The Exchanges

Q algorithm-driven differentiation. We, we, you know, used our fancy ML box to do something with that data nobody else can do. You know, it's like so many different configurations. What, what's the one that you think, first of all, like, what's, how do you tell the story around Arcadia's moat? But I guess more broadly, like, what's your view on, Building a, building a long-term moat in a data enterprise.

A Yeah, so when you look at examples from other industries, uh, and I, you know, the, the most similar and analogous to us is Plaid in the fintech space. Um, they're not inventing any new data, right? They're simply exposing data in a much simpler, easier way, uh, for, uh, In that example, fintech companies to build new innovative customer solutions. So that, that's sort of, I mean, from, you know, being honest, like we, we're exposing data that is already there within the multi-billion dollar investments into online digital accounts and metering. Um, but making it simple and easy for the new energy company to focus on what they're doing, pull up a single API, um, And get access to multiple, you know, geographic monopoly utilities, right, without knocking on every, every utility's door at once. And that is just wildly valuable. The moat, um, you know, a lot of people, um, they, they assume outside looking in that, that this Simple task. But in reality, the way a kilowatt hour is presented by Duke Energy versus, uh, Mid-American versus Pacific, it's radically different. Um, and then when you talk abroad, it's a, it's a tough data model to standardize. Um, but I think that network effect, part of the reason we are so excited about bringing in Urgenet, who's spent, you know, more than a decade, uh, You know, building connectors across providers, you know, 9000 plus utilities around …

AI assessment note: “I think our moat is, if I'm a new energy company... I just want a single API.”

Q not the ones who know what to do with it exactly. And so somebody else like Arcadia or an Arcadia customer Um, might be able to, to do something more valuable with it. There, there's been this big movement for years around like green button, basically letting people open up their accounts to third parties. What's been the challenge historically in, in third party access to utility data at scale?

A So there's been a lot of attempts and you probably have seen, uh, you mentioned green button as maybe the most high profile, uh, attempt at exposing, uh, Uh, data. There's a lot missing. And this goes back to, like, um, you know, I think for folks who've been in and around the industry, let's say ESCOs or new energy companies for, like, the last decade or two, they always run up against this problem, uh, that is, um, asking for a PDF printout of the bill, right? Or, or just some, some really sort of manual guesstimates, uh, On, uh, you know, how someone's using energy, uh, because it's so hard to sort of extract the data. Um, there's a lot still missing today. So like, there have been a lot of attempts, um, EDI, uh, which is electronic data interchange, used by a lot of retail energy companies, has a very specific use case. It's not very real time in a lot of sense. It's, it's not applicable. It doesn't have sort of full, uh, account, uh, visibility, uh, that a lot of sort of Businesses need. Um, you know, Green Button, uh, and some of, some of the other attempts, at the end of the day, they all suffer from poor customer experience, limited data availability, um, Fairly weak service quality, and sometimes all three, uh, you know, wrapped and packaged, uh, and, and utilities were sort of promised to do green button, um, uh, but haven't actually executed on it. And the reality is…

AI assessment note: “they all suffer from poor customer experience, limited data availability”

Q What's the point of it? What can you do with it if you have it and put it in the right place in the right situation? And then like, what don't we have actually, what would be more valuable? So let's start with what, what is the data? So talk me through like what, as you think about building a business around becoming an energy data platform, what data matters?

A Yeah. So, uh, I'll talk a little bit about, uh, the different layers of data, um, Um, and a bit of, you know, what matters, especially to people who are probably listening to this, um, focused in DERs and renewables. Um, you know, the, people talk about, uh, the, the different layers of the grid, right? The distribution system, uh, generation, transmission system. And each one sort of within has just a complex set of data, right? Largest, the grid, uh, as the largest sort of engineering marvel in the world, uh, already has billions of data points. Um, and the fact is, we have spent, uh, at the distribution level, billions of dollars on upgrading the telemetry around the systems that are in our homes and in our businesses, which is the meters. Um, specifically around where we focus, um, Around the distribution system is the data that you and I produce when we use energy, right? So in our homes or, uh, our buildings, uh, you know, small businesses and commercial buildings, um, there is a lot of rich energy data that is produced by the meter, but it's also sitting in your online utility account, right? So like the incredible, uh, you know, Network we've already built, uh, already has amazing telemetry, and it's actually exposed to you in an online account. Most people don't spend a lot of time there, right? Um, uh, I don't know how many of your listeners, uh, actually log in month…

AI assessment note: “historical energy usage, uh, depending on the utility and the AMI meters, real-time energy usage”

Q What data don't We have access to that would be, make it even more valuable. Like, what are we missing? Obviously, there is a lot of data unlocked by just access to the customer's utility bill, but not, not everything you would want, I assume.

A Yeah, we, we've still got a long way to go. So with the combination of Virginette, we're able to pull, again, you know, 52 countries, like 9000 utilities around the world. We're able to pull whatever the utility is. Provides. Which is still pretty radically different, obviously, of utility to utility, and depending on the AMI structures that exist, and how quickly they, they publish that data, because that is what we're relying on. And so, um, look, time of use rates, also not everywhere, right? And then we're seeing a huge growth of that, I think, you know, utilities and regulators understand why, and so that's growing. Um, I think faster data access of, even in the places where AMI exists, um, Uh, is absolutely necessary. Um, we talked about carbon intensity, but I think what, what will be difficult here, similar in some respects to, like, the offset and rec market, is, like, there needs to be a single standard, and I actually see this as where the utilities role is, of, like, what is the single standard on a utilities distribution grid of the carbon intensity in an intraday period? Like, I think they should own that. Someone should put a stamp of approval on it, that this is what it is, not Six different companies saying it's six different things. Um, And so, all those things matter. I think, you know, we also look at other things around, like, the, the payment of a utility …

AI assessment note: “faster data access of, even in the places where AMI exists”

Q So there's all this data. You've got my historical consumption data. Maybe you have interval data. You have a bunch of metadata about me. Like you said, you know, have I paid my bill on time? Have I moved, et cetera? Like, how do you think about some of the sort of scalable use cases for that data that allow you to build a big business on top of that?

