This is the full transcript of Jay Kapoor's SXSW 2025 panel on building Physical AI in dirty, dusty, and dangerous industries, with Alex Hawkinson (Founder/CEO, Bright AI), Rebecca Hu (co-founder, Glacier), and Theo Hooker (Cambium Carbon). This is original Dirty Jobs content. It has been lightly edited for readability from the live recording.

Jay Kapoor: Hey, thank you all so much for joining us for this discussion, cheekily named Dirty Jobs and the Warming Planet. My name is Jay Kapoor. I'm a general partner at VSC Ventures, and I have the immense pleasure of being your host today for this discussion with three amazing panelists, all of whom are startup founders working at the intersection of labor shortages and climate change.

Jay Kapoor: Before we dive into some of our prepared questions and then some Q&A that we'd love from the audience, I want to set the stage for you about where we stand today regarding climate adaptation and labor availability across dirty, dusty, and dangerous industries in America. So a quick show of hands, just so we get a sense of who's here with us today. How many of you are startup founders in the audience? How many of you are investors? How many of you work in these industries, from construction to operations to logistics? Okay, great. And how many of you believe that we urgently need to address climate change? All right, good. I'm in Texas. I have to make sure, you know? Just make sure we're in the right place.

Jay Kapoor: So a couple of interesting stats for all of you to set the agenda. By most recent count, we're about two and a half to two point six million workers short to keep America's growing needs in construction, manufacturing, logistics, and maintenance met. And the hourly cost of construction labor especially keeps going up, but the reliability keeps going down. If you've ever tried to build a house or get yours renovated, you know what I'm talking about. The average age of workers continues to go up. It peaks year after year. The average age of a telephone pole electrician is now fifty-nine years old, and the average age of a heavy machinery operator is fifty-three years old.

Jay Kapoor: We can disagree on the reason for this. People talk about educational choices and folks going to higher education versus vocational training. It could be a generational shift. Maybe it's just that backbreaking labor is not interesting to people anymore. But what's not in debate is that this shortage is not cyclical or seasonal. It's systemic, and it's generational. And yet the improvements that we're trying to make when it comes to energy security, material security, and climate adaptation will require a combination of human innovation and technology innovation. So how do we do it all? Well, to help me answer that question are three people way smarter than me.

Jay Kapoor: Please give a warm welcome to all of our panelists. I'll start with Alex Hawkinson, CEO of Bright AI, Rebecca Hu, CEO of Glacier, and Theo Hooker, COO of Cambium Carbon. Please give them a round of applause. And I am Jay Kapoor, general partner of VSC Ventures, where we invest up to a million dollars in startups innovating in dirty, dusty, and dangerous industries, and we help them break out with our expertise in brand, marketing, PR, and storytelling. So that's enough foundation laying for me. Let's get started.

Jay Kapoor: Alex, I want to start with you. Bright AI wants to transform traditional labor-intensive processes into efficient data-driven systems, and you're using sensors and artificial intelligence to do that. So lay the foundation for our audience. What exactly is happening to our critical infrastructure today, and how is Bright addressing some of the labor issues in that critical infrastructure?

Alex Hawkinson: Thanks for coming. Happy to be here and to participate in this discussion. It's a big topic area. In infrastructure, I like to think of it as not just traditional physical infrastructure, but all the essential services that we rely on every day, which are born outside of your home but we all rely on, whether it's water, power, energy, key services like HVAC, pest control, logistics, transportation, construction, all these things.

Alex Hawkinson: You have pressures on them as a result of a number of different systemic things. You obviously have climate-driven pressure that we'll get into in a bit. But in the labor markets, you have places where these are highly repetitive jobs in some cases. People don't like to have their time misused, so they'll go to some spot, they have to periodically inspect some part of the infrastructure. Nothing feels worse than doing something you feel isn't necessarily high utility. You have dirty, dangerous jobs where it's very physically intensive, and it's a big economy. There are other opportunities out there for these people. And I'm maybe the oldest panelist here, but these systems are also in some ways becoming more complex.

Alex Hawkinson: So to have the expertise to diagnose what's going on with a computer-driven machine or whatever, the learning curve can be quite brutal. Take HVAC as an example. It takes 25 years to become a master technician, and who has that time? So you have all these challenges, and if you look at the specific industries, take our water supply, which was built after World War II. It's crumbling now beneath our feet. Thirty percent of the fresh water that we treat leaches back out into the environment because of leaks in the pipes and other things like that. You cannot find the labor to go address that problem fast enough. So what you have to do is respect people's time. You have to bring automation to some of these dirty, dangerous, repetitive tasks.

Alex Hawkinson: You have to shorten the learning curve and give tools to the frontline workers so they can go after this stuff more efficiently. And it's such a critical need. If you look at American infrastructure, we give it a C-minus or D grade in all these different segments. So many people talk about job loss as a result of automation. We literally cannot get to the problems that need solving yet. So at Bright AI, I won't spend too much time talking about a physical AI platform, but we're trying to address those core things. We allow you to understand what's going on in the critical infrastructure across these different industries, so in real time you know when you need to send the person.

Alex Hawkinson: We do a lot in robotics, so we assist the humans in the places where the danger is involved. And we build co-pilots so that the frontline worker, somebody with less training, can get after it and do the work with the qualification of the best expert in the fleet on their shoulder, the angel on their shoulder, so to speak.

Jay Kapoor: That's great. We started touching on the idea of robotics and AI replacing labor, and Rebecca, this is something that gets talked about a lot when robotics is brought up. Is it taking dirty jobs? Is it replacing a human that's doing that job? What's the situation in the recycling industry today, and how are your customers thinking about your involvement when it comes to labor?

Rebecca Hu: Sure. Maybe before I get into that, a quick show of hands from the audience. Who here has actually physically been to a recycling facility, also known as a MRF? Nice. Okay. I'm liking the numbers I'm seeing. Before we start talking too much about recycling, and those of you who have been to a facility will immediately understand this, I actually think it's important to start the conversation around manufacturing in this country.

