Latest / Autonomy Insiders / AV Charging Infrastructure: How Terawatt Is Building the Physical Layer Autonomous Fleets Need

Transcript

AV Charging Infrastructure: How Terawatt Is Building the Physical Layer Autonomous Fleets Need · Autonomy Insiders

• Daniel Abreu Marques: Welcome to Autonomy Insiders. I'm your host, Daniel. And today we are talking about a topic that doesn't get nearly enough attention in the AV world, charging infrastructure. So everyone talks about the sensors, the software, the regulation, but charging is often overlooked. And my guest today is building the necessary physical infrastructure. Today I'm joined by Peter Cohen, a general manager for AV and rideshare, Terawatt Infrastructure, a company that has raised over $1 billion to build to day with Peter Cohen. Welcome to Autonomy Insiders, the show where global industry leaders in autonomous driving unpack their real-world insights. fleet charging hubs across America. So Peter, welcome to Autonomy Insiders and so happy to have you on the show today. Peter Cohen: Yeah, thanks, Daniel. Happy to be here. Daniel Abreu Marques: So Peter, let us start at the beginning. So I think many of our listeners know the AV companies, the Waymos, the Zooxes of the world. ⁓ the infrastructure layer behind them is much less known. ⁓ let us fix that. So for people who have never heard of Terawatt, in your own words, what is Terawatt and what is the service you provide Peter Cohen: Yeah, that's a great question. So we're a full stack developer of EV charging hubs for fleets. As you mentioned, were started five years ago with the idea that, you know, critical infrastructure, which we believe charging is for these fleets, ⁓ a lot of space, power, capital. It's very complicated to develop. So data centers or cell phone towers ⁓ for relevant industries, we decided that, you know, starting a company such as Terawatt would be crucial to support these fleets as they grow and electrify. So we are present in the United States today building hubs across the country where we acquire real estate, make sure it has the proper zoning permitting, is in a location that's relevant to a fleet, which for autonomous and ride share means, you know, dense urban areas near airports, depending on, know, what a city looks like. That could be a couple of different types of locations. Make sure it has a lot of power. And then we fund and design and construct the site and then operate it to a 98%, you know, uptime. for the chargers to make sure that this is working every time that a vehicle needs to plug in and is ⁓ in ample amounts of power in that relevant location. Daniel Abreu Marques: So if I'm running an EV fleet, what exactly changes for me as a customer when I work with Terawatt versus say just buying the chargers and installing them by myself? Peter Cohen: Yeah, I mean, simplification. think what we provide is a replicable platform to work with our partners. You know, maybe you have a small scale operation and you have a couple of chargers and you have a facility, you know, converting to an electric fleet. Sometimes you can have an electrician and come and you get, know, that can work just fine. But for scaled operations, you know, our sites can be anywhere from five to 10 megawatts of power, which is as much of a skyscraper. That's really hard to find and develop. So we take all of that complication on ourselves and let our customers focus on their fleet operations, what they do best. don't have to figure out how to build out this charging infrastructure. They can just focus on their core business. Daniel Abreu Marques: maybe you could also walk us through a typical Terawatt charging hub and how does the charging hub actually look like so maybe what how big is the typical side and also what range because I think there will be differences for sure. Peter Cohen: Hmm. Yeah, it depends on what we can find. I think the real estate and location drives a lot of what the ultimate design of the site will look like. But for simplification, you can imagine looks like a parking lot. You know, these are trying to be as fleet focused as possible and as similar to existing operational hubs, but obviously with a lot of electrical infrastructure, but then somewhere between 30 to 90 charging stalls or parking stalls, depending on how many actually have a charger or some are just parking. Gated. and privatized, so security, as well as making sure that only the right vehicles can get into those facilities. Having the necessary data offload infrastructure, which is very relevant for autonomous vehicle fleets. That can be done in a number of ways, but is required element of the site, as well as an amenity building, which houses a small break room for drivers or operators of the fleets, depending if they're manned or autonomous, which can have vending machines and lockers and a place to sit while the vehicles are charging. as well as bathrooms, obviously, and then space for the IT equipment and spare parts for chargers. That's a key one. So all of our facilities have spare parts on site, which helps with the managing of the charger uptime and maintenance. really what you get is a nice private Sometimes they just blend into the surrounding ecosystem and you would never know that they're there. ⁓ We to make sure that these are working really closely with the local authorities having jurisdiction to ⁓ meet code and just ⁓ be facility on the street that you would never know is not some scary special facility. obviously there is a lot of electrical equipment from the chargers and the necessary upstream equipment as well. Daniel Abreu Marques: Yeah, that sounds like a huge project to build. So I'm really curious how much does it cost to build a site like that? And maybe also how long does it take from a decision to build that to the first vehicle actually charging? Peter Cohen: Yeah, so the price will be dictated mostly by the real