Energy Tech Startups

0:00 Transform your startup journey with the energy tech nexus. Connect with fellow founders, access critical resources, and be part of a community shaping the future of energy and carbon tech. Your

0:10 path to building a Thunderlizard starts here. Learn more at energytechnexascom. Welcome back to the

0:16 show. Today, I'm really excited to have Sean Kelly, who's the CEO of Amperon. Amperon provides AI-powered electricity forecasting and analytics solution for every electron on the grid. Welcome to

0:28 the show, Sean Thanks so much for having me. So excited to have you here. Maybe we could start off with, tell us a little bit about Amperon and what problem are you really solving for the energy

0:40 transition? Yeah,

0:43 great question. The problem that we're solving is that what's coming down the pipe, the future forecasting is very difficult. If you look at our website back in July of 2018, there's a paper or a

0:55 medium article that my co-founder Abe Stanway wrote. and literally laying out why we thought Amperon was the company to start. And thankfully for us, we've had no pivots along the way because we

1:07 knew EVs were coming. We did not know that we'd get the free tax credit and people would be very incentivized to do that. We knew PV was coming. Probably didn't think that 20 gigs of utility scale

1:20 solar in Texas would be on the grid. At that point, we knew smart meters were continuing to come. They started back in around 2007, '08. With the Obama stimulus. And so we knew we were getting

1:33 better data. It's funny that us and energy call big data one meter, or one meter read every 15 minutes, whereas the tech people that we have are like, what is our iPhone hearing from this

1:44 conversation? One data point every 15 minutes? I don't think so. We just knew that it was gonna get significantly harder. And so that's how we started with demand forecasting. And then we

1:54 realized that the story now is not demand forecasting It's net demand forecasting. is we're sitting here in Houston, Texas. If it's a really windy day, prices are cheap. It really doesn't matter

2:05 what the load is gonna be. If there's 25 gigawatts of wind and another like 15, 18 gigs of solar, it just doesn't matter what the load is. So yeah, that's where we've, we knew it was coming. We

2:18 didn't also know that everyone's meter was gonna change with COVID and everyone was gonna start working from home. So our thesis could have been more spot on and there were things in there that were

2:28 correct that we didn't even know would be that correct. Yeah. So tell us about how I guess complicated this problem is. Are you going down to the like generating asset level for? Well, you said

2:41 demand. So it's not supply and necessarily. Yeah. Yeah. On the generating asset level, yes. So we're doing independent solar utility scale solar assets as well as we'll be launching later this

2:53 year, wind. And we also do on an ISO level, the wind and solar forecast us versus ERCOT versus KISO versus PJM. And so we are looking at it on an asset level, but then on the meter side, we'll

3:08 look at it as the kind of all of the ISO, so we're in all seven ISOs in the United States and both balancing markets, so Western imbalance as well as the southeast and 16 countries in Europe because

3:24 a meter's a meter, a megawatt's a megawatt, it doesn't matter what language it's in. So we're doing all the way down, we've got roughly 10 million meters under management. Yeah, so how does that

3:34 work? It's not like public data, I take it. Correct, it's not public data, we're going in and getting the data either through a skater read from our client who they'll upload to us or we'll get

3:47 it from EDI data, which is old archaic data that utility companies get I think it started in like the '60s or '70s. or we'll get it through the ISO directly. So we'll go in and look on behalf and

4:00 we'll be like, oh, Jason, energy, we'll go in and get a like load serving entity certificate and get the data that way. So there's a few different ways to get it, but at the end of the day,

4:10 it's still really dirty and it's

4:13 continuing to get better, but there's gaps in it, there's zeros in it, there's negative prints in it. And so a data cleaning becomes a lot of the story too, which is nice that on the engineering

4:24 side, we're hiring from the Metas, Amazons, Google, Microsofts of the world, 'cause they know data. Yeah, interesting. So who then is your primarily your client then? Yeah, so we've done a

4:40 good job of diversifying our client segments. And so we started off here in Texas with reps, retail energy providers. So for those of you who don't know, there's 14 states in the United States

4:50 where you get to choose your own electricity, Texas is unique in that. As we know, if you don't choose one, your lights are off. There's no power. Whereas when I lived in New York, you could

4:60 choose one, but you just automatically had Con Edison as a backstop. So that's where we started. The goal was to get meter data as quickly as possible and dealing directly with the retailers is

5:11 significantly faster than utilities. 'Cause they have to measure it anyway. So they want to bill. They got exactly. That's exactly why they have to have it. 'Cause they got to know what you used

5:19 and now you have all these cool different plans So I'm on, I'm personally on Tesla's plan. And so

5:27 it's great because my car just automatically, as long as it's plugged in, starts charging at 10 pm. And so they need to know not to build me for that. And so you've got to have these actual like

5:38 intervals and they've got to know to turn the car on at 10 o'clock. So that's where we started. Great place to start. It was definitely hard to watch like Winter Storm Erie and live through that

5:49 because those really, that really hurt the retail industry. Next up for that is traders. That's my people. That's my background. That's what I did for 11 years. And so as we were talking about,

6:01 before we came on, that's when having a D-Shaw as an investor is really helpful, just 'cause it kind of gives you that street cred of, Hey, this is a big powerful fund, is not only using but

6:13 investing in Amperon. And so traders came second. We went from there to public power. So munis and co-op, kind of the forgotten ones There's 3, 300 municipalities and cooperatives. I lump CCAs,

6:26 community choice aggregators, primarily in California into this. And so we've done a really good job of working with some of the aggregators. So like AMP, American municipal power has 134 munis

6:40 under it. And so we're able to kind of go work that way. And then the next we went to is getting back to your earlier question is IPP, so the independent power producers This is anyone who produces

6:52 power, so. We lump batteries into here. Obviously you've got wind, solar, but even natural gas. Anyone who needs to know kind of how best to dispatch their plant. So fifth, we went to CNI. So

7:06 commercial and industrial clients. And so this is anyone who has some megawatts of load. And so we worked with someone very forward thinking, this is the space I'm really excited about as we start

7:19 having regulations coming down of like scope two and things of that nature, which we've gotten rough guardrails around a 20, 28

7:28 time frame. I think that people are going to really take off because in the CNI space, it's like LNG. That's why everyone's in Houston for gas tech right now. You look at all these data centers.