Rebecca Hu: When you think about where our manufacturing infrastructure is today, we've achieved incredible things. Think about plastic bottles. We're literally capable of pumping out billions of plastic bottles every single day, and each one is identical to the one before it. We've got insane quality control and inspection criteria. We're able to really tune in to make sure that all of the stuff we're producing is exactly to specification. Now, a lot of people think of recycling as a nice-to-have alternative to landfill as a method of disposal. And at Glacier, we're really thinking about recycling as another crucial pillar of manufacturing, because at the end of the day, you're making new feedstock that can get turned into new stuff.

Rebecca Hu: When you go into one of these recycling facilities, also known as a MRF or a materials recovery facility, imagine what I call the most demented form of manufacturing. You've got this commingled stream of recycling and waste material coming into a factory, and the job of this factory is to literally sort all of that mixed stuff into beautiful bales of aluminum, plastic, cardboard, metals, et cetera, that can be sold to manufacturers to make new stuff. That's how recycling works.

Rebecca Hu: But imagine now in these facilities literally dozens of people whose job it is to just stand there and sort stuff on a conveyor belt, "I Love Lucy" style, as quickly as they can. And this is not the stuff that you buy pristine off the grocery store shelf either. We're talking about really, really dirty material, torn, crumpled, mixed in with all kinds of crazy things. We hear stories about people finding ammunition, hand grenades, all kinds of crazy stuff coming through these recycling facilities. So when you think about dirty, dull, dangerous, this is the perfect storm. As a result, the injury rate among workers who are doing the sorting is actually twice as high as in construction, and there's a 500% year-over-year turnover. So if you're running one of these facilities, you have to hire for that same sorting workforce about five times per year.

Rebecca Hu: Now, against this backdrop, the requirements for recycling are increasing rapidly, both in terms of how much recycled stuff we need to capture and also the purity rate of that material. So against this backdrop of higher requirements and a dwindling workforce, this is where Glacier is really excited to come in and plug a crucial gap with our AI-enabled recycling robots. They're helping to not only automate this very challenging sortation task and upskill existing workers into more safe and trained jobs, but also provide really important data on what's coming through our recycling and waste streams, and feed that back to producers so they can make more circular packaging, and also to municipalities and other government agencies so they can actually understand what is in our waste and recycling streams and do something to improve upon it.

Jay Kapoor: That's great. The word that Rebecca's used with me in the past is gnarly, and it really does apply. When you go into one of these MRFs, if you haven't been in there, it's like Christmas lights and tennis shoes and half-eaten whatever, and people think stuff is recyclable that really is not. So gnarly is the word I would use.

Jay Kapoor: So Theo, we've talked about labor shortages here. What doesn't get talked about enough is the creation of new jobs. So I'd love to hear from you, as you introduce Cambium, how are you creating new jobs in the world of lumber, and what do your customers look for from you as you're addressing this labor concern?

Theo Hooker: I'll start at a high level, what we do and what informs what we do, and then I'll get down to the specifics. At the highest level, I think in the US and the West in general, we've done a really good job of exporting some of the dirty, some of the unsightly aspects of our supply chains, regardless of the ecological or social impacts of those things. Now we're seeing there are actually some benefits of bringing manufacturing back to the US and integrating those into more secure, more transparent supply chains. So at Cambium, our goal is to make it really easy to use material we normally waste. We do that in a few different ways, but one of the core elements is working across the supply chain.

Theo Hooker: So from the fiber coming into the processing, to the actual producers and wood products companies, to the buyers, the architects, the designers, the furniture companies, the developers. By working across that whole system, we're able to both increase the amount of feedstock coming in, creating jobs, enabling folks to process more material, which requires more hands, and also tell the story, educating and bringing awareness to this material as a viable replacement to tropical hardwoods, to plastic composites, to other things that maybe have a bigger ecological impact.

Theo Hooker: What we see on the job creation front is really that integrating technology, promoting awareness, and telling that whole story matters. Early on in our work, a lot of folks thought that was a disadvantage. We were told, "You're doing too much." As we've grown, and we're about five years old now, that has become a real defensibility component in the market. We can speak the language, from loggers and tree service companies doing wildfire mitigation on the West Coast, or recovering material from Hurricane Helene on the East Coast, to large, small, and medium-sized sawmills, and then to the architects, the designers, the buyers of that material. So each person in that supply chain plays a role, and we can speak to them to help grow the industry overall.

Jay Kapoor: So I talked a little bit about some of the stats around the why now, but one of the things I'd love for each of our panelists to talk about, and I'll start with you, Rebecca, is the why now of your technology. What necessitates Glacier today? What necessitates innovation in circular manufacturing today? And what technology is available today that maybe wasn't five to ten years ago that allows you to do what you're doing at the scale that you're doing it?

Rebecca Hu: Absolutely. One of the reasons we were really excited to start Glacier about five years ago was that there was this perfect storm of factors that enabled us to build Glacier as it currently exists. From a macro perspective, our recycling infrastructure in this country, while not perfect, has existed for about 50 odd years now. If you rewind to 2018, 2019, China, who was previously the world's largest buyer of recycled commodities, passed very rapidly something called the National Sword Act. It was basically because they said, "Hey, everyone's sending us recycling, which is actually just trash, and we're sick of landfilling or incinerating all of the trash we're receiving from around the world. So if you really want to do this, you're going to have to quite literally clean up your act and send us better material for us to do something with."

Rebecca Hu: So what that meant was that for a lot of these MRFs, these recycling facilities, they suddenly lost their biggest customer overnight, and they realized that the new way to engage in recycling is that you have to find a way to sort much better at much lower cost in order to stay in business. Now, this coincided with a huge technological explosion and new innovation that allows us to build things like AI tools and robotics to actually do such a crazy, gnarly, long-tail job. So Glacier has built a purpose-built robot that uses computer vision, so think AI cameras that sit above these conveyor belts, and they're literally tracking in real time every single item that's on that belt.

Rebecca Hu: So we can say that's an aluminum can, that's a toothpaste tube, that's a cardboard box, and then we can feed that information to our robot, which is sitting downstream. And this is a purpose-built robot. In other words, you can't just order this robot out of a catalog. We've made it to be really good at fitting into these facilities, being really robust, being really cheap, easy to install, all that good stuff. And then the robot can actually know, all right, if it's an aluminum can, I need to pick it up and put it here. If it's a plastic bottle, I'm going to put it here. So it's also doing that automation. And of course, we're collecting data on literally billions and billions of items to continue that flywheel effect.