estate costs. So we do acquire property, which strategically for us and for our customers gives us a long-term view on the property. We can maybe phase in power over time, it does allow us to amortize things on a longer horizon than if we were to lease a property or something of that nature. So, but the real estate cost does drive a lot of the answer to that question. But could be, 10 plus million dollars of infrastructure costs. The charges obviously are you know, not, not cheap. so it does scale pretty quickly. And that's again, another point of why we exist is to turn that capex burden into an operating cost for our customers. So they pay us a fixed monthly fee for access to the facility. ⁓ we take all the capex to actually invest in the property upfront. And so they can offset that cost to their operations. You know, if they're running a fleet, they have monthly revenue. Now this is a part of their line item of costs, but not something that they have to put upfront before they even start operations that. And that's again, another point of why we exist is to turn that capex burden into an operating cost for our customers. If you're scaling five, 10, 20 facilities can become quite a big number. As you mentioned, we've raised a billion dollars, which is a great starting point, but it's definitely going to need to be a bigger number than that for the scale of this industry as it grows. So that's kind of the cost piece in terms of the timelines. assuming ⁓ have found a site, which obviously the first thing we do when we're either engaging in a search on behalf of a customer, if they say we want to be, I'm in San Francisco, if they want to say, I want to be in San Francisco, ⁓ with a site these three circles, or ⁓ the we want to be closest to, we can go look for them on their behalf, kind of like a bespoke search. And in case, obviously we're finding what's available, ⁓ the zoning, power, as I mentioned, the other layers of need for a site to be viable. That can be an open-ended search depending on how quickly we can find something. But assuming we found a site, the various timelines will be dependent on design process and working with the city for the associated permitting. of the layers of construction, as well as the utility. The power timeline is a big one. We've gotten lucky where we've found power and it can be ⁓ pretty quickly. But that can be the long pull in the schedule. So rough could be anywhere from a year to two years from site acquisition to actual commercial operations. Daniel Abreu Marques: Makes sense. you mentioned now a lot of filters. So for an AV charging hub, you need then the location that sits inside the AV providers ODD that overlaps with rider demand, has sufficient grid capacity, also the zoning rules. And it's also close enough that dead heading to the depot does not kill the fleet economics. So how many potential sites typically survive that filter and what is the biggest constraint of them? Peter Cohen: Mm-hmm. Yeah. I think we have a statistic where only 3 % of the sites that we look at actually get acquired, which is quite a low number. that's taking consideration everything you said. I think power is probably the driving one of what kills a site and kill meaning something looked at that actually isn't worthy of being acquired. Either the timeline acquire that or to develop the power is too long for it to be relevant for an investment and commercial operations. You know, maybe it's only four years away and the customer needs to be ready in less than that, or we want to be ready in less than that. Um, that's probably the driving one. Um, but there's other design elements that feed all of this feeds into, you know, the cost per vehicle mile that the fleet, you know, I guess it gets by nature of using one of our facilities, which is all of the pieces go into that. That's how much charging they can have at the site. Uh, the cost of all of the pieces of the infrastructure, as well as the utility cost on top of that to see how many vehicles can they push through in a given day. ⁓ the more they can obviously lowers their cost per vehicle mile as well as deadheading, as you mentioned. So having a site. In the center of their ODD is optimal from a deadheading perspective versus if they have to drive 30 minutes to get to the charging site, that's 30 minutes of that they're consuming in the vehicle. That's cost. It's not generating revenue as well as 30 minutes there and 30 minutes back to the center of demand. So an hour essentially of lost rider revenue. So ⁓ trying to, you know, be really thoughtful with our partners to understand what does that whole calculus look like? Because sometimes it might be worth paying a premium for a facility to be closer to demand because ultimately the dollar per vehicle mile that the fleet experiences is lower. And that's a lot of thoughtful partnership we do with our customers to understand what that looks like and where the sweet spots are. Daniel Abreu Marques: Okay, so I think we have now a good understanding of the product and what it takes to build it. And I think it's now worth get a deeper understanding to understand you are building for. So maybe ⁓ quickly explain who are Terawatts today when you look at portfolios or what types of leads are you serving? Peter Cohen: think you do a great job, Daniel, of kind of looking at this ecosystem. It is rapidly evolving. you know, first and foremost, ⁓ have a really great partnership with autonomous vehicle fleets such as Waymo and others. And some cases, they are the ⁓ ones decisions on the charging locations. But there's obviously a rapidly evolving ecosystem where operators, other fleets that are not autonomous, whether it be rideshare fleets, you know, I personally oversee the autonomous and rideshare division at Terawatt we do have a trucking division, which looks at logistics class eight large semi trucks here in the US, which is a growing