost, reliable 3D sensing and grasping.

    Now, on the sensing side, so we'll just jump up here. This is, and I'll stop this so you can get a feel for what's happening here. Okay. So these are robots at FedEx. So these are robots deployed in Memphis. This was a job of a person at a station, and their life was get all those packages out from the right and put them on the left conveyor. And had to do that 25 to 28 times a minute for the next six hours.

    People don't want to do this anymore, high churn job. So bring automation to bear. But you can see the arm is just a standard industrial robot. That's a typical Yaskawa robot you would see in an automotive plant. But what's different is these sensors here. So this is a 3D low-cost, high-fidelity 3D sensor. Plus One didn't make that sensor, Intel did. Those are the Intel RealSense cameras and they're used a lot in AMRs and in picking robots like ours.

    So the technology was the sensor and the AI that's evaluating the sensor data. And that's what you see happening here, is the robot has to take and pick and place. And what the vision system is doing is evaluating that entire pick bowl as we call it, looking at which parcels are available for pick, meaning not occluded and that type of thing. And then its job is just going to be to put them out one at a time. But look at this, because this happens on occasion. I'll stop it here.

    About a half a percent of the time, the vision system will see a scene and not understand what it's supposed to do. And here's an example of that. It's looking and it says, "You know what? I don't see any single parcel that's completely unoccluded. What am I supposed to do here?" And it does not pass confidence threshold.

    The AI says, "I'm not sure." When that happens, the robot raises its hand over the cloud. It knows enough to phone a friend, and that friend happens to be in San Antonio. And a human being, what we call a crew chief, then takes, and with their mouse tells the robot, "Yeah, go pick up this one. I know it's occluded, but you can pick it up there." Same scenario here. It's like, where does one package end and another begin?

    This is the type of thing that supervised autonomy lends itself to. And we'll look at this. This is an AI torture chamber. They're all flat, similar, shiny, reflective, occluded. The AI is going to fail. And when it does... Jeremie, you immediately knew which one the robot should go get. So the human tells it, "Go pick up that one from there." And that is the notion of supervised autonomy. These are the technologies that have come to bear. It's the sensing to generate the point clouds, the representation of the world. It's the AI to sort of determine from those images what's valid and what isn't. And then it's the grasping.

    Now, for our world, vacuum is enough. Because I deal with parcels. If I were doing each's out of an auto store, it wouldn't work and I would need a different type of gripper. One that articulates, kind of like RightHand Robotics or Soft or somebody else. But that's the notion of the technologies coming together at the right time. This is just another example. This is a standard robot with a standard vacuum gripper, but its 3D sensing allows it to do picking and placing basically depalletizing real time.

     

    Jeremie Capron:

    And so tell us about your customers that are now deploying those systems. I think they went through an experimentation phase. And it seems to me, just looking at some of the major announcements in the market, that the biggest players, you mentioned FedEx, UPS, DHL over in Europe, and potentially postal services around the world. Where are we at in terms of the adoption? Are you seeing any acceleration here?

     

    Erik Nieves:

    For sure.

     

    Jeremie Capron:

    And also would love for you to talk about how they think about the return on investment when they acquire a new tech.

     

    Erik Nieves:

    Yeah, happy to. To give you just a sense of scale, this is a live view of the picks happening in the world as we speak, using Plus One's technology across the globe. So where it's over half a million picks so far, by the end of the day to day, there will be two commas in that number. It's over a million picks a day. And we're really excited at Plus One. Because sometime in October, we're going to surpass a half a billion picks all time.

    And this is just far and away, the leader in actual production. And the reason for that is exactly what you're saying. It's that the top of this market is rolling out this technology. So I can tell you, FedEx is public about the work they do with us. They had a vision partner, it wasn't working. Plus One came in, the human-in-the-loop made a huge difference. We deployed four robots there initially, and now it's way more than that.

    It's the same in e-com and it's the same across all parcels. What'll be interesting, and this is something that will happen I expect before this calendar year is done, is the US Postal Service. Because the postal service put out an RFI for 2,000 systems, and since then put out an RFP for a portion of that for about 400 robots. That decision is due anytime now. And that will immediately drive big acceleration in the number of deployed systems globally. So we're eager. At Plus One, we're confident that we're in a good place to win some of that business. We'll see if they'll be multi-threaded the way others are, or if it will be single sourced.