Rebecca Hu: Now, I'm not going to linger too much on the innovations in AI because a lot of people here know that stuff is changing every single day. Suffice it to say that we're staying abreast of all of these innovations in AI that allow us to build a growing catalog of information about all of the stuff in the waste stream as it's evolving. But I also think what's remarkable is when you think about what's possible now with hardware and with robotics compared to a decade ago. There are huge advancements in the versatility and, frankly, the price point of off-the-shelf components. There are technologies like 3D printing, for example, that have made it much faster to innovate with shorter iteration cycles than traditional hardware manufacturing.

Rebecca Hu: All of that combined means that we were able to bring literally a purpose-built robot into one of the dirtiest environments out there in a matter of just about two years. One specific example I'll give is that our robots rely on suction. You can imagine almost like a vacuum that's sucking up a can or a bottle or whatever it might be. But in order to get that to not clog, you have to use basically these filters. So you're sucking, but you're not sucking too hard such that a plastic bag gets up there and gunks up your machine. Those filters, if we were using traditional manufacturing methods, would probably take at least one to three months for each new iteration that we were testing.

Rebecca Hu: But with our 3D printing technology, we can literally print out a new design in minutes. So over the past year, we've iterated our way through literally dozens of these designs to get to something that really works. Think about how long that would have taken a decade ago, literally years, to get to the same point we've been able to get to in less than a year. So we're really excited to continue to see that rate of innovation speeding up and applying increasingly into hardware spaces as well.

Jay Kapoor: I love that example, and I want to bring Theo into this on the AI technology front. The lumber industry is probably not the first place people think about software and AI. So why are they excited about innovation, and what's the why now of why AI in lumber makes sense?

Theo Hooker: Great question. I'll ask another question of the audience, following Rebecca's model. How many of you have thought about how to get food quicker to your door so you don't have to leave your couch or your bed? Probably a lot. There was a need, and then people spent a lot of time addressing it. How can restaurants talk to delivery drivers to get you what you want? And I think the why now is that no one has really thought about how we can get wood that has fallen from a storm or from a wildfire clearing into the supply chain.

Theo Hooker: No one has sat down with loggers, with truckers, with sawmills and said, "Hey, how can we make your job easier?" So when we think about it, the lumber industry gets labeled typically as sort of tech-averse, or not native, and my pushback to that is, well, maybe. But also no one has really spent the time to sit down and ask them some questions. So again, looking at the entire supply chain, we provide services to get material across all the players, and we offer physical products. So CLT, cross-laminated timber for mass timber buildings, so low, medium, and in some cases higher rise buildings, structural and non-structural, and then millwork components, so your flooring, accent walls, ceilings, decking, fencing, you name it.

Theo Hooker: On the software side, we look at each part of the supply chain to help not only get material into the supply chain quicker, but also move it through quicker and with more transparency. For lack of a better example, we have a product that is basically Uber for logs. We work with tree service companies and arborists. When they get enough logs to fill a truck, they can post that to our platform. We have a partnership of sawmills across the country, they can claim that, and then we work and coordinate logistics through third-party trucking. So again, not a revolutionary concept in the age of Uber and DoorDash and Lyft and all the things, but just applying that to an underserved industry is really where we see the opportunity, and we're seeing the folks on the receiving end of these solutions also see that opportunity.

Jay Kapoor: Yeah, and I think that's the biggest misconception I see as an investor. Whenever we invest in these businesses and we talk to some other funds, they say, "Oh, the market is not big enough." And that's always a very silly thing to me. I go, no, the market is there, it's just that the technology serving it hasn't spent the time to get there, and it almost becomes a self-fulfilling cycle. If nobody is building those solutions, then the technology market is never big enough, and that's why you see these industries being 20 years behind sometimes when it comes to tech innovation.

Theo Hooker: And just a quick note on that too, listening to the customer is a big part of it, because we work with a lot of truckers. In that sense, they don't want to be turning on an app. They don't want to be glued to their phones. So all of our coordination of logistics, when to pick up, where to drop off, is all text-based. It's all SMS-based. That's what's easiest for them. And integrating AI for voice-to-text, voice-to-form for customs. Getting stuff over the border is hard and complex. If you could just say, "Hey, I have a load of this, this, and this, and it's going here, here, and here," and there's your customs paperwork, there's your bill of lading. Those are the types of things that we're starting to see and uncover and make efficiencies.

Jay Kapoor: Alex, on that topic of meeting the customer where they are, I'm going to ask it in somewhat of a provocative way. Sensors and IoT have been around for 10, 15 years. So what is it about Bright's technology today that gets your customers excited? There's the demand, obviously, that's there, but what about the technology is so radical and so different that it necessitates that?

Alex Hawkinson: I'll ladder it up to that in a couple of different ways. One thing before I go to that is on the size of the addressable market. If the worldwide economy is $120 trillion or so, 85% of it is physical industries. The TAM is there. It's just been difficult to reach those places with the technology. So it's an enormous opportunity for us. But stepping back for one second, I previously founded SmartThings. It was a big consumer IoT platform with billions of connected devices, and I saw this phenomenon.

Alex Hawkinson: Let me pose a question for you. If you could understand what's going on in the world at all times, if you had a digital twin of the world, theoretically, what would the value of that be? If you could know, is something there right now when it shouldn't be? Is it vibrating weird? Does something smell off in that spot? Whatever it is. If you could wave a magic wand and understand what the state of the world is in a given spot, what would that unlock for you? I had a big aha moment over recent years, that if you think deeply enough about the problem, it turns out I think at least 50% productivity gains across all these physical industries could result from state understanding.

Alex Hawkinson: I can dig into why. Capital efficiency, you can spread your dollars twice as far because you're seeing problems in advance. You're getting in front of the catastrophic breakage otherwise. And actually, most of the sustainability goals, wherever you fall on climate change, they fall out of the bottom of that 50% productivity gain. You have a doubling of productivity, a doubling of capital efficiency. You're not sending trucks when you don't need to. You're not replacing broken stuff that didn't get out of hand. So within that, the key has been how do you unlock state in the world? The emerging technology matters here, because simple sensors don't get it done. You have to understand and interpret the pattern that you're seeing in a complex way, the way a trained technician would see it.