market as well. So we are agnostic to who the customer is, but we found, you know, a really strong fit is with high powered need and concentrated locations, which for every one class eight truck, is a big, that is a big battery that can charge very quickly, needs a lot of power. And, you know, locations can be in warehousing hubs or near the port, for example. And then obviously autonomous vehicle fleets can need a centralized location near an airport or downtown district, as I mentioned, and push a lot of vehicles through. So we are not doing public charging is I guess where we don't focus. But for fleets, we are but have found a lot of really strong fit ⁓ with those vehicle fleets as well, like I mentioned. Daniel Abreu Marques: Yeah, makes sense. And you mentioned that you also are in trucking. So where are biggest differences between building charging infrastructure for an autonomous robot taxi fleet versus say, a heavy duty trucking fleet? Peter Cohen: Yeah, I would say the end result of what the site looks like is similar, just different scale. The trucking size parking stalls are much larger than the autonomous vehicle ones. And the resulting property might be, you know, two to four acres of a size where our autonomous vehicle fleet sites might be half an acre to an acre and a half. So about half the size depending on what we find. But ⁓ data infrastructure is obviously much more specialized for autonomous vehicles, ⁓ vehicle fleets and the location is the biggest difference. know, trucking have to be relevant for trucking operations. Autonomous vehicle sites have to be relevant for autonomous vehicle operations, which is either ⁓ like I mentioned, or near an airport or where ride hailing happens, which I'm sure you and all the listeners are familiar with versus trucking is much further out of location, much more industrial in terms of the areas that those sites are located in. Daniel Abreu Marques: And are you also looking to expand then into the combination of both of your customer segments? So in autonomous trucking verticals, Peter Cohen: Yeah, I mean, we'll see over time. Definitely, you there's a lot of excitement in the autonomous trucking industry in the US, especially those those fleets today are diesel trucks. I think, you know, they're obviously trying to solve the long haul problem where going autonomous can add the most value. But over time, as those vehicles become electric, for sure, we think that, you know, having the expertise and understanding what an autonomous ride handling fleet needs for a facility will translate somewhat nicely to what an autonomous trucking fleet will need as they electrify. ⁓ So yeah, we definitely want to support all fleets as they electrify and become more sustainable. We just found a really a lot of traction with the autonomous ride sharing fleets by nature of what they need and how quickly they're growing in the US. Daniel Abreu Marques: Makes sense. And now without naming anyone, you cannot name them. But how does a typical customer relationship start? So does an AV company come to you and say, we need a charging this city by this date, or does Terawatt build speculatively then also try to attract the customers? Peter Cohen: Yeah, I think we see, you know, charging infrastructure is becoming the crucial bottleneck for expansion of these fleets and operations. And so we're trying to, you know, we have the flexible capital to do our part to go out and buy as many sites as we can and build in the relevant locations and cities to support the, you know, the build out of these fleets as they expand and receive regulatory approval and launch operations, which is extremely exciting. You know, as I mentioned, these facilities can take a year to two to build, is a lot. longer in some cases than the adoption and growth of this technology and how quickly someone can go from a 10 vehicle pilot to real commercial operations in a city, which this year is very exciting to see. So many announcements that are translated, historical announcements that are now translating to actual operations. so we don't want any fleets to be in a pinch where they're ready and they don't have charging. So we are making speculative bets in certain cities. However, are finding that with most of our customers and the throughput. They are aware that charging is a bottleneck and are looking to us to help them grow and work, you know, in hand in hand to say, here are the cities that I want to be in. Can you find sites for us? And we have like a, you know, a bespoke relationship where we can provide design feedback and find the sites that fit their exact needs. So we do both as the answer. We both we find sites and are pushing forward in the industry to offer up inventory to fleets where it's relevant to them, as well as taking their input. and looking for things proactively with them together. We can do both. Daniel Abreu Marques: you mentioned that when you build this charging depot at the scale you operate, you need massive amounts of electricity, which is probably comparable to a small industrial district. from what I understand, just getting that grid connection approved and built by the local utility can also take a lot of time, many years probably. This means you essentially have to decide today where to build, even though nobody knows for certain which cities AV companies will enter in 2028 or 2029. How do you manage that and how do you then decide where to place those bets you mentioned? Peter Cohen: Yeah, we're thankful to have an amazing team which draws upon experience from past industries, whether it be fleet related, data centers, construction of high powered ⁓ And ⁓ comes with that is an understanding of how to work with the local AHA, the authority having this jurisdiction, as well as the utilities in a number of markets. So what we've been doing for the past couple of years and proactively are expanding is working