    But either way, the adoption rate of robots in parcel handling is definitely up into the right. And it has to be, because there's not enough labor available. Think about Memphis and Louisville, the two busiest airports in America overnight. Well, that means they need to have a ton of people to come handle all of that volume. And there's just not enough rooftops in these metropolitan areas. Today, FedEx is busing people in from Little Rock, Arkansas to do sorts at night. That's two hours each way just to be able to sort packages. It's not sustainable. Automation they understand is their only path forward.

     

    Jeremie Capron:

    Got it. I want to jump in with maybe a little bit of a different question. Because as you know, our community here on the webcast is primarily people in investment, management, business. I'm curious, now that you've raised I think over 40 million today at Plus One, so you've interacted with the investment community. For a few years, what is your impression of how investors approach the world of robotics and AI? Anything that surprised you perhaps?

     

    Erik Nieves:

    Sure.

     

    Jeremie Capron:

    Anything that investors should think more about?

     

    Erik Nieves:

    Well, the early questions back in 2017 when we were first starting was, "Hey, is this a big enough problem? What's the TAM associated with warehouse automation? Or at least the different applications within the space." Those questions are largely behind us now. Everybody has sort of reconciled themselves to, logistics is more than just trucking. And that within the four walls of the warehouse there is great opportunity. And so that's not the constraint it once was.

    The questions that you get now are really about business model, Jeremie. Because the investment community, a lot of them and particularly in the growth stage funds, are really indexed to SaaS metrics. And so the business model that they prefer, that they give the best valuation, multiple on, et cetera, is what's your recurring revenue. Well, that's pretty orthogonal to a lot of the way that automation has classically happened. Even within warehouse automation, T-MATIC doesn't sell you a system on a performance basis and you buy productivity. And you have this ongoing relationship with them on a month to month basis.

    No, it has been, like all system integration, classically it's in and out in 26 or 30 weeks, and here's progress payments along the way. I'll see you when you have your next project. But that's not "SaaS y". That's not recurring. Even if that same customer comes back a year from now and says, "Okay, now I'm ready for the next facility." That selling motion to the investment community, feels bespoke. And so they don't really give you credit for the same customer coming back and buying again. They want that to be contracted recurring revenue.

    It's going to take some time before this market does that. I'm not sure frankly that it ever will completely. And again, it goes back to the market being bifurcated. If you're at the top of this market, chances are you cash flush and you are looking for places to take depreciation schedules left, right, and center. And an OpEx model is not in your interests. Down market, when you have the 3PLs and the DTCs, they are more capital constrained and the CapEx dollars are going to be used for their network, facilities, et cetera. And they don't want to tie it up with equipment.

    They are more open to RaaS, robots as a service, or some sort of operating expense model. But this market, I would argue is never going to pivot to look like a SaaS world. You are still going to have it be sort of this two-headed beast. And at least from Plus One's perspective, you better be prepared to engage with both of them, or you're just cutting out two substantial apportion of the market. So to the investors, I would say, be thinking about whether SaaS is the right way to think about warehouse automation or not.

     

    Jeremie Capron:

    Well, there's no denying that the SaaS model provides answers to a lot of the uncertainty to the business of investing, and it makes investors' jobs a lot easier. But I think we also went through a phase of almost exuberance around the SaaS models, where we saw evaluations of anything SaaSy as you described it. Explode on the way up, and more recently explode on the way down. And so I think there was really exuberance around that.

    I think also the robotics as a service model that you described, probably has a lot of potential perhaps starting by some of the pieces of the services that you offer. You mentioned the supervision of robots by-

     

    Erik Nieves:

    Crew chief.

     

    Jeremie Capron:

    ... crew chief that would certainly lend itself to such a model. So I'm curious if you're seeing any other, perhaps parts of the problems you're solving through your customers that could follow such a model.

     

    Erik Nieves:

    Sure. So again, our users will typically buy the vision system and that'll be a CapEx deal. Because in their view, the eyes of the robot are an extension of the robot, and the robot was CapEx. So the eyes should likewise be CapEx. And we'll see you next year for 18% maintenance on that software. So it's a typical enterprise software transaction.

    But the human-in-the-loop piece, that is a subscription. Always has been. So that's the recurring revenue piece for Plus One, is folks subscribing to the uptime that Yonder, that's what we call the human-in-the-loop service, provides. And so that will continue to be an important part of our business model.