Alex Hawkinson: So the couple of breakthroughs that are powering what we're doing, and we have a platform called Stateful. One thing is what's happened with AI in the last several years. We all know about ChatGPT and large language models. They hallucinate, and there are other issues with them, but they're incredibly powerful. If you apply that same technology to objective data, not human language data, sensor data, you get very different results. You get very predictive, no hallucination, better than human prediction of what's going on. So that technology's come online. It's a big deal. The second problem is the energy of computation. Out in the physical world, it's hard to do the pattern detection there. So how do you address that problem if you can't plug it in or you don't have networking? You've had to have a breakthrough in energy efficient computing. At Bright, we've got a unique technology that dropped the energy cost of computing by a couple hundred X. So it's let us create next-gen sensors that fuse multimodal AI at the edge, that can operate in the environment for 10 years and do that pattern detection. So we're trying to unlock state in that way, and those are a couple of the building blocks.

Jay Kapoor: And the value, the way I take it, is that instead of having to send a technician out there, middle of the woods, three hours out from any sort of human civilization, you're now able to get real-time information in a way that you never could before.

Alex Hawkinson: Exactly. You don't send the people when you don't need to. In a lot of industries, that might be 50% of the time. You send the right person with the right part on the right truck at the right moment instead of going back. So you're quicker to know what to do. Your job at the site is quicker and more surgical. And then out of the back of that, of course, you're doing it in advance. You're not having catastrophic breakage. You're avoiding all this waste that comes out of the back of those processes. So that all adds up to those huge numbers that we talked about, and we're seeing it across seven different industries already. Power, energy, water pipelines, pest control, you name it.

Jay Kapoor: Yeah. So the second half of the panel idea is dirty jobs and the warming planet. I think we've talked quite a bit about the dirty jobs. I'd love to talk about that second half with you all. So Theo, let's talk about wood and why wood is so important to carbon sequestration. What do people not understand, or maybe what do your customers, as you're educating them on this, finally grok about the importance of carbon sequestration when it comes to lumber?

Theo Hooker: How many of you have thought about wood lately? Okay, some hands. Good room.

Theo Hooker: How many of you use wood on a daily basis? Okay, some more hands. I'm biased, but one of the things I like about wood in particular is that it integrates so much into our daily lives, from the desks we work at to the products we use, et cetera. But it is one of those unsung, quiet industries. It's not sexy, certainly not in the news as much, despite COVID price hikes and supply chains and things like that. I think a lot of people don't understand or aren't aware of where wood and the lumber industry actually fit into our lives.

Theo Hooker: And in particular, the wood that we don't see. The trees that fall, the trees that are not used, the material that is sitting in our backyards but underutilized in lots of ways. And going back to my previous statement at the beginning, on that out of sight, out of mind idea, a lot of the companies that we're working with across the built environment, across furniture, they really appreciate the story and the purpose that wood plays, not only in consumers' lives, but also in the economy, in the industry, et cetera. And certainly from our perspective as a startup growing, we have these early adopters, folks who appreciate that. Not only is it important to their business and their ethos to bring to light some of the supply chain issues or sustainability challenges, but also to promote that for their customers.

Jay Kapoor: Are these brands? Are they, like, carpenters? Who is this?

Theo Hooker: Yeah.

Jay Kapoor: Yeah.

Theo Hooker: A few of our notable customers, Room & Board, as an example. So we have a few different product lines with Room & Board. Sustainability is at the core of their ethos. They want to have 100% sustainable wood sourcing by the end of this year. We're a part of that. That's important to their customers as well. And helping them tell the story, helping them make this process easier, is certainly a benefit to us, but also something that they're really focused on. Another customer of ours, Patagonia, we've done a handful of stores for them. No one can probably think of another company that wants to put supply chains and transparency at the forefront of their products. So again, a very aligned mission, vision, in the world that we want to build. As we're growing now and getting more into the mass market adoption of those two products in particular, we're starting to build relationships, one of them with Amazon, using salvaged material into bigger applications across their operations.

Theo Hooker: That's where we're starting to get into bigger adoption, more price competition. So it's not just a sustainability thing. It's not just a story. It actually has some tangible value. And just to give you a tactical example, throughout our supply chain we have a lot of red oak. Red oak is sort of out of design these days. A little bit red, a little bit rustic, people don't like it. It's typically thrown away or used as pallets pretty much.

Jay Kapoor: And when you say thrown away, this is going to landfill, you're saying?

Theo Hooker: That's typically where it would end up. It's either landfill, chipped and mulched, or in some cases burned, just because that's the easiest way to get rid of material. We can turn red oak into a viable outdoor product through a process called thermal modification, which is just heat and pressure that actually bakes off some of the starches and sugars that bugs, bacteria, and fungi eat, which leads to decay. It turns it into an outdoor stable product that competes with tropical hardwoods and plastic composites, and is price competitive on the market. So that's an example of more than just the story, more than just the sustainability element. As we're building the supply chains, we're finding more of those pockets that customers are going to continue to resonate with.

Jay Kapoor: And I guess on that note, Rebecca, Amazon is one of your partners, also an investor in your business. I'd love to think that Amazon is doing this for purely altruistic reasons, but where does sustainability, and where does manufacturing and recycling, fit into their mindset? Is it purely altruistic, or is there another reason why they care about climate forward technologies?

Rebecca Hu: Yeah. This is where I think one of the reasons why Glacier's business gets really interesting. So far we've mostly been talking about these MRFs, these recycling facilities where all of that magic happens, the sorting, the recycling, the sending it off to the end markets. But if we rewind back to this idea of manufacturing, historically we think of manufacturing as a linear process. You make stuff, you consume stuff, you find a way to get rid of it. This idea of circularity becomes so much more salient when you have this type of AI and automation that's now in the mix, because for our customers like Amazon, like Colgate-Palmolive, like many other major producers who make a lot of the stuff that all of us buy at the grocery store or online and then throw out, up until very recently there's been a total black box around what happens to all of that stuff, that packaging, once consumers are done with it.