with cities upfront. And I think this is the most important thing to understand. ⁓ we know we're going to come to a city eventually, do the work today to understand where can we actually place these facilities. know, ride share is in every city in the US and we can look at understanding where those patterns of behavior are today for where a customer we think might want to go. Work with the city to understand their zoning and where appropriate sites should be placed. And then also be very proactive with the utility to understand the power maps, their design process, to have a lay of the land so that when we're ready to pull the trigger and buy a site. We've already done the work to understand where we should be looking to have a much more efficient process for our own search, any customer searches and directing them so that no one has to find a great site and then later realize it doesn't work for zoning, it doesn't work for power. So we're trying to be very thoughtful and proactive. And another benefit, you know, back to your earlier question of the value that our platform brings to our partners is understanding that process at scale across many cities. It can be very painful to try and go through that. And another benefit, that ⁓ our platform brings to our partners is understanding that process at scale many cities. It can be very painful to try and go through that. city by city and aggregating that knowledge. The first site we build in a city, we'll learn that process. The incremental second, third, fourth, fifth site we find, we can leverage that existing experience and relationships with those parties to understand to make it faster, better, and more ⁓ efficient for city by city and aggregating that knowledge. The first site we build in a city, we'll learn that process. The incremental second, third, fourth, fifth site we find, we can leverage that existing experience and relationships with those parties to understand to make it faster, better, and more ⁓ efficient for Daniel Abreu Marques: Makes a lot of sense. And I have a thesis, but I think I know your answer already, but I just wanted to get this from So the thesis charging infrastructure could become the single biggest constraint on how autonomous vehicles scales, not the software, not regulation, but the physical infrastructure. So do you agree with that and ⁓ why or why not? Peter Cohen: Okay. Unsurprisingly, I agree. know we talked about this in person as well. I ⁓ it's all the layers of things that have to go right from what we've been talking about today. ⁓ are a number of pieces that 3 % success rate ⁓ looking for property to then ultimately find something that we think works for our customer's use case is a very low hit rate. We're trying to be more efficient and drive that up by nature of taking our learnings and looking for things that are only appropriate. But a lot of that is things that are just out of control and have to go right. by nature of all of those things, in different layers of process that just take some time, we found a lot of ways to pull in that, call it two-year development timeline closer to a year or even faster if we can. But a lot of these overlapping layers are processes that just do take time. so while technology is rapidly evolving, finding this critical infrastructure, once you find it, You know, and you have an operating site that is a strategic resource that enables operations, like I said, we're acquiring these properties in perpetuity, securing the grid connection and having that much power is a permanent scarce resource. And part of why we believe, you know, there's value in the solutions we provide to our customers for a long-term partnership. But yeah, we do believe heavily that having access to these sites and finding them and securing them and building them. is crucial for ongoing success. As well as we haven't really talked about so much today, but actually, you know, once we build the site, operating and maintaining the charging infrastructure at a high uptime is a crucial part of having successful operations at a facility. You know, if you have something and the chargers don't work, it's not really a helpful site. So having all of those stacks together and being able to operate them at scale for a reliable, consistent charging experience requires a lot of complexity that, ⁓ thankfully, had, you been building for the last five years and are now really just expanding very rapidly. Daniel Abreu Marques: Yeah. And now I want to bring up a cost factor that I think ⁓ most of our listeners have never heard of. Demand charges. So can you explain what demand charges are, why they matter so much for a business like yours and also how Terawatt deals with them? Peter Cohen: Mm-hmm. Yeah, so essentially, just to simplify, working with utilities, know, this much power, any power, you know, they have a lot of infrastructure that they've developed and they want to make sure that, you know, they're putting up electrical capacity for you, as well as in general ecosystem that the site might be part of, you know, in downtown warehouse district or near the airport, et cetera, that they're managing load from customers and also having an efficient use of, you know, providing electricity to everybody. It's very complicated and there's a lot of factors that go in. So when you pay for electricity, you know, have your, your new normal rate where maybe during lunchtime it's cheaper than at dinner when everyone's home and turning their lights on, et cetera. So the cost of electricity can change during the day as well as over, over all of that. And it varies utility by utility. There is a cost for just the maximum draw you have, uh, throughout the, throughout the day. Um, so essentially there could be charges on top of your utility bill for your peak usage at any given time. that to say is this is obviously getting very complicated, even as I'm describing it ⁓ gets