 Okay, good afternoon everyone. This is the winter 2024 innovation showcase brought to you by the Tomcat Center for Sustainable Energy. My name is Matt Cannon. I'm the director of the Tomcat Center on behalf of our entire team, myself, Donica, Brian and Elizabeth. I want to welcome you to this event and thank you for joining us. I know several or maybe most of you have attended one of these events in the past, but just just very quickly for those of you who are new to an innovation showcase. What we do is we feature three companies started by Stanford students whom we have supported through our innovation transfer program at the Tomcat Center. I should say that the innovation transfer program has supported over 100 teams that have gone on to form companies over the years. You're seeing just a sampling today. And if you're intrigued by the innovations that we support, then I encourage you to check out our website where you can get information about all of the different companies that have come through the program. Donica will post a link to that in the chat. But we feature three founders in this event and what they're going to do is they're going to give a short presentation about their company and about their technology and talk about its potential impact on energy and sustainability and GHG emissions, which is really central to our mission here at the center. We have a about five minutes for a Q&A session. So please add your questions to the Q&A function throughout the talk or during the session I will do my best to accommodate them. But if I don't get to them, the speakers will be able to see all those questions. And I really encourage you to follow up and connect with them either through us, the Tomcat team, or there'll be mechanisms for you to connect directly with them. And thematically today, all of these teams are really dealing with efficiency. But in three very different arenas, as you'll see, and we're going to kind of progress sort of along the supply chain, if you will, starting with the earth with mining. And then going to manufacturing in the factory. And then finally, in buildings where we reside and work. So without further ado, I'm really pleased to introduce our first speaker, Tyler Hall from ExploreTech. He's the co-founder and president of ExploreTech. ExploreTech increases mineral exploration efficiency with a scalable AI platform and just to kind of set the stage for the impact and the problem here. Critical minerals that are needed for energy transition and defense. And for those is growing at a very rapid pace estimates of increasing demand on the order of several, a multiple of three or four times what we extract today, which is already an enormous amount. And so the ability to find resources and extract them efficiently is really going to be critical for many different industries. So Tyler, take it away. Sure. Thank you, Matt. Let me go ahead and share my screen and we'll get started. Can you see my presentation properly? Yeah, looks great. Thank you. All right. So thank you, Dr. Kanan. So today I'd like to talk about a common problem in the physical world and this common problem affects so many things, including natural resources, infrastructure, disasters, governance and the environment. And so for the next few slides, what I'm going to do is focus on natural resources and specifically copper. There's a big supply constraint coming up just like Dr. Kanan had said. And that is that to achieve net zero by 2050, which is in 27 years. It's not so far away. It's 2024 today. Humanity requires twice the volume of copper. The world has produced over the last 3000 years. So what does this look like? Well, on the left is a cube of all the copper metal mine since the start of civilization. And so I'm comparing its height to the height of the World Trade Center. Well, the one World Trade Center in New York. So it's just about equally high. And here's the amount of additional copper that we need to achieve the stated net zero 2050 goals, according to the IEA. And so the question is, how do we get here? This is a huge problem, especially considering that it takes 15 or more years to go from discovery to actually getting the metal out of the ground and distributing it. And so here's a simplified value chain of mining, starting at discovery and looping to distribution. Again, this takes 15 or more years to actually go from discovered metal in the ground to extracted processed and sold. And so we need to speed up this loop from 15 years to a lot less. And so all of that involves complex engineering problems. And so from a software perspective, it's possible to address aspects of all these using computing, right? And that's what we've done at ExploreTech so far. We've been working on speeding up early stage metal discovery, right? And so since incorporating in August, we've been working with early stage explorers to speed up the discovery of metal deposits. Explore takes geophysics data and outputs an optimal drilling plan. And so it does this in 1 week rather than 3 months. And so you can read about it more at our website, our brochures right on the landing page www.exploretech.ai. And so there you'll see a brochure with a little bit more information. But that's directly for customers because we'll be going to PDAC next week in order to meet more people there. But since incorporating, we've also been applying the same computing technology to a problem in a completely different domain, but extremely similar problem formulation construction and data, which is real infrastructure modeling or monitoring, believe it or not. And so here people are trying to use physical simulation to test the strength of railroad ties. And so this machine will drive down this track and test railroad ties repeatedly. Every single one of these ties will need to be tested for its integrity. And so there's millions, tens of millions of miles of this railroad track. And so you can imagine there's millions upon millions of these ties that need to be tested. And so it's important to be able to perform those computations very, very rapidly. And so that's what we were doing in streamlining that physical simulation. We're able to save an entire month in data processing. Right. All of this is to say that it's a common thing that we're observing in industry is that the software that people use is inaccessible. It's unable to solve highly computationally intense problems. Right. So what do I mean by that? Well, the software that's the state of the art. It all looks like this. Men use buttons, scrolls, windows, everything. I mean, look at this, right? It's the fundamentally inaccessible way that resources are discovered plan for mind process and everything else in the