And you can't really optimize for an outcome that you can't measure. So take Amazon as an example. Credit goes to them to think about, we're putting all of this material out into the world. If we want to be more sustainable as a company, we need to find a way to actually measure those outcomes so that we can make changes and see if they actually have a positive impact on things like the recovery rate, coming back into the end market and being recycled.

Now, to your question about why these massive for-profit companies would care about this, there are a few reasons. If you want to set aside the sentiment that everyone should be a steward of their resources and should help preserve the planet, we'll set that aside for a second. There is a huge and growing body of evidence for all of these producers that the extent to which consumers buy their stuff is heavily correlated with the sustainability footprint and the climate footprint that they can put forward. In other words, for these massive companies, being able to prove that they're taking strides toward better circularity and more sustainability in their practices is directly tied to their revenue base.

So this is a very powerful reason why we should all be choosing with our wallets, among other things. Now on top of that, when you think about recycling as manufacturing, we get back into this question of a cost equation. Where are you sourcing your materials from? Amazon uses a ton of cardboard. Colgate recently put out a really novel toothpaste tube that's made out of entirely mono-material HDPE plastic, which means you can actually put it into your recycling bin for the first time ever.

As these massive companies are thinking about where to get material to feed back into their supply chain and how to justify that cost basis, that also brings to light another reason why they would care about knowing that, for example, there is enough recycled cardboard on the end market for them to incorporate back into their supply chain. So even setting aside the sustainability angle, there is a real dollars and cents operational reason why all of these companies are investing in making sure that we have access to enough recycled feedstock, and actually visibility to make sure that it's coming back into circulation.

Jay Kapoor: So with the last panelists, we talked about materials. But Alex, I'd love to have you chat about energy and grid, and that being a big part of how we think about climate resilience and climate adaptation, in a world, or at least in a country, with crumbling, unmaintained infrastructure. What role does physical AI play into the energy and grid side of the equation, and what are you seeing from your customers on that front?

Alex Hawkinson: That's a great question. It's obviously a big and complex value chain and ecosystem. But on the supply side of energy, just take that for a second, obviously climate change is putting two pressures on it. It's creating events that are damaging, more dramatic weather that can bring the grid down. If there are vulnerabilities that you don't understand, it can trigger wildfires, it can trigger outages. Even without the damage such as wildfires, every time there's an outage it has a huge economic impact and also a life and death impact for a lot of human beings. We require it for everything that we do.

And of course there's greater demand on that supply as well. It's not just hotter, it's colder and hotter, it's more dramatic weather swings. You need to cool spaces down to keep them safe and comfortable. You need to cool down food supplies to keep them safe for consumption, all those things. And then AI is trotting in behind and putting all this demand on the grid. So you have these vulnerabilities in the distribution and the supply of energy, and they all have costs associated with them.

So one of the things is visibility to that. We literally can't send the people out. The comment came up earlier about 500% worker churn. That's happening in the energy grid as well. People have to go and dig out the base of every wooden power pole every 10 years. It's a back-breaking job. Nobody wants to do that anymore. They have multi-hundred percent churn in that workforce. So you don't get out to inspect it as much, you don't know where the vulnerabilities are, and you're weaker. If you can know that state information around those things, you can get after it and maintain it for resiliency in the right ways.

The other side is really as consumers of energy. So that state information, I'll take another example, would be HVAC. It's the number one source of energy use in households, for all of us, by far. If you have a well-maintained HVAC system, it uses like 20% less energy, even older systems. And if you can monitor the state of those things so you can do proactive maintenance, you can actually make an appreciable drop in the wasted energy, which is great for everybody.

So lots of examples like that. I like to talk about businesses that are opportunities that aren't a whack-a-mole. I don't know if that means anything to anybody, but when I was growing up there was the carnival game where you hit something and another thing comes out of the other place. Often there are businesses where, okay, we'll take something away over here and it's going to be worse over there. All of this stuff that we're doing is a win-win-win. It's not a whack-a-mole kind of game. You make it more resilient, you also make it more efficient, the quality of the services is higher. Everybody wins.

Jay Kapoor: So for the startups and investors in the room, one of the things I want to underscore with all three of our panelists is that we didn't talk about carbon offsets. We didn't talk about carbon accounting. I say that just to underscore that addressing climate change doesn't have to come down to figuring out how you let one person pollute and not the other. I think for all three of you, what you've shared here, and I think is so critical and why we're excited about this space, is that as you create efficiency gains and as you address labor issues, you start to address climate as a result.

That's probably something that more investors are catching on to, and all of you have raised venture capital. So Rebecca, I'll start with you. Robotics has been kind of the kryptonite for a lot of VCs for a decade. What did you say to your VCs that convinced them to give you money? Because a lot of folks have lost money in robotics over the last decade, and yet Glacier has been successful in that space. So what did you say?

Rebecca Hu: Gosh, I said a lot of things. I don't know what specifically resonated with each particular investor, so we'll have to ask them directly. But in general, as I alluded to earlier, there have been so many powerful inflection points in the ability to commercialize something like a purpose-built robotics company as we have done. Among other things, a lot of investors are extremely surprised, in a good way, by our gross margin profile, because it is becoming increasingly inexpensive to actually make and deploy these types of technologies, especially compared to a decade or even five years ago.

So it is really becoming increasingly possible to make a venture-backable business even in the realm of robotics or industrial robotics. On top of that, one other really powerful layer of Glacier's business is this concept of physical AI, as Alex was referring to earlier, which is to say, with the way that a lot of AI tools are going, everything that is virtual is becoming so rapidly commoditized.

And so there is an incredibly powerful moat, especially as you think in the long run, in building an AI business where your core data is in the physical world, because that is information that we have access to that no one else does. Our scanners are installed in recycling facilities looking at materials that we, as the sole custodians, have insight into. There are many ways to monetize that type of first-of-a-kind data set, but it is a really compelling business to think about how to translate that unique opportunity into revenue. So there are a lot of fascinating things that are changing in the paradigm of how investors typically think about hardware-centric or robotic-centric businesses.