even more complicated where thankfully we have a very sophisticated utility team managing all of this. But there is a lot of work that you can do to manage how you use the energy at a site. ⁓ can integrate batteries to ⁓ down and use energy that you charge up the battery with at a ⁓ time and dispense it onto the site at a more expensive time or back to the grid to help them with their peak demand. It's all about managing supply and demand of electricity on the grid. so there's a lot of strategy you can take by how you use the chargers, how you work with the fleet to make them aware of ⁓ time usage, how they're managing towards their utility bill, et cetera. It a whole nother factor of, we've talked so much today about just getting sites in locations and getting them ready is complicated. Then there's obviously all the efficiency to be done of when you actually have the site ready and you're using it. How do you make that a very cost effective operation? There's a lot of variables of how you can do that. Obviously, right now, I think a lot of the industries just get to market, get customer experience going and make sure they have market share. But over time, I think we're going to see a really strong emphasis and already are starting to see that on optimization of fleet operations and making sure that you can bring that cost per vehicle mile down as low as possible, because ultimately that will translate to ridership costs, which is helpful for the whole ecosystem. All that to say for our facilities, we're very aware of this and trying to make sure we're picking the right tariff. So a tariff is like when you have a facility, some utilities have EV specific tariffs. It's like the actual rate plan you're on and those have different usage. How they charge for the electricity depends on time of day usage in these demand charges. Some programs do not have EV specific rates, but picking the right rate can have a lot of impact on what the customer. ⁓ experiences at that facility. Daniel Abreu Marques: All right, this sounds like a real complex problem. And so your customer probably really happy that your team of specialists is handling that for them. So good offer. And ⁓ another point that your CEO, Neha Palmer often draws the analogy that AV fleets need charging hubs the same way AI models need data centers. And I really like that comparisons. But data centers have also ended up with like three. Peter Cohen: Yeah. Daniel Abreu Marques: hyperscalers that dominate and do you think fleet charging infrastructure will consolidate in the same way? Peter Cohen: It's interesting. think we kind of touched on this a little bit of who is responsible for, for, you know, finding the charging infrastructure. Is it the technology provider? Is it the TNC, you know, like an Uber or Lyft that's operating the customer app? Is it these operators that are emerging in the ecosystem that are taking over responsibility of the fleet to provide services to the technology providers? I think it's so early to tell, but I think what's core and all of that is you need a place to charge. So. You know, we've worked with all sorts of folks across the ecosystem and it's very exciting to see different business cases emerge and different use cases for who's responsible. So I'm not, think it's too early to tell honestly. think it depends on where ⁓ can provide the most value in the stack as that stack evolves for, for who's operating and who's responsible. But I think what we have seen is that there is a strong recognition that charging infrastructure is complicated as we've acknowledged so far. ⁓ is critical for operations as well as a strategic resource to enable profitable business. a facility downtown in a certain city that's having that power and zoning is scarce, but having that site for your operations ⁓ that if you didn't, you'd be at a disadvantage. You might be having fewer rides or more expensive platforms. So all to say is I think we're seeing a lot of interest across the value chain for folks who want to figure out how they can incorporate the charging in the right places into their offering. Daniel Abreu Marques: Yeah. So you already touched on Uber. So also recently announced that they have a hundred million plus investment to build its own EV charging hubs in cities like San Francisco, LA, Dallas. So that's a ride-hill platform deciding to own the infrastructure layer. So how do you see that move? And what does it mean for a company like Terawatt when a potential customer decides also to become a competitor, to speak? Peter Cohen: Yeah. I wouldn't see this competitive. I think the pie is so large that, a hundred million dollars can get you a couple, a few sites. We talked about, you know, these can quickly scale past $10 million per site. So for Uber scale of a platform and how large they need to grow, you know, that's, that's definitely it's helpful to start, but it's, there's going to a lot more to come. Like I said, like we were personally raised, you know, a billion dollars so far and think there's, a lot more need beyond that for the platform and for the ecosystem. ⁓ I think it's a recognition of the strategic importance of what charging is. the pie is so large that no ⁓ player is going to do this themselves. Similar to data centers, there are ⁓ and Amazons, whoever all would build it themselves and also rely on third parties. It's a combination just because of the scale demands so much supply. ⁓ Daniel Abreu Marques: Yeah, makes sense. And I also was wondering if there's also like a takes all or winner takes most dynamic in charging infrastructure where the first company to look up real estate in key AV markets create an really insurmountable moat because there is not so much free land so much ⁓ real estate buy. I think if you the first one you could probably a huge moat So what is your opinion on that? Peter Cohen: I think there's a combination of the first