infrastructure and built world, right? Ends up using software that sort of looks like this and is fundamentally inaccessible. Nobody without an advanced degree will ever be able to use this tool. Let alone that it doesn't have the capability to benefit from advanced computational techniques like the ones that are invented by people like Alex and myself for mineral exploration and beyond. So we need to do two things. The first is we need to solve for the unique complexity of the physical world. Right. So we need to unleash this power of high performance computing for these tasks where it's not currently applied like mineral exploration, like rail infrastructure monitoring. We're the only ones doing it right now. And number two, we need to do it in a manner that's accessible, scalable and repeatable and understandable. Because we know that we're running out of the engineering expertise that's required to actually carry out these tasks because people are retiring and we're running out of experts. So explore tech. Where are we so on the left is Alex. He and I met in 2017 when we started our PhDs is a geophysicist and a graduate of Stanford, formerly consultant to Rio Tinto. He's the co-founder and CEO and on the right is me. I'm a geologist co-founder and president of explore tech. And finally, I worked with the Freeport Mac Moran copper and gold and Glenport. And so, both of us met at Stanford and big shout out to Tom cat, where I graduated as a graduate fellow. And we've enjoyed a transfer grant and we're looking forward to an intern this summer who we just extended an offer to earlier this week. So our first market is in natural resources, early stage mineral or early stage metal discovery, which itself has about a $200 million annual revenue of just the niche that we're targeting. Other solution domains where we work is in rail infrastructure where the serviceable addressable market for us is about $420 million of just rail infrastructure monitoring and doing that in a very scalable manner. And third in geospatial analysis. So all of these markets have the same common problem, which is this common problem of needing to marry expertise and compute in an elegant and accessible way. So we incorporated in August and since then we've had seven contracts either approved complete or in procurement at a value of about $470,000. You can read about them in two customer press releases. One of those press releases will be coming out either tomorrow or early next week. So stay on the lookout for that one. And so here we're looking to connect with people with experience in physical industries. And so if you're working in construction chemicals, semiconductors, biosciences and otherwise, please we're looking to get in touch with you. And so, again, we want to talk with experts, advisors and possible investors as we're looking towards growing our team over the next few months. So with that, happy to take any questions and thank you. Great. Thank you, Tyler. Really interesting to see the progression of your journey here. I guess I'd like to start just asking a little bit how you've, how you saw this connection between the early metal early metal exploration problem and the rail infrastructure problem. And you sort of connected it as, okay, we need tools where there's expertise that that maybe is being lost and it's complex computations, we need tools to streamline that. And so I get that connection. But how did you, how did you sort of make the leap from from the geology focused exploration to the to the rail infrastructure. It all came down to geophysics. And so geophysics is like a remote sensing type of way to be able to characterize materials. And so if I scroll all the way back to one of our slides showing this here, we were using geophysics, which is this helicopter data to sense what's in the subsurface, which is this ellipsoid. And sorry, I'll put my mouse down here, which is this ellipsoid. Right. And so you're trying to characterize material in the same way you're trying to characterize material of these railroad ties. And so if you're able to speed up that computation of detecting what is in whatever material you're testing, you're able to do that very, very rapidly. And so instead of your expert spending time on all of this data processing. It is only processed for a little bit of time. And so what they can do is then spend more of their time on more of these cycles. This cart can go farther down the track before they have to stop and process the data and on and on. So we're essentially unlocking the amount of track throughout the nation that we could actually sense. And in the same way, we're unlocking the number of targets that can be assessed in mineral exploration so that experts cannot spend their time on processing, but on deciding which targets to go after and so on. So do you make any of the hardware or sensor technology or is it all a software tool now? It's all in the orchestration of the compute infrastructure and delivering it in a way that people can use it rapidly. Okay, okay. And for the geophysical data collected by helicopters mostly, I guess, how accessible is that? So can you use your tool sort of anywhere or do you have to find somebody to collect a certain kind of data to help you provide the optimal information? Best places in the world for commonly available geophysics data or just generally like spatial data related to mineral exploration is Australia and Canada. There's a lot of opportunity for more government surveys in the United States, but we've worked with government survey data in the past. It helps to have had like an explorer already have this survey performed. Like this, they had done this survey already, but just looking at that, their staff doesn't know which target to go to. Do they go over here? Do they go over there? It's hard for them to make that decision. And so a lot of times the people that we work with have way too much data and they don't know how to assess it in a rapid manner. And so that's the problem that we're solving. The data already exists. It's just people aren't processing it fast enough. And so that's the problem we solve. And can you, I'm sure the analysis is complex, but can you give us a sense of how the algorithm translates this kind of data into a drill plan and maybe what do you mean exactly by a drill plan? Yeah, so the technology is in this case for X flare, it's two components, which is what we call inverter API and then Driller API inverter API makes sense of the geophysics data and goes from this data to the number of ellipsoids that all explain the signal at the surface. So each of these possible ellipsoids explains whatever signal it is in this target area. And then we design