And the last thing I'll say, which speaks to the theme of this panel too, is ultimately, what are the problems that you are solving in the world? And even beyond today's state, what are the new opportunities that you are creating going forward? Sometimes investors might justify Glacier's existence against, for example, the number of existing sorters that are needed to sort at all of the recycling facilities. But what I would say to them is, if you can get this automation to be good enough and cheap enough and easy enough to deploy, how many new fully automated recycling facilities could you build? Or how many more communities across the US could you offer financially prudent curbside recycling access to, which then grows the overall total addressable market for everyone? So there are many new opportunities that this type of technology will unlock, even beyond making what's available today more efficient.

Jay Kapoor: The second order effect is what you're referring to. Now, instead of having very large facilities located miles away from communities, you can start to have smaller ones, and you also cut down on the transportation costs and the climate externality of that.

Theo, you guys just announced an 18 and a half million dollar Series A, somewhere around that. I'm going to flip it and say, congratulations obviously on the raise. What was the biggest pushback that you got from investors when you were raising for your business?

Theo Hooker: A few things. We got a handful of nos through the process. I would say the things that we got pushback on from one set of folks were the exact same things that resonated with others. So maybe I'll flip your question back on you. The reality that getting wood that would be wasted into the hands of people who can use it is not a revolutionary idea. We joke that using wood that falls in your backyard is kind of how humans started to do stuff. We've just drifted away from that a little bit. It is not an if problem, it's just a how. We know that there are millions of tons of material that gets wasted every year. It's how can we get that to the right people at the right time to use it in the right way?

And so to some folks, that is too much. There are too many moving parts, there are too many players in the system, and they didn't want to touch it. On the flip side, folks who have agreed to join our journey with us appreciate that and understand the value in these messy problems and how they do require complex or holistic solutions, and how we are situated really well to do that. And echoing Rebecca's point on the data, the lumber industry in particular, and as Alex was mentioning, throughout the built environment, throughout our physical infrastructure, there is so little data that is widely available and widely utilized for more efficiencies, for more innovations and things like that, that we have a real opportunity to not only access but also then utilize in bigger ways.

Jay Kapoor: And before I go to Alex, we'll start taking Q&A right after this, so if you've got some burning questions, please be ready, and we'd love to have you come up to the mic. Alex, you guys took an interesting approach of doing private equity partnerships before going out and raising venture capital. Specifically within these dirty jobs industries, these dirty, dusty, dangerous industries, why was that an appealing approach? Why did that approach work for you?

Alex Hawkinson: So a couple of things there. One is a little pithy comment on the earlier questions. Like him or hate him, with Musk, he's the richest man in the world. He's building real shit. This comment I made earlier, if the worldwide economy is 120 trillion dollars and 85% of it is physical industries, and all of the answers to sustainability, climate change, and quality of life improvements lie in those hard problems, it's really worth all hands on the flywheel going after it. They're great businesses to build.

So within that, for me, I've built eight companies over my career, lots of exits. I am a big believer that you raise venture capital once you know the pattern. You can judiciously use the dollars to accelerate the pattern that you've recognized in the world. And so for us in the beginning, it was a complex, ambitious thing that we were trying to build with the platform, and so we bootstrapped with the customers. We found willing customers that took a little longer, that wanted to work with us with a shared vision, and got through our first 80 million in revenue that way, and locked in on the pattern then.

And then we were able to raise a round last year. We could have raised it before, but we raised a 20 million dollar seed last year. We're going to raise again this spring. These are enormous business opportunities. You raise venture when you've got the pattern down, not before. A lot of people celebrate the raise itself. It has nothing to do with the outcome. It's got to be acceleration capital for a pattern that you're implementing.

Jay Kapoor: And I'll say from the VC perspective, it's one of the things that founders don't, especially first-time founders, don't fully appreciate when they go out to raise venture capital. You're signing up for a relentless pace and speed of execution that your investors are expecting from you. And if you're able to deliver on that, as Alex is saying, then it's a fantastic use of capital. Then the dilution is worth it, because the speed at which you're going to be delivering, potentially raising more money or getting profitable or going public one day, is worth the capital you're raising.

But especially at the seed stage and Series A stage, there are founders who think they're ready to sign up for that, and then they raise the capital, get six, nine months in, and realize that it's very hard to sustain that pace of execution, to do it reliably, and to grow on the scale of weeks, not months. And so I commend you on recognizing that and designing your fundraise around it.

Alex Hawkinson: Can I add something to that? I think the question to ask yourself as an entrepreneur or a business person implementing these things is, with these dollars, will I be pursuing my highest and best purpose more quickly or not? And so for us, as an example, the judgment call is, even if I had dollars, I actually think I would just fail at delivery if I tried to push too hard. We're still discovering that pattern. Once you get it, we've grown 600% in the last six months. You find it, and you know when it's the right time.

Jay Kapoor: Please introduce yourself and address our panelists.

Audience Question: Hi, my name's Eleanor. I actually have a question for Rebecca. I love hearing you pull back the curtain on the recycling industry and this task that we probably all do every day when we put the can in the bin. But what happens next, and what I think would be really interesting to hear you share more of, is that buyer persona. Who is it that you sell to today? Are these family-owned businesses? Are they conglomerates? Are they private equity? Who are the people behind the facilities? And when you think about, if you're five years in now, five or 10, 50 years from now, how do you see that buyer evolving, or that space or the opportunity in the facilities themselves?

Rebecca Hu: Thanks, Eleanor. So to answer your question most immediately, currently Glacier's core customers are these recycling facilities, these MRFs. But as you allude to, there are many different types of MRFs out there, owned by many different types of entities. One common misperception is that a lot of people think these MRFs are public, that they're owned by government agencies, when in fact about 80 to 85% of them, at least in the US, are owned by private companies.

Now, if you double-click within that even further, there are a couple of these really big waste companies, Fortune 500 companies like Waste Management, Republic Services, Waste Connections, and so on, that might own dozens up to 100 MRFs across the country. These behave much more like enterprise buyers. If any of you entrepreneurs in the room have tried to sell into a Fortune 500 company, you know that comes with a lot of multi-threading, a lot of understanding and aligning stakeholders, proving out the financial justification, and so on and so forth to many different teams.