being, you know, obviously having a good access to what is available today, more so like there's this flywheel of understanding the process with the city and utility and being able to replicate that. If you do have a good understanding and are able to establish relationships, understand where power is or is not. It'll help you move faster versus, you know, someone who's coming later to the market and trying to replicate that process. Is that a disadvantage because they have to learn for the first time versus, you know, someone who's already been there and is replicating experience. I think there's a value As well as obviously trust with partners. This is a huge part of requiring success for the fleet operations. There's lot of risk to trying a new partner if there's not to be capable of providing the actual service. And we take it very seriously, the trust that our partners put in us to make sure that we are delivering these sites at the dates we tell them that they will be ready. and that they'll continue to operate at the high uptime that they need for their operations. That's a lot of responsibility that we take extremely seriously. And so there's a question of if you're shown to be able to have replicable success, that's a lot of what we're providing is this ⁓ playbook of being able to scale into multiple cities. So I think there's a question of as this ecosystem shapes out, one, obviously, as you mentioned, having the right resources is scarce. ⁓ so being to move quickly and find the right locations. not a... Power can be created over time. It's just a question of how long that'll take. So, and finding the right zoning for it is also very, very challenging. I think there's definitely some first mover advantage for sure. And a flywheel of success that comes from repeated project creation. Daniel Abreu Marques: So oftentimes there's also a debate built versus buy in the AV industry. and maybe some fleet operators also are saying ⁓ is so mission critical. So I need to it myself. And ⁓ what would be your pitch to an AV executive who's skeptical about outsourcing something this important to you? Peter Cohen: Yeah. So I think we touched on a little bit. Part of that is, you know, there's fleet technology providers, there's fleet operators, there's TNCs, there's other people in the ecosystem. None of that has to do with charging infrastructure. So that's a strong deviation from what that company does day to day, their processes, their focus of their core team. This is all we do all day. So we go out and like I said, we're doing in many markets where we've had that repeated playbooks of success. We know how to work, especially with the cities, with the utilities to find the right locations that can be functional and have raised capital for exactly that purpose. Not, our capital is not going towards developing an AV tech stack. so I'd say that there's the question of, can we develop a team? If I'm the executive, can we develop a team that's going to do this at scale, ⁓ repeatedly, ⁓ can we rely on a third party that can act as a trusted partner to us and leverage their skillset? Which I believe that is how we are operating today. and so I said, also on top of that though, there was the pie of how many sites are needed is so large that ⁓ it's just there's so much opportunity out there for everyone. So even if there was a company that was doing some of this themselves, having a third party to augment or to be their core and find both is definitely going be needed. There's going be many, many of these sites in all of the cities. And so there's just such a large opportunity that, you know, I think there's many strategies that can be successful here. Daniel Abreu Marques: Makes sense. And maybe you also mentioned now several times that you raised over $1 billion, but this was in 2022. And it was before the current AV deployment wave really accelerated. you also mentioned that you think there's much more needed. Can you give us a number what you think would be a number that you need to raise to fulfill all the needs you have? Peter Cohen: Oh, I mean, there's going to be many, many rounds of fundraising, billions of dollars. The industry here is going to be tens, if not hundreds of billions of dollars of need in the next couple of years. As you mentioned with the rate of adoption accelerating, it's very exciting. And so that initial billion raise definitely is not going to get us to the ultimate need for this industry by any means. We're going to have to continue to funnel capital in and we're excited to do so. So I think what we've developed here is the platform that is successful and repeatable and excited to continue to leverage our strengths and build ⁓ cities and ⁓ powerful sites as time goes on. Daniel Abreu Marques: ⁓ to topic that is central to this podcast. So focus ⁓ a lot on markets ⁓ outside the US, particularly Europe, but China and Middle East. most of the AV infrastructure conversation is very US centric. So let us change that also today. as far as I know, and Terawatt exclusively in the US today. But is Europe on your roadmap? And if so, ⁓ What does the timeline look like? Peter Cohen: For sure. Europe's definitely on a roadmap. I think what you're seeing is a lot of activity in European that are large ride share markets, with announcements are popping up as quickly as they are in the US. So we want to be a good partner our existing customers, as well as ⁓ future customers who present in Europe, but not yet in the US. ⁓ So it's really to see all of the announcements that will translate to operations and then very quickly scaled operations. So ⁓ the time is I think, the biggest takeaway. I say it's never too early to think about charging infrastructure you ⁓ to be proactive as opposed to reactive and might be in a tight spot if all of a sudden you need five megawatts of charging in London, for example. That's going to happen overnight. So ⁓ we're