the borehole or the drill hole that will optimally intersect all of these ellipsoids, right? Not only do we do that, but we do it in a way that can say, okay, if we had three boreholes, how would we drill? If we wanted to drill a sequence of three boreholes, which one do we drill first and so on? That first component is Alex's dissertation. The second component is my dissertation. So it's essentially this one to combo that ended up teaching us about like the value of architecting these algorithms in such a way that they're rapidly scalable. And so you obviously have connections to the mining industry from, I think even from before Stanford, how difficult is it or not to really connect with potential customers here and explain your product to them and get traction in terms of getting someone to try it out. Yeah, so it's actually kind of nice when we can talk to people, they understand what we're trying to do and we solve a problem that they resonate with. And so they say, look, we have exactly the same problem where we have this target area. We don't have the expertise to actually know where exactly to go ahead and drill. So we've had good traction when we can get in front of people. What limits that is just getting in front of people. And so we've been going to conferences and physically meeting them. I mean, we find that that is the best way, right? Getting hits on our webpage. People are busy. It doesn't really resonate as much, but that personal connection, which we've been to now three conferences that Tomcat through the innovation transfer grant has helped us sort of get to. And so meeting people in person is the best way. Yeah, there's no substitute for the interactions. And then last question. So you highlighted these sort of three markets that you're going after. What do you see as sort of the ultimate vision? Do you think this is a tool that has a sort of much more general set of applications or is your goal to really try to specialize and capture those key markets and then build from there? We're trying to make this as general a solution as possible. Because what we're finding is that the world is running out of expertise. And the world is running out of people with extremely specific degrees and experience and stuff like that. We want to make it so that people can perform this drill planning in a way that makes sense to them without having to go through, you know, five years of experience working at a copper mine, going to Stanford for five and a half years and getting a Ph.D. And on a hunt, that shouldn't be a barrier to it. And that's what we're trying to solve. And we're trying to do it in a way that everybody can access it, not just a couple of people. Fantastic. Thank you, Tyler. I guess the last thing I want to add is it's really nice to see someone that we supported as a graduate student for, you know, we have a graduate fellowship for basically to allow students to really sort of develop an application of their research. And as Tyler mentioned, he was one of our awardees and then has, you know, really taken off with it. So it's been great to see that. Thanks for sharing your story with us today. Yeah, and thanks to Tomcat for the opportunity for the funding. And just for the excellent collaboration to really appreciate it. Thank you. Fantastic. Okay, so next is our next speaker is Lauren Dunford, she is the founder and CEO of a company called Guidewheel. Guidewheel is on a mission to empower the world's 10 million factories to reach sustainable peak performance and just sort of one fact that may be surprising to set the stage here. And manufacturing and factories accounts for about a third of GHG emissions as a huge opportunity here. Lauren, it's great to have you join us today. Thank you. It's an honor to be here and so grateful to Tomcat. Let's see, I'm sharing my screen. Are you seeing it the right way? Yeah, it's great. Great. Well, just so excited to be here together. I'm going to take us on a journey into the world of how the stuff in our day to day lives is made. And specifically what Guidewheel does is AI power factory ops. And I'll share more more about that in a moment. But first wanted to just start with a problem. If you are a manufacturer making things or stuff. The time that your machines is running is really darn important to you because if your machines are not running, every second or minute or hour that you lose is revenue you're losing from an existing asset that you have. And of all of the factories globally, they are losing up to and sometimes a lot more of 20% of that production time on their machines a huge amount of waste. It's not just to catastrophic machine breakdowns, you know where everyone's running to fix something, but far more frequently to the daily human stuff shift starts late. An operator is changing over the product on a machine from product a to product B and the material for product B isn't right there and so they have to run to the back to grab a forklift and suddenly have lost 40 minutes on that changeover is all this kind of daily human stuff. It's really important, and that a few of the world's factories, the ones we see in the newspapers and you might recognize as this image of a factory have very advanced the spoke expensive and complicated systems to solve and manage very closely. But what is often quite surprising to people outside of manufacturing is that the vast majority of manufacturers and of machines in the world are managed far more manually on yellow clipboards that look like this or Excel spreadsheets that hopefully don't look like that. And if that's a surprise to you for folks in manufacturing, it's usually not but for this audience I wanted to run through just one real example of why that can be a company called pretty and I'll share what it was like for them before they started working with us, they make these kinds of plastic models you probably recognize a lot of these from your kitchen and in all of their their plants they have many many and 25 plants, many different types of machines to transform raw material into product ready to be sold so lots of different types of processes across those plants. They also have big and expensive pieces of equipment. So there's not typically a time when it makes sense for a plant to toss out or get rid of all of their equipment in a single year and replace it with all brand new equipment. Instead they're usually buying a machine if a machine breaks, or they're buying a machine if they need to add capacity in a certain area. So what that contributes to is different ages makes and models of equipment within each of those patchwork of processes as well. Average equipment age in the US is 20 plus years and increasing and I've spoken