A Totally. And, um, you know, something I say around the office is like, uh, I was sort of red-pilled, uh, eight years ago, uh, as sort of a reference to, you know, once you, once you see something, you can't unsee it, uh, from the matrix, and, and that was sort of, if you're going to be in this business of new energy, you know, selling energy services, DERs, The data I just described, energy usage history, um, underwriting, cost structures, that it's completely foundational to DERs as a valuable grid asset. And I think a lot of companies, as including me in a past life, like, maybe sold energy services or even, or even DERs as widgets, right? But everyone loves to talk about a big game of like, what can these, uh, you know, you know, responsive devices or services, uh, provide over the longterm. And in reality, you need, you need data on how a customer is using energy, how it's costed, how it's priced, uh, and then maybe be able to finance it using this underwriting. So some, some like very, uh, simple examples of people using our platform. Uh, today, um, you know, our biggest vertical to date has been community solar. Um, for those that don't know, the short story of community solar is sort of a developer building an asset and being able to chop it up into thousands of offtakers. Now, without insight into the fact that, you know, Shale, you live in a mansion with a jacuzzi and …

AI assessment note: “our biggest vertical to date has been community solar”

Q platform is doing is allowing the OEM who delivered me my EV to Help me optimize my charging. Presumably you could also be offering the OEM who delivered me my thermostat the same thing. Do those two things need to talk to each other, do you think? Or is it okay that we might just end up with a network of different devices each getting optimized against the same tariff?

A Yeah, I mean, that is a fundamental question you're asking about where the market goes, and I think we at, at Arcadia, uh, decided that we just wanted to be, uh, the, the picks and shovels while folks sort of figured this out. Like, I'm, I'm not sure that there, there will be an opportunity in the future for a single manager, uh, and someone should take that, uh, path, but Again, it's not possible unless you, uh, know the rate structures and every device is sort of optimized against it. Um, I think there's a future where there will be sort of whole home managers that pop up that are, are able to sort of, you know, you asked earlier about the different layers of data. I actually skipped one that is sort of device level data, what's happening at the distribution system, and then on up. Um, and...

AI assessment note: “I think there's a future where there will be sort of whole home managers”

Q of things, for example, you can imagine like weather data being really important or, or maybe, you know, some grid data, data, right? There's been some, a bunch of attempts to Sort of internalized distribution system data to try to do things. Well, what do you think of that is interesting, at least whether or not it's currently integrated into the Arcadia platform? That lies outside the customer's utility bill.

A Yeah, and I'll say from our perspective, um, you know, we're, we're focused on data where we have a competitive advantage. You know, one of the reasons we acquired Urgenet is this account level data, you know, that to, to be able to pull it in globally across residential, commercial, serve any customer, and get rates and tariffs. That's where we wanted to focus initially. Um, so you mentioned two, um, Uh, weather data, of course, incredibly useful for all sorts of applications. Um, uh, and I think similar to carbon intensity data, which I would also say is, like, an incredibly valuable, um, you know, source to pull in, data source, it, it's somewhat subjective. It's actually a pretty tough physics problem, uh, to figure out what the right, um, uh, as you know, there's, there's a few different ones out there, eGrid, WattTime, a few other private companies, uh, Um, it's, it's hard to sort of know what the right benchmark is without, um, frankly, I think a utility, uh, stepping in and sort of saying this is, this is what's happening on our, our system.

AI assessment note: “weather data, of course, incredibly useful... similar to carbon intensity data”

Q for like what the generation profile is going to look like. In the future, you, you know, assume some benchmark average for savings from whatever device it is, and you're probably off by some relatively significant fraction, but, you know, if it's enough to kind of make the decision no matter what, does it, does it matter? Like, what level of fidelity do we really need in making these decisions?

A Let me give you an example that we're seeing and then pull out a bit. So some power uses our data, um, Around their solar proposals. When you look at any rooftop solar company's, uh, funnel, there's a huge drop-off when the salesperson, you know, they come to your door, they say, we'll get you cheaper, uh, solar, put it on your roof. Then they say, can you print out 12 of your past PDF, uh, utility bills? And there's this huge drop-off. Um, and oftentimes, you know, you're taking swags on, like, how much can I save this customer? How should I size their system? So, The world we're imagining is if you had this level of granular data, could the product actually deliver significantly more savings, um, and significant long-term savings and, and live up to promises to the customer? That then sort of feed back into better sales, faster sales velocity. Because I think the problem a lot of DER companies have today is sort of unmet, uh, promises, right? On long-term, you know, what are long-term sort of assets going into and long-term solutions going into buildings and homes? Um, you know, you, you've probably heard of some of these stories of like, you know, uh, Uh, you know, people sort of miss price, uh, misallocating savings on rooftop solar as an example, but, and so I think this is actually crucial, Shale, is like, if we're gonna, like, increase the sales velocity and more and mor…

AI assessment note: “we kind of need that granularity, uh, to deliver the savings and, like, live up”