All the way down to, we also work with family-owned operations. For example, a recycling facility that's been owned by a family since the 1970s. They ultimately are still in it for a financial return, they need to succeed as a business as well, but their incentives might be much more tied to the performance of a specific site in a specific region versus optimizing across their entire network of facilities.

Now, ultimately, what Glacier is really excited about is what I call democratizing access to this type of technology. So whether you are a huge Fortune 500 waste company or you own a single site and you're optimizing for your own family's future success, we want to make sure that our technology is cheap enough and easy enough to implement that it can deliver real financial results for anyone that we work with. Because at the end of the day, one of the other insights that not a lot of people realize is that recycling fundamentally is a business. A lot of folks are in the industry because they want to do good for the planet, but you can only do that if you can sustain yourself as a business. And so what we find most exciting is this idea that automation and better tools like AI can actually fundamentally change the economics of recycling, and in doing so, we make it more financially profitable, which then incentivizes more people to go into the business, build more of this infrastructure, and proliferate access for more and more people. So that's our ultimate goal.

Audience Question: Thanks, this is an amazing panel. Thank you very much. My name is Bridget, and I'm a fractional CFO for tech startups. My question is about the mentalities. You guys are doing your share, clearly. You're visionaries, you're early adopters. You're clearly moving things forward. But to me, it feels like the technology is there, but I don't feel the mentalities or the will is there in the overall population. So what thought-provoking thoughts could you share on how you do your share? How can we do our share? What are the best avenues? Is it social media, regulatory? I know we're in Texas, but government-private partnership, what can we do? Or higher education, what can we do to accelerate pivoting the mentalities?

Jay Kapoor: May I take that a little bit? A big part of what our fund focuses on is helping our startups with their storytelling. So it's exactly your question. How do you find the people who care, and how do you reach them? One of the most important pieces of this that we've seen for business leaders is that in the age of AI, in the age where the cost of robotics has come down by a factor of five X in the last 15 years, a robotic arm that used to cost 100,000 costs less than 20,000 now, and it's startup affordable.

The messaging that has worked, to put it simply, is FOMO. You go to a business leader and you say, if you're not implementing this, your competitor is. They are implementing AI. They are implementing labor automation solutions. They are addressing those labor shortages using technology. And so part of what we do in helping our companies is help tell those stories so those case studies are out there. Social media works to a degree, events work to a degree. I don't think there's one silver bullet. It has to go across those channels.

But once you create those proof points as a startup, you can go to those business leaders and say, hey, I'm speaking to Colgate-Palmolive, but I'm also speaking to J&J, and everybody's thinking about this, and it is becoming a part of how they build their cost structure. FOMO works. It really does. Nobody wants to be left out or be the last one to adopt those technologies. And so, as Rebecca said, you have to make the cost case, you have to make the business case. Just leading with the do-gooder altruism, the bleeding heart, might get you a conversation, but it's not going to get you through the approval if it doesn't make business sense. And so I say lead with the FOMO, lead with the fact that other folks are adopting this already, and you'll start to see the change happen a little faster.

Alex Hawkinson: Can I add to that? Thanks, Bridget, for the question. I'm not a totally calloused guy in terms of my view of the world, but people are good in their hearts, and often as an after effect, you have to have it make business sense at the end of the day. It has to be, this is the path to profitability. And I'd say the two things we run into that drive the conversations are the opportunity and the fear.

The opportunity is that we can drive 50% productivity gains, double the life of capital dollars. And the sustainability is an after effect, so you should feel really good about it too, and you do have the true believers that pile in for that as the front reason. But if you can say those other things, it's very exciting to anybody in the business world. You can build up a great business.

The second is fear. Digital technologies tend to create winner-take-all dynamics. Like the Roman Empire, I like to say, why did it end at Hadrian's Wall? Because it was the maximum extent of communications inside of a centralized organization to actually be effective. So it got a little dumber with each mile that it grew. But businesses that technologically scale get better with every customer, every site that you have. You're better every time, and they tend to create these winner-take-all dynamics. So to your point, if a competitor is implementing this before you are, you are effed on the other side. And so you can use those two things. Business profitability, it's obvious. And by the way, if you don't do it, it's coming for you very soon, and you're not going to like what it feels like when you're two years behind a curve that looks like that.

Audience Question: Hello. This is one of my favorite panels of the week. Hello from the architecture, engineering, and planning industry. I'm so excited about this. My name is Bea, I'm a social scientist. I run a research and development program at a large integrated design firm, and we try to do on our end the things that all of you are describing. We try to do circular materials. We've designed a cost-neutral timber hotel. We work with the energy grid and monitoring-based commissioning. And one of the questions we get across all of these is something you all referenced at the beginning, and I have a how might we.

We are always asked, or in the case of the timber hotel, we were told by developers we worked with, this is a great idea, we can make it pencil out, but we wouldn't deploy it first in a market with union labor. And when you talked at the beginning, there's a lot of fear of job loss that just doesn't line up with the reality of the math and projections. So how might we show organized labor, which personally I feel is a great thing and very compatible with innovation, as not a block that people need to fear as they try to deploy these innovations? I think yours is a perfect example. I would love to hear from any of you.

Alex Hawkinson: I would love to add to this. This is like a trick question, but not a rhetorical one. Who here thinks the world is as awesome as it could possibly be? Like, we're at peak awesomeness. Nobody, right? There are so many ways to make it a better place. There is plenty of work to go out and be done. So then you demonstrate that in every single one of our seven industries, there's job growth in every single one of them. People want to have these modernized jobs. They want to make a difference in the world. They want the quality of the services that we consume to be better.

So you just have to show that unlock, from my standpoint. You demonstrate, here are 100 case studies of job growth in every single organization at scale that's implementing these technologies. And by the way, there is plenty to do. It's not a zero-sum game. Wouldn't it be better to have better water? Wouldn't it be better to have higher quality materials that weren't releasing more carbon into the ecosystem? Wouldn't it be great if the energy was reliable and also less expensive? The world is way off of its awesomeness peak.