with manufacturers you'd be very surprised where the average age of assets in their bleeds of 10s of thousands of them is 60 plus years old. But to make it even more of a challenge, kind of getting that real time visibility and managing that day to day, all of that equipment is being run every single day by real humans. Everyone from the operators running the machines to the maintenance technician to the plant manager to the CEO is, you know, every single day and before they started working with us using manually tracked data on these clipboards transcribing into Excel and emailing around Excel. And this is a billion dollar plus extremely professionally run operation. So if it was surprising to you that there's an opportunity within manufacturing hopefully that that example is a little bit helpful. What we've done is take the visibility that was formerly so expensive so cumbersome and really require the kind of expertise and investment that was only accessible to a few and made it accessible and democratized to be something that that any factory can clip in and under 24 hours. Pays for itself in the first month and can be intuitive and use across the entire factory floor. Basically taking that fast curve you can see here of a traditional investment and pulling it a lot faster. This is a sample of what that can look like for a particular plant. And the way that we do that you've noticed I haven't talked about sustainability yet but don't worry the way that we do that is we use a sensor like basically a smartwatch for a machine that clips around the electrical heartbeat the incoming power to any machine because if you think back to all of those projects models ages of equipment, they all use power, the simple and elegant truth. So we clip around that power draw. And that's where we've got algorithms and software that turn that heartbeat like your heartbeat monitored by a smartwatch might show if you were having a problem, whether you're running or walking, what speed you're moving, etc. We can show how the equipment is running running idle off how fast it's producing how many cycles and output, and if it's starting to have a problem and get a remarkable amount of power from that simple layer of visibility, which also means we're building in from the start and energy is such a core driver of emissions all that energy management information in a way that's built into the operating fabric of the entire manufacturing team. What we've built is workflows that really integrate that that energy and sustainability into how the factory runs so getting that sensor data into the cloud, reaching them on scoreboards like this tablets of the machine, the systems they're already using and and integrating into the meetings they have shared source of truth to make those meetings really efficient make the manage, help them manage out any loss from from machines, alerts and recommendations to address issues before they get worse than the tools to manage those incidents. And the workflows to continue to improve continuous improvement is a real thing within manufacturing and just fitting in to that existing process. With a system that's built from the start to get smarter and better and faster every moment at predicting problems before they get worse, diagnosing and recommending action and benchmarking based on all of the data in the system already. And this is where it starts to get both pretty interesting and very relevant to the topic for today, which is we've built that platform from the start. So that it is getting smarter and better with increasing returns to scale because that scale when we think about impact on climate and supporting manufacturing has such benefits. Not only is it a net new data set on the most important or if not one of the most important metrics in the factory. It's also layering on top of that second by second sensor data, all the information to make more and more meaning from that production quality reasons for loss production time in a flywheel of more and more impact from that information being layered on. And here's just one example you can see our users are actually tagging these incidents for us, not just because they're going to benefit from the amazing algorithms we will build for them, of course, and have built based on based on all of that tag data. But also because they want to roll it up in the morning and see where was our 8020 where is our top cause of downtime on loss with the eventual goal not that guide wheel has to build everything for every user or factory, but that customers have easy building blocks to build on top. And eventually at scale and ecosystem of developers can accelerate improvements for everybody, even faster. A very concrete example of what this can look like. We always work directly when we can with the VP of ops the person responsible for production and industry week one of the top outlets in the manufacturing space did a really nice detailed example of how pen color was able to increase runtime on their key machines their key bottlenecks by 50% by using guide wheel at first the operators didn't believe the data they were seeing and then oh my gosh what an opportunity to improve. And really importantly engaging the team not just in those cost savings but in using their time in a way that was much more fulfilling and productive and got the whole team feeling aligned and and working closely together and what's not in this industry week article which makes me really happy is the work that pen color is also doing to use that same foundation to really be a leader in sustainability and emissions reduction because they have it already part of their factory ops workflows about to close here and can't wait for lots of questions but just in terms of who we are. Originally, Tom Cap was, we got one of the innovation transfer grants they funded our first very early prototype, backed now by great craft and breakthrough energy ventures to the top investors in the world have built a really unprecedented team in terms of expertise across software, AI and manufacturing plant for experience and live already with hundreds of manufacturers across US Latin America, Canada, Europe and Africa. So just really honored to see how needed this is and how fast things can grow towards the vision that we can do this. Like you said at the beginning manufacturing is close to a third of emissions, and being able to be that platform continuously improving integrating sustainability into those data day workflows can be a really transformative impact. And I'll stop there. Thank you, Lauren. That was really fantastic. And I have to say, I'm a little bit biased because I happen to love manufacturing as a as a problem area, but nonetheless