    So there's recurring just from, "Here's a portion of our deliverable." I like to think of Yonder as a managed service, and that managed service as a subscription. Then you've got the folks that just buy the whole solution, robot, gripper, safety, installation, all of that as a service. So there's your RaaS model. That's going to be a portion of the market. And Yonder, the human-in-the-loop is included in that performance contract. And then you're going to have the folks that are just straight CapEx.

    It is true. We do have installations where they are using our vision system and don't need the human-in-the-loop. It has to do with the variability of the input stream, et cetera. So in that case, it's CapEx and I'll see you next year for the maintenance agreement. And that's the recurring piece.

     

    Jeremie Capron:

    Got it.

     

    Erik Nieves:

    Yep.

     

    Jeremie Capron:

    Okay. Well, I think it's time for us to open the call to questions from the audience. So feel free to use the box at the bottom of your screen, the Q and A box, to send over your questions. I see that we have one already that's come in around the percentage of potentially economically valuable warehouse automation that has been achieved so far. I think you touched on that a little bit, there's various estimates out there in the market. What do you think, Erik?

     

    Erik Nieves:

    There's a lot of headroom to go in this business. Warehouse automation is a nascent industry, I would argue. So for that million plus picks a day, that doesn't scratch the surface of the amount of picks that happen in the world. So I'll give you a sense of that here.

    We talked about the three classes of labor, the people on wheels, the people that are running up and down the shelves, and then the people that are standing still. This is them. So this is nearly a million and a half FTEs that are at a station in warehouses in US and Western Europe.

    This pie chart is what they are doing. 46% of them are loading some kind of conveyor, 14% of them are breaking pallets apart, 13% of them are building pallets, 13% are doing what are called packout operations, and there's the 10% that's G2P, goods-to-person picking, Alloy, Amazon Kiva. That's only 10%. And then the 4%, which is going to continue to grow, is the returns processing.

    So there's a million and a half people doing this right now, and there's only a million picks in the world. That tells you there is a lot to go. For us in specific, we do parcel, we do induction onto the conveyor, and we do depalletizing. And that's about 60% available application space to the type of work that we do. But that should give you a sense of scale as to where the TAM is yet to go in this space.

     

    Jeremie Capron:

    I see. We have a comment from Michael around the RaaS model. And I think it's a fair point that he raises that RaaS can be very attractive to the user in terms of not having stranded cost risk in a downturn situation.

     

    Erik Nieves:

    Sure.

     

    Jeremie Capron:

    I think it goes back to the beginning of a conversation and whether robotics industry first flourished or in the auto industry, which is as cyclical as it can be. And where it's true that the very first signs of slow down and demand, the CapEx budget were just trashed and very quickly robotics companies felt the pain. What do you think about that, Erik?

     

    Erik Nieves:

    A hundred percent true that robotics historically, automotive gets a cold, robotics gets a flu. It is a leading indicator of softness in the robotics market for everything you just talked about, Jeremie. So yes, that was part of the impetus for us at Yaskawa trying to find something countercyclical to the automotive space.

    The comment is right that RaaS lowers a lot of barrier to entry. This is why it's working so well for AMRs. Because hey, if in the end those robots worked well for me from January through August and now I know I'm going to surge, I'm going to call the AMR provider and say, "Hey, I need another dozen robots to show up next week. Add them to my bill. And great."

    And the flip side, come February. February now, not January. Because where the surge used to end at the end of December, that's no longer true, because January returns again. You have to deal with all that. But come February, that same operator can call up their AMR of choice and say, "Hey, thanks. Those worked great. I'm back now to steady state volume. Come get your robots." So yes, the OpEx approach has real benefits and such. I'm just telling you, not all users want it.

    And there is a difference between AMRs and manipulation systems. Just think of the COGS. Think of the COGS involved in an AMR versus the COGS of a conveyor induction system with a robot and conveyors and safety and all this other stuff. It's just more stuff. And if you're going to subscribe to all of that, then somebody's got to finance that. That's a consideration also. But again, down market, the 3PLs do the whole thing on RaaS. But I'm telling you, I could not walk into Memphis today and say, "I will only do conveyor induction for you on an OpEx model." They will tell us, "We won't be doing that."