Rebecca Hu: Yeah, and just to add to that, from Glacier's perspective, we've run into this question before as well, whenever you're dealing with anything related to automation in the workforce. I do think we're, in some respects, fortunate to be in a space where there is this kind of 500% year-over-year turnover. We work with a lot of facilities that are seeing anywhere from 20% to 50% labor shortage. So in some respects, it's almost a moot point. But in cases where it does come up, I've found that one really powerful angle we've leaned into is, the fear comes from, I might be losing my job. But a really great alternative is to say, well, what new avenues might it be opening up for the workforce instead?

One thing that Glacier has really leaned into, for example, is this idea of workforce development as part of rolling out our technology. A lot of the recycling facilities we work with have trained technicians or staff in-house that are accustomed to dealing with all kinds of industrial equipment. We've actually really leaned into creating a proprietary learning and development program, where every time we roll out our technology at a new site, we'll train their full staff on how to engage best with our robots, how to troubleshoot them, even how to do a lot of basic maintenance and repairs themselves.

Because for us, this is also beneficial. We don't necessarily want to be in the services business. We want to be innovating and building new technology. So the more we can hand the reins to these workers and make them feel in control of the way this technology is being introduced into their working environment, and work with them to upskill themselves into more rewarding, safer, longer term careers, we've found that it tends to be a win-win.

Jay Kapoor: I'll use another real world example that I think the shipping industry did really poorly last year when the longshoremen strikes were being talked about. They talked about doing blanket automation at ports, and obviously the longshoremen pushed back. You guys might have seen this very colorful character named Harold Daggett who was like, I'm going to end all of you. I'm going to end your careers. And part of it was he was trying to rally his troops in the longshoremen union to say these jobs are going away. But if we look at what's happened in Asia, for example, most of the ports in China are automated, but they do have remote operators. They have people that sit at a desk all day and move a joystick to make sure that shit gets loaded correctly.

What the shippers didn't do a good job communicating is that those jobs do need to be created. Now, you could go hire a 22-year-old fresh out of college that can play a video game and move it around, or you could take the existing labor and say, if you're willing to learn, we want to have the programs to teach you. And that just wasn't talked about, or if it was, it wasn't publicized well enough. So I think that's a really great point that you made, that if you go in there with the understanding that your job is going to change but it's not going to go away, you're going to end up retaining a lot more of that institutional knowledge that otherwise would just be walking out the door.

So I think we have room for one more question. Sorry, go ahead.

Theo Hooker: Just quickly to add on the mass timber side of things in particular. Compared to Europe, which has around 70 mass timber manufacturing facilities, the US has 17 or way less. The beauty about mass timber is that the machinery needed, people are not assembling these big sheets of mass timber or CLT. There are already machines built into the process. Where we are is actually getting more prototypes into the process, into the mainstream. So shooting that as, if we do this, more jobs will come, this industry will grow, it's already growing, do you want to be a part of it? So a little bit of the FOMO answer as well.

Jay Kapoor: Question for the day.

Audience Question: A final question, I hope this is a good one. Hey, thank you all so much. So the question is around the gentle friction between values and pace. When you talk about venture capitalists and the need to raise that money and see turnarounds in weeks, not months, oftentimes there are all sorts of cases that we all can pull from history where, when you move that quickly, you really don't know what the consequences are of the actions you've taken until down the road, especially with communities that are in some cases so vulnerable. How do you balance that? How do you think about that as you weigh the needs of your investors and the money you need to raise and the pace you want to keep, and then the impact on the communities that you might know you're directly serving or impacting, but then others that you might find out about five years down the road, my God, I didn't realize we were going to impact them?

Theo Hooker: I can take a first stab at that. From an investor perspective, having freshly gone through this process, a lot of the folks that we have in our room now appreciate the nuance in the work that we're doing. If they wanted a quick money and hockey stick outcome, they're putting their money in the wrong place. And that might be counterintuitive, but the folks that we have brought in understand and want to be in this solution for the long haul. In our conversations, we're talking about the five, 10, 15-year roadmap that we have, even though we understand that results are needed in the near term. And there is a lot of opportunity, and more folks are becoming keen on exactly that, the tension between those things. So one piece of advice would be that there are people out there who share those values and understand the implications if you don't consider or appreciate the balance between those two.

Jay Kapoor: And I would echo that. Venture capital is not a monolith. I think currently you're seeing the industry grapple with itself of, what is venture capital in a world where AI is moving this quickly? And is there, are there any moats in software left? And we're sitting here with three founders that interface with the real world, but interfacing with the real world is messy. It takes longer. It has many more externalities than a piece of enterprise software. And so the question, I'm not going to have an answer for you, but the question that the industry's grappling with is, what does the venture capital return profile look like if it takes a couple more years?

And then from our perspective, we raise our money from LPs like endowments and fund of funds and institutions. What are they expecting from us when they're not going to get the great enterprise software SaaS-type returns of the last decade and a half, because AI has changed all of that? You and I today could go build a CRM tool in about five minutes using no code whatsoever. When you can operate at that scale, to me, selfishly, because I invest in this stuff, the real opportunity is in the physical world and in addressing these dirty jobs. And so I think you're going to need a venture community that understands that and responds to it.

Alex Hawkinson: I think I agree with everything these two just said. I wanted to think about the community aspect of what you were talking about. It's a tough one because things are moving so fast now. I'm 23 years old and look at me. It's been a rough road. No, there's a lot of energy being on the front lines. But you kind of have to evolve quickly, and be adept, and it's a challenge we're all facing.

To me, there's a big role of storytelling in all this, going back to something you said earlier, and not about the FOMO. I think it's really important that we as a society make it cool to work on really important problems. It's not about making a quick buck. Is what you're working on making the world meaningfully better in a way that you can understand? And so the prize is worth it, whether it takes more time or the urgency to move really quickly. Is it an important enough problem?

And if we tell the stories and you get specific to the individual problems or industries, like, hey, we can eliminate 90% of the pesticides used in pest control with this proactive technology because it's killing the world, you can get people on board if they understand the meaning of it, whether it's slow or fast. And so I think there's this huge movement we all need to push toward, which is the importance of doing really critical work.

Jay Kapoor: That is a wonderful place to leave it today. Please give a hand to my panelists. Thank you all so much for joining us. I think we'll stick around in the room a little bit until they kick us out, so if you guys want to come say hi, please do. Thank you very much.