really really showed the tremendous opportunity there. I guess I want to start just asking how you personally got to know about this, this opportunity and became interested in this space. Yeah, absolutely. And I just got to say for everyone listening this space, if you're not in it already come in. It's just the opportunity is so massive. So I kind of first got obsessed with that operational efficiency. Here's something that's good for business good for the planet. Actually, when I was an undergrad at Stanford back in oh five to oh nine so I let students for sustainable Stanford founded a group that was specifically focused on operational efficiency within the operations. And then, and that was the green fund so we did this type of project actually in a ton of energy efficiency work obviously not manufacturing but got that bug. Then continue diving into supply chain got a full right to study supply chain actually in New Delhi, and then spent five years working out of a fresh food manufacturing plants in East Oakland and was lucky enough, although I was on the partnerships team I was in charge of 65 million of the West Coast business but physically had to walk through the plant floor to get to my desk, and certainly, you know, was managing very closely our customer relationships so saw how critical manufacturing efficiency and productivity was, and how, how oversized all the solutions that existed for real time visibility were for plants of our size so headed back to get the MBA at Stanford and focused every second that I possibly could on how do we build a really impactful scalable and profitable business in this area by approaching the problem in a smart way. That's great yeah so clearly the passion is backed up by a lot of time and effort and personal experience, so that that's really nice to see. I want to dive into the technology but so you mentioned, I mean, so you can basically connect to all of the equipment in, you know, in a day or so. And, and you gave the example of basically, you can instantly connect a sensor to to monitor the power consumption by each piece of equipment. I think this analogy with the health monitoring watches is great to get yeah okay so you can you can record heartbeat and that's that's super useful. How do you go from that to sort of workflow optimization I can tell how the user can look at that and say okay, maybe get some valuable insights into downtime and uptime and but but how do you how does the user sort of take that and kind of learn how to, or does it sort of suggest reconfigurations to optimize workflow and minimize downtime. Yeah, yeah and there are some things we can do and then some we can't. So the things we can do is if you're on key equipment and with that equipment you tend to care about three things is it running at the right speed at good quality for any manufacturing nerds that's overall equipment effectiveness availability performance and quality. We are really darn good with just that simple power draw at is the machine running at the right speed. We can predict quality anomalies but we don't track yet good parts automatically people can enter it in and layer it on and we can build lots of algorithms that can associate with that. We can get those two things of is the machine running at the right speed for a lot of manufacturers that is a huge improvement, so we can give them that for all of their equipment right away. We don't typically use our software to guide around what the bottleneck is, you know, flow, etc. It's more giving their continuous improvement teams and the folks who already care a lot about that stuff, the accurate trusted real time ability they haven't had, rather than, you know, having people do that job over time, and we've already built a bunch of algorithms that of course make that data much more useful and insightful benchmarking, being able to automatically calibrate and derive cycles for example from small fluctuations of different components of the power draw different equipment. So normally detection being able to get really predictive when a hidden pattern change happens and send the right alert to the right person I can give a little visual of what that can look like. And of course there are some alerts as well within the system even AI powered it's just hey this machine started to draw weird amount of power compared to usual. What we're doing is we're building assistance in for the humans within the manufacturing environment, not trying to automate every different piece of the continuous improvement will close that they have. Did that answer the question though. Yeah, no that's been that's fantastic. And then, with this impact on sustainability. So can you walk us through an example of of how I mean I can see how you want to you obviously want to, you know, maximize the efficiency for the for the machines. How do I think about the opportunities to improve efficiency and I guess a little bit more specifically. When I think of processes in a factory. It, it seems like there's not a whole lot of room for dynamism like you basically want things to sort of be running constantly. Then linked with the sustainability outcomes or are you finding opportunities where you can actually power something down to save on energy and save on emissions and then and then ramp it back up and without a negative effect on the on the overall process. Yeah, the short answer is that there's lots of opportunity. I'm realizing I should have, I just I wanted to take direct screenshots from our customer talking about this and I should have captured more detail here but I can certainly share lots of examples of customers talking about the specifics because the, the data empowers a lot of different solutions depending on the industry but the core principle is the same which is we've got the ability to get not just the core production lines visible, but all the utilities compressors chillers pumps dust collectors. You know, cooling towers all of the equipment that surrounds the auxiliary equipment surrounding the production lines as well. And that's often where there's good opportunity to so compressed air for example. But, you know, it's not a question of are there leaks and your compressed air system and how much are you losing so that type of thing chillers insulation. And that is where we can certainly make recommendations and surface things and there's lots there with production equipment it's usually let's minimize idle time and of course these are huge machines so that can have a huge impact with compressors with chillers we have all these playbooks around those types of things pumps. And that's all within the daily operating rhythms of here are the easy process changes that cost very