     

    Jeremie Capron:

    Mm-hmm. Well, as we look across the various applications of robotics and automation, and I'm thinking about the healthcare sector here, where we've seen meaningful pickup in terms of the leasing model for things like big expensive surgical robots, but also less expensive solutions around pharmacy automation for example. Clearly the hospitals also appreciate the OpEx model. So that's just one thought.

    And another one is that certainly right now, if you look at the price action in the public equity market and particularly the one that I showed you around logistics automation sector that's really been bound over the last six to nine months, I think it's clear that investors will just rush out of those companies at the first signs of potential downturn or recession. And here we are with our interest rate curve inverted, and the typical behavior has happened again and those stocks have been sold off. I just want to point out that historically, those have been the best times to get involved. I see we have one more question. It comes from Dean. How many FTE is forklift drivers in the US?

     

    Erik Nieves:

    Ooh.

     

    Jeremie Capron:

    That's a good question. I don't have the answer on the top of my head.

     

    Erik Nieves:

    I don't either. But boy, there are sure a lot of folks trying to tackle that problem. So all the big MHE companies are trying to do automated forklift of some kind. Seegrid is doing automated pallet jacks. Fox Robotics, automated fork trucks. Phantom doing effectively Plus One for forklifts, meaning they have a remote driver. So there's no lack of people trying to tackle that situation. And not just within the four walls of the warehouse, because the marshaling yard also presents itself as an opportunity for automated driving. So think of the marshaling trucks that move the 53 foot dry vans around. There's companies Outrider, ASCE, et cetera, that are trying to automate that as sort of a stepping stone to automated, autonomous vehicles on public roads.

     

    Jeremie Capron:

    And I see a note from Aaron. Thank you, Aaron. That according to the Bureau of Labor Statistics, 700,000 people are employed as material moving machine operators last year. So that's definitely is a large pool. And I know that on top of the companies you mentioned, Erik, the big players in the forklift market of course are paying a lot of attention. So I'm thinking about one of our portfolio companies, KION that owns T-MATIC, but also have a huge... I think is the second largest truck lift provider in the world. They're working really hard on this problem. And about five years ago or so, they really bifurcated the offerings to include electric forklift. So you have the electrification that's already well on the way, and the next step is probably the autonomy.

     

    Erik Nieves:

    Mm-hmm. Agreed.

     

    Jeremie Capron:

    Okay. Well look, Erik, anything else you want to discuss or present today before we wrap it up? We're getting close to the end of the hour now.

     

    Erik Nieves:

    No, I would just say that everything that we've talked about here this morning, you can just lather, rinse, repeat, for construction, AgTech, et cetera. The central conceit that I make is that for manufacturing, AI's role is going to be limited. We've been successfully building cars for a long time. For outside of manufacturing, AI is an important tool and I argue insufficient. That the rate of change of the real world is such that you're going to have to have a human-in-the-loop.

    And that is the secret sauce, is AI plus supervised autonomy. And that's not just constrained to warehouse automation. So when I think about AgTech, when I think about construction tech, et cetera, those are the things that I'm looking at. I get asked regularly, "So what's the role of AI in these emerging markets and how should investors think about that?" And I guess I look at it differently.

    And that is I evaluate an industry for, would that industry be benefited by comprehensive immigration reform with a guestworker program? And if the answer is, yes, that would relieve a lot of labor constraint and allow the industry to grow, then you have to look at robotics. The Economist did a special report a number of years ago about robotics and called them immigrants from the future. And that's really the way that I look at it is, would this industry be benefited by immigrants? Because comprehensive immigration reform is further away from us today than it's ever been. So robotics is the only way forward, and that's true writ large across a number of sectors.

     

    Jeremie Capron:

    Thank you, Erik.

     

    Erik Nieves:

    You bet.

     

    Jeremie Capron:

    Well, it's time to wrap up here. And I want to remind everyone, if you want to learn more about investing in robotics, automation, AI, you can visit our website roboglobal.com. We share some of our research on companies and the ROBO, the THNQ, and the HTEC portfolios. So thank you very much, Erik, for sharing with us today. Good luck to Plus One and to you all who joined us today for this call. And we look forward to speaking with you again soon.

     

    Erik Nieves:

    Be well. Thank you, Jeremie. Take care all.

     

    Jeremie Capron:

    Bye.

    Back to All Entries