excited to be looking at various cities to help the partners that are scaling operations there, which there's a lot of announcements, which is very exciting. it's never early to think about charging infrastructure if all of sudden you five megawatts of charging in London, for example. That's not going to happen overnight. Daniel Abreu Marques: Yeah. And from your perspective, what are the biggest operational differences for a company like Terawatt that you face in Europe compared now to the US? Peter Cohen: Yeah, mean, a whole new regulation environment as well as utility environment. That's obviously we're saying Europe, but obviously country by country, city by city that changes. So that'll be the biggest difference. Obviously in the United States, have every state is like a different country and there's thousands of utilities in the United States, which is a very fragmented market. So while Europe is much different in the United States, it's not that the United States is all one way of doing things either. being able to iterate and be dynamic in our platform. We're going to have to do the same in Europe as we are in the US. Just new parties to learn from. again, our team has in past lives worked across Europe. So we're excited to pull in that leverage in those relationships to be able to find the same success we have in the US as we do enter the European market. Daniel Abreu Marques: And if a European robot taxi operator calls you tomorrow and say we want Terrawatt in Frankfurt or London or Paris, what is the first thing you would need to figure out before saying yes? Peter Cohen: I'd say what does success look like? What is your operational need so we can understand we're looking for the right thing for them? Obviously we have some design standards and assumptions of what we would think to start looking for, but the first thing would be understanding where do they want to be? What does success look like from a stance of how many vehicles are the site sizing or different characteristics to make sure that as we do go approach the city and the utility to understand how we can find those sites, we're actually looking for the right thing for them, not just what we assume would be the right thing. But then the first step would be in those cities, you know, working with understanding zoning and power to inform where our search can go. The ride share market, understanding that obviously is more established, you know, figuring out from the customer where they'd want to be. But where a customer wants to be and where you can actually put these sites are not always, they're definitely not 100 % aligned ever. But we want to make sure we're finding the most overlap in those circles as possible. Daniel Abreu Marques: And maybe one question from someone who is not that into the real estate market, but I could imagine that if you think about charging hubs in London, for example, there is really not much space. So you probably need to have several different smaller hubs instead of one big hub. Is this correct? Peter Cohen: Potentially, I think it's a question of where you can find the right amount of size. Is that kind of sweet spot calculation of is it better to have a facility that's a little bit further out ⁓ than the of the man center, but maybe gets you more efficient operations if you have 50 charging stalls and you can run a of vehicles through them and do data inspection. What are the other things that happen at these facilities? ⁓ Or is it to have five. smaller sites throughout the city that are giving you more coverage. depends. London, obviously, like you mentioned, is very ⁓ city with a lot of old small streets and different things that are a little different than how things are running here in the US, obviously much older, which would maybe a different answer to that question. If you ask me for San Francisco, what's the strategy? something that as work with our partners over time will definitely ⁓ evolve. ⁓ There's not perfect might be you need both. It depends on how much coverage you can get, where you can find the power. I think a lot of this, there's no set in stone playbook for what site is perfect for what city. It depends on an evolving strategy, what you can find, which to the complexity and the need to make sure you're deeply in these markets for what opportunities can come up and making sure that it's the most efficient for the fleet. Daniel Abreu Marques: let's ⁓ zoom out our final part of our conversation. I want to get your read on the industry and where things are heading. So there is a scenario AV technology is ready. ⁓ You have regulation in The customer demand is also but the grid cannot keep up. And the charging hubs are not built yet. So how real do you see that risk? Peter Cohen: It's very real. mean, we're trying to make it less real every day by going out and proactively ⁓ to the speculative side of our business and making sure that we can eliminate those potential pain points for fleets. But that's real scenario for sure. Like I said, these things do take time and where industry has, I think, expanded more rapidly than anyone could have predicted and probably will continue to do so. If you look at charts that were drawn a couple of years ago and where the number of players, how quickly they're scaling, it's really exciting, I think. Obviously lot of that also driven by a lot of the AI push that's amplifying that growth rate. that's real. There was definitely some strategies you can take to try and ⁓ some temporary solutions, but they tend ⁓ to be less sustainable, ⁓ both as well as for the environment. But there's you can augment ⁓ getting a site on ⁓ with a Daniel Abreu Marques: Yeah. And everyone is also in the industry focused on the next, I don't know, 12 months and how many cars and so on are on the street. But I want to ask you about the next five years because you have a more long-term view, I would say. So what