little work but once you see them you can easily have the right person take the right action and guide will can support a lot. The other thing that's been interesting to see is is the catbacks impact of being able to size and finance more equipment, more efficient equipment, because you have the right measurement systems in place to actually see what a more efficient air compressor efficient injection molder would do in terms of payback and then also to measure and verify it so you can get the utility credits that often are out there but waiting because so few folks have those the measurement and verification systems in place. So there's tons I wish I had the short answer of this is always a bullet. Yeah, lots of lead bullets. But when you talk with the manufacturing engineers they usually know what they are, and it's about giving the right person the right system. And also, we feel very strongly it's about having a system that's integrated into production. That's why the the premium example of go back to the one I shared about the before guide wheel. It's been really inspiring to see this has been lots of other customers as well but visiting their clients and seeing the engagement of the operating team and maintenance team and production team and saving energy and driving sustainability goals. That doesn't happen if you've got an energy system or a carbon system that's off to the side sustainability as a side project. We're big fans of let's integrate this in and that's where making that data powerful for all the teams, especially the production team because production wins every fight on the plant floor is really critical. And so, are you involved in making the sensors at all or is that sort of off the shelf technology and then related to that. Do you sort of see in the future, going after other sensors to try to get some other data from certain pieces of equipment to aid your algorithms. Yeah, we actually already have lots of other sensors that can connect in so temp, humidity, flow pressure, lots of others we built the system to be sensor agnostic we just love starting with current because boy that's powerful. In terms of driving really fast bottom line impact which is what we usually see is the key thing for our customers. In terms of whether or not we build the sensors, we actually love that sensors are really commoditized and falling in cost like crazy. And so we built from the start to be sensor agnostic because of that can use lots of sensors and love. We did not invent the sensor that the smartwatch you know, using the electrical power as a sensor called a current transformer that we can actually sell remotely to manufacturers because we can say to the maintenance person. Hey, have you used a CT sensor before a current transformer and they usually say yes and recognize that it's not intrusive and safe and risk free and that this is something you can clip on the equipment without any any security risk as well. Well, thank you we are unfortunately out of time this has really been fascinating and we really appreciate you sharing a guide will story with us today so thank you learn. Thank you. And now so far our final speaker is rough con from a company called vanilla. So now we've moved from the factory to to buildings for for residents. And vanilla is a software platform really at the intersection of climate tech and AI and fintech that enables energy efficiency at scale buildings I think it's as many of you may know are huge energy demand. I think something to the tune of when you when you count energy losses in the in the generation of primary power something to the tune of 40% of us energy goes to buildings so rough it's great to have you and we're excited to hear about vanilla. Thanks so much. Hopefully you can see my screen. Everyone I'm Ralph, the founder of vanilla, we're building an AI powered marketplace for clean home upgrades and reinventing how clean tech is bought and sold. My co founder Max and I were both involved in clean tech companies in the past. About a year ago. In a Stanford class we took together we learned that if you were born in a community with high pollution. And especially if you were low income or diverse, you're actually likely to have lower SAT scores. Now there are many ways to impact that but what was clear to us is that everyone deserves access to a clean environment. Technology like EV heat pumps and more can improve those communities, but they're still out of reach for most. And so we came together with a mission to make clean tech accessible for all. A little bit more about us. So I'm Ralph from Stanford MBA, former materials engineer, where I worked on fuel cells. I've worked across climate tech where I helped UPS basically electrify their fleets and their depots. And so the challenges that even a large company like UPS faced. I also worked in FinTech at a series B company that helped small businesses like trucking fleets and plumbers. Max is our CTO and he's an engineering master student at Stanford was at Berkeley CS program and previously was a software engineer at Amazon and machine learning engineer at Volvo. And another leader at a climate tech startup as well. We're lucky to have the support of several organizations, including Tomcat and where we've also won a few awards on campus as well. So let's zoom out for a second. There are 100 million buildings in the US that make 40% of our greenhouse gas emissions. And we have to make these clean. There's lots of great progress in retrofits. Certain states, for example, are targeting increasing the number of heat pumps sold by forex by 2030. There are tens of billions of dollars available through the IRA to fund clean tech, electric vehicles and more. But after hundreds of interviews with companies and people, we know today that clean tech adoption is really hard, specifically buying and selling it. And we've distilled those challenges into three areas. One is discovery. The second is economics and the third is process. I'll share a little bit more about that before I move on to what vanilla does. So in terms of discovery, there are 80 million homes in the US and every single home is unique. This creates a lot of friction. Let's imagine that you're a homeowner or landlord upgrading your home. We've heard these questions consistent me. You know, which product should I buy? Why should I even go clean? Will my energy bill go down? Which contractor should I use? And can I actually afford it? There are generalist platforms like Yelp and Angie's List, but we've learned homeowners don't trust these for larger renovations. Instead, what they end up doing is asking their friends and neighbors for recommendations. We've also learned that platforms like Angie's List are not actually liked by contractors. We've heard stories of contractors paying $400 per home before they even get to the front door. And when they knock on the door, sometimes that customer isn't suitable for clean tech. There are new clean tech companies trying to reach these homes directly to avoid these platforms, but they have a really hard time doing it because they're aggregating all of that demand themselves. So one question that we thought about is can we connect homes and contractors in a better way? The second big challenge in our space is economics, and that's because upgrading to clean tech is really expensive. A new water heater, for example, could cost $20,000. It might help you save energy, but it could have a payback period over 10 years. And if you're a homeowner, you may know that there are lots of incentives and tax credits available, and actually many of these are targeted towards lower income communities, but they're super complex to learn about. They're super hard to apply for. They change often, and contractors believe that they create unrealistic expectations for the customer. And for contractors, remember that $400 that they paid to get to one door? Well, because every home is unique, they often find that a home may not be suitable to add, for example, a heat pump. The home in this image in the top right has lots of air leakages, making it a poor candidate for heat pumps. Even having windows in a certain orientation or even the material that frames a window can make a home unsuitable too. A contractor usually only learns that after buying the lead and then visiting the home, which wastes their time and money. Contractors actually describe this sense of a home's suitability for clean tech as intuition and a feeling. They do calculations to refine that feeling and add some science, and they also do online research to validate their view. But if they get it wrong, or worse, a homeowner insists in getting the heat pump, even if they're unsuitable and laser becomes unhappy, that deal ends up losing the contractor a lot of money. So we thought about, can we help them eliminate the first visit and improve their unit economics? And then lastly, contractors do lots of manual work in Edmund, for example, filing those rebate applications. We learned that they actually hack together lots of different tools to evaluate customers. Many of these tools don't give the right answers, and they actually learn how to adapt the answer to get to something that's more accurate. A very consistent quote we've heard is, you know, we just want to install, we didn't sign up for paperwork. And that also makes sense when you realize that paperwork is not billable time. So we thought, you know, can we automate the workflows so that they can basically spend more time installing? And that's the heart of our product vanilla. So we're helping match clean tech contractors and homes, and we're using AI tools to basically increase the sales velocity. Our big bet and hypothesis is that AI can really help this process. And we use it throughout our product. And I'll explain more now. So the way that vanilla works from a homeowner's perspective is that they take photos of their home. We analyze those images and recommend upgrades, prioritizing clean ones with all of the different incentives included. And we target at those recommendations to their home specifically. That solves that discovery challenge that I mentioned. We then match the home with a contractor and we manage all of the different interactions between them. This is an example of something that has actually proven quite popular with some of our initial customers, which is a portfolio of their incentives that they've captured where they can see and keep in touch with all the different incentives and, you know, contractors that they're communicating with. What's really cool is because new incentives are launched every few months, their desire for up and because the desire for upgrades changes over time. They actually use the app quite regularly. So onto some of our features for contractors. Well, we've created a queryable earth GPT we're working on that name. But it helps contractors basically query the community and decide which areas they should prioritize to and sell into. So for example, a contractor could ask, show me all the homes in the Bay Area that have solar already and have lots of rebates available. They could even ask which homes on Tolman Drive face the sun optimally and have fewer than 12 windows. We've trained our models on public and private data and they are quite accurate. With AI, we can also do other things. For example, trying to eliminate that first visit and replace that intuition feeling. So with AI, we can see what upgrades are possible. We can even find new opportunities for upgrades that the homeowner or even the contractor wasn't aware of. In this case, lighting, for example. We can predict the energy savings from those upgrades. Customer service teams can actually from the contractor can see if there are any red flags that would make the home unsuitable down the line and deprioritize those leads. And with AI, we can even create 3D reconstructions of the home so that a contractor could plan their visit before they ever step foot in the door. Before this was really hard, but with new technology, we can make this much easier. After a contractor analyzes this, they can then choose whether to accept the lead or not. Homeowners like this too, because if you're a homeowner, it means fewer visits, less scheduling, and a contractor can be in and out quicker, which basically means less billable time. And they can manage all of the interactions on one single platform as well. What we're really excited about as well is because we can help contractors find the right home and qualify them much more easily, they can reduce their prices to serve that home because they are really confident in that deal. And hopefully that increases access as well. So we have some, you know, early traction and we're getting into homes and hearing some great feedback from different sides. We also believe that we can help communities not only become cleaner, but hopefully create more jobs for trades people and contractors and help them basically run better businesses and even learn how to install new types of clean tech, which should also increase adoption. We didn't have much time here so there are many things that I didn't discuss like our relationship with renters, OEMs and other clean tech startups, some of our business procurement services, network effects and so on, and I'm happy to speak about those offline. So we're live in some homes now we find some contractors and we're growing. I think