does Terawatt look like in 2031? If everything goes according to your plan. Peter Cohen: Yeah, I mean, we definitely think two, three, four, five years is a future we have to for our business. so, know, I think where it's right now, maybe the phase is getting the first site in many cities. It'll be getting the fifth, the 10th site in the existing larger markets and making sure that, you know, we have that network operating very across multiple facilities. And, you know, today we're providing to customers one or two sites that becomes a network and really deeply integrating with them to provide. What's the most efficient to kind of the London question? How do you route vehicles across multiple sites really effectively and get this ecosystem? Maybe there's shared sites as the industry evolves and there's more customer willingness to share the resource or maybe not. I think that's an interesting one that could either go, folks want to hold on to these strategic resources and not share because it's valuable to their business or there's so much demand in so many sites available that you can share across and there's then efficiencies there, I think time will tell. But I think the biggest thing that we see is many sites in many cities across the world as this technology really takes off. Daniel Abreu Marques: What could be one thing that could stop you from getting there? What is the biggest hurdle that you need to overcome for that? Peter Cohen: It's a good question. Obviously it's going to require a lot more money, but been really ⁓ not bottleneck so far. think just working with regulatory ⁓ to make that these sites are ⁓ in the locations that work with the cities ⁓ is a that's very delicate and takes a lot of coordination. wouldn't say that's not going to be like a bottleneck, but just more so something that call out that will require a lot of ⁓ proper coordination across, not just ourselves, obviously, it's all the players in this ecosystem to make sure that know when we approach working with these cities that AV is embraced and obviously we work with not just AV fleets, work with EV fleets as well that are manned, but just these fleet charging facilities are located in areas that work for the city and the citizens of the cities that is not causing any issues. Daniel Abreu Marques: Yeah. And you have been in this AV ecosystem also for a long time now. So is one thing that the industry is getting wrong right now that nobody's talking about? Peter Cohen: I mean, first I would say you kind of did the intro, just not enough focus on charging, but obviously very selfish in that view. But I think it's true. It's a required factor for the industry where a lot of the focus so far has been get to market, get the vehicles on the streets. And it has to be, but that'll quickly turn to, okay, now scale the operations. you can't just ⁓ snap fingers and change that overnight. You need to have a lot of thoughtful strategic planning. So I In the ecosystem, a lot of the press and news and focus tends to be at the top layer of the market where we see ourselves quietly building in the bottom. I think that's rapidly changing as evidenced by our conversation today and as fleets are really scaling up. So I know if they're getting it wrong necessarily, but I would say that the focus needs to be larger this being the factor on growth. Daniel Abreu Marques: Yeah. do you think then this is basically the reason why ⁓ enough people are talking about ⁓ because they have like a more view because they want to get the on the road first and then think about what comes later or what is the main reason for that? Peter Cohen: I think it's understandable. mean, you need the vehicles on the road to need the charging, right? So if you're only focused on charging and you don't have vehicles on the road, maybe that's not making much sense. But I think it needs to be equally part of the conversation as we're to get the vehicles on the road and scale. And you have to think about both at the same time, which is a challenging and a lot of work we do with the customers understand why they should be thinking about this, you can plan effectively to phase your operations up. But a nice value ⁓ of partnership is In the two years that we're building a site and you're growing operations, you haven't put up $10-plus million into a facility. You're focused on your core operations. back to how I would pitch that executive is put your capital where it's most effective in scaling your operations. Let us focus on the infrastructure. And when the has scaled to a point where ⁓ the is ready, if that process works really seamlessly, it can be a really effective offering. Daniel Abreu Marques: one last for today. So if there's one thing you want our listeners to take away from this conversation, what is it? Peter Cohen: It's definitely never too early to think about charging infrastructure for any fleet or someone in this ecosystem that is responsible for scaling fleet operations. It takes some time. There are folks like ourselves who are here to help make that a lot easier and more seamless of an operation, but definitely being more proactive will lead to success versus being reactive could lead to potential pain points and slowed growth. Daniel Abreu Marques: Peter, thank you for being so generous with your time and your insights today. So I think this conversation made it very clear that the autonomous vehicle revolution is not just about sensors or software, it's also about megawatts and infrastructure and real estate. if this episode made you think differently about the infrastructure side of autonomy, check out Terawatt's infrastructure side. And I will link the website in our show notes as well as Peter's LinkedIn account so that you can connect with him as well. So Peter, thanks again for your time. was a very insightful conversation and I enjoyed talking to you again. Peter Cohen: Thank you, Daniel.