just some asks from this incredible audience that's here today. So if there are any introductions to home and contractor associations, you know, state and city governments, other clean type companies or even advisors, that would be amazing. And if you are thinking about upgrading your home, let us know. I'll pause there. Thank you. Great. Thank you, Ralph. Just tremendous possibilities here just thinking about all the things that you can do. I guess I may have missed it, but maybe you could just tell us a little bit more about the business model. So, so, you know, is it a subscription service for the contractors? Is there, you know, basically a fee per transaction or how do you envision this working? We're trying to build supply and demand. And so our application is free to use for homeowners and contractors can accept leads at first for free and to use some of the incremental AI qualification tools, we then charge our subscription. That's the business model today. In the future we can look to add things on the payment side and take rates and things like that. But at this point in time, we're just really focusing on helping both sides of that marketplace. Okay. So it's just sort of focusing on growth right now and then figuring out how to monetize optimally down the road. So, I can see how it's relatively straightforward. This is coming from someone who knows very little about software but relatively straightforward to train to identify incentives or maybe you don't even need to train. That's, it can be sort of hard coded in, but in terms of identifying problems, because you gave this great example, the contractor goes out and, you know, they're using their intuition and experience. And as they look at the home, they realize, okay, maybe this is not a great candidate for, you know, a heat pump because of leaks and whatnot. So, can you give us a little bit more about how you, what sort of data you need and how you train the algorithm to spot some of those problems that would sort of alert either the homeowner or the contractor that okay, maybe this isn't the right upgrade for you. So, we combine two different approaches. One is using satellite imagery and spatial data on the building. And then secondly, data from those images that a homeowner may take. For those, we pull something like 10 characteristics, which can, for example, the dimensions of a room or the thickness of a wall or things that may influence heating and cooling loads, for example. And combining those different approaches, we can basically give a quick answer to a contractor as to whether this is a good candidate or whether they should think more about taking on this lead. Sales velocity is really important for contractors because they run, you know, in some cases quite difficult businesses. And so, helping them find, helping them serve that lead much faster, helping them maybe find new opportunities if they came in for one specific upgrade and leave doing maybe two or three upgrades. And also finding more incentives to basically increase the size of the pie. It creates a much more favorable interaction and ultimately hopefully helps the homeowner as well. Great. And I want to go to a couple of the audience questions. These are a little bit more specific. So, one person has said that they've upgraded with heat pumps recently. The incentives are confusing and asking if you've considered partnering with organizations like the Silicon Valley Clean Energy. That's one. And then the second one is, what is the reception among program operators retained by utilities examples being up light. Clever's old ICF and Franklin Energy. Yeah, thanks for those questions. I'm smiling because they are things that we're thinking about. And so I think partnerships for us will be super important. We're speaking with some of the CCAs, some utilities that are local, and also some city governments. And that's because incentives operate not only on the federal level and the state level but also on the city and even utility level. And so combining all of those incentives, confirming eligibility, doing all those different types of paperwork, making sure that it won't eliminate you from getting a tax credit in the future is complex. And we're hoping to partner with those organizations to help them meet their sustainability goals as well. That's why they created those incentives in the first place. So that's definitely something that we're thinking about. I'd love to speak to anyone offline who has any ideas on that as well. In terms of the second question, so I think that ties to something broader, which is helping other businesses use our platform in the future to accelerate their own go to market motion. That's something that we're thinking about and companies like the ones that mentioned may be great partners to do that. What we've noticed now is that incentives mainly focus on the residential side. There are commercial incentives, but it's a bit few and far between and also have more complex application criteria. And so where we think we can get the most impact fastest is on the residential side. And again, are happy to speak to anyone on that as well in terms of partnerships. Great. And just last question, really big anxiety for homeowners is finding good contractors. So is there a vetting process so that homeowners who use your tool have some assurance of the quality of the contractors. So we didn't get to speak about this in the presentation, but I mentioned network effects and and basically what we're learning is each community wants to share their contractor or their good experience with others, whether it be neighbors or other people in their community. And so we're actually getting and creating systems for them to do that scalably. Another example is when someone decides to upgrade their home and add a heat pump, they may require an electrician to make a panel upgrade, for example. And we also help sort of create that plan and recommend the best other contractors that would be involved in in that plan. So that's how we're capturing learnings from different contractors. Perfect. Well, thank you Ralph for telling us about vanilla I want to just thank again all of our speakers Tyler Lauren and Ralph for really giving us some fascinating content today. And as hopefully you saw in in the chat, we would really encourage everyone to connect with the founders and the larger Tomcat ecosystem by joining our LinkedIn Tomcat network, networking group that link is in is in the chat so please do that if you're interested in connecting with any of these founders or any of our other teams. And on behalf of the whole Tomcat center. Thanks again very very much everyone for for joining us today. Have a great evening.