7:39 That falls into this data center is the hottest topic right now. And so just kind of continuing on that path. And then last and definitely not least, but utilities And so utilities, we took a

7:51 unique approach. in that we knew that it was a slower moving process. And we also knew that if you're early stage, they should not trust a like year old startup that has no track record. And when

8:06 I first started pitching them, it raised860, 000, not over30 million. And they should not, for a critical infrastructure, trust an early stage startup. So we didn't actually go sell the

8:18 utilities until after year five. And we finally like, we had money, we had a proven track record, and then you get to go in and say, Hey, this is an amperon-shaped peg, and there's the

8:27 amperon-shaped hole, as opposed to kind of going in. And the innovation competitions are great, and it works for some companies. But for us, we just realized that doing these kind of like long

8:39 pilot cycles can be hard, but now utilities are kind of flying off the shelf. And also that's in part because we have Microsoft as a channel partner they are going in and like introducing us to

8:53 utilities and we're introducing them to utilities as well and just a really strong partnership set up by our chief revenue officer, Alex Robart, who had been who had been multiple years in like the

9:06 executive team of both energy and sustainability, Microsoft. Yeah. And I so that what you're providing them is is forecasting. Yeah. But I assume each of these constituents use that data in

9:18 different ways, especially the traders make money differently than the RAP. So how do you do you do you have to change the tools you give them or are you giving them a fire hose of data? How does

9:30 it work? Yeah, definitely a variety of variety of different ways. And you're very right is that some people traders are trying to make money where retailers are trying to mitigate risk, right?

9:44 And so their job is not to make money on the trade. Their job is to basically say, okay, I got your home. And now I need to go lock in that price because I agreed to set price for you for two

9:55 years. You need to go and hedge that. And obviously the electricity market is like the stock market, it moves around. You can't just say, Apple stock is going to be worth the same forever. And

10:05 so people ingest data definitely different ways. The short-term forecast runs 360 hours, and that's because that's how long the weather models run. And so the weather models, you've got the Euro

10:20 or ECMWF, you've got the GFS, you've got those all, they're government models that continue to run like four times a day. The shorter-term ones run every hour, every three hours. And so that's

10:34 why our predictions go out that 360 hours. And then we have a longer-term model that runs out 10 years that we run normally once per week, but we'll still give the client hourly granularity. And

10:46 back to your point, people want data, how they want data. And so we've done everything from starting off with what I think is best in class UI UX for them to just kind of visually see what is going

10:60 on. Then we also have API that people can look at. We also have partnerships with like Snowflake so that they can go in and ingest data that way. Really love what Snowflake's doing. They've been a

11:11 great partner for us. And then we'll also some people want to just be a power BI. And so you kind of got to having over 100 customers, you got to cater to what they need because at the end of the

11:23 day, the data is the data. And then summer, never, some of the traders are really hilarious. They ask for the log in every six months. I'm like, Oh, you need to reset the API? Yeah, pretty

11:33 much. And then other people, we can sit and they're like logging in every single day. I guess my question was around, you know, for someone who doesn't understand this market so well. You know,

11:43 retailers were your first customer target. Tell us a little bit about. Why is this forecasting so important for us? And how does it actually work? Yeah. Yeah, 'cause I just like, I just slipped

11:58 the light on. How are you gonna fix my future actions? I don't even know what I'm doing. That's why it's nice to aggregate a bunch of people. We've got, I think around two and a half million

12:07 deregulated meters in Texas. And so you can kind of say, okay, this offsets the other person. But the reason it's important is you need to know what's coming the next day. So every morning said

12:22 retailer is bidding into a day ahead market. They have to put in their bids by 10 am. Most of them probably do it around seven or eight just to make sure there's plenty of time. And then they go

12:31 say, dear Urquat, this is how many megawatts I would like to buy out of need one. So like midnight to one am. Two, three, all the way to out earning 24. And then if they get picked up and they

12:43 go in and get a price, they now know they get to rest easy If they got their number correct, the price is already set. If it's not, and they need to buy more, then guess what? The real-time

12:54 price, as we all learned during winter storm theory, can go a little wild. And so that's why people need to know, it's really going back to the, not making money, but mitigating risk, because

13:05 they now know I'm locked in at this price, but they have to get the number right. If they bought 100 megawatts and it comes in at 120 megawatts, they're gonna have to go buy that 20 megawatts from

13:14 the market. They're gonna be buying that 20 megawatts from the market Everybody else is probably at the same time, and because it came in three degrees hotter, which we know three degrees in summer

13:26 with, I think, a thousand percent humidity, it feels like it's definitely gonna add a good bit to the price. And so that's where it is. It's just really about that price surety. So this is, I

13:39 guess, where the question was going, is on the one hand, they got to pre-buy the power in a day ahead. And if they don't buy enough, they have to buy on the spot market I assume if they buy too

13:49 much. they resell it or they just lose it. It's the same way, if they buy too much, then it might say you bought too much, or maybe everybody else bought too much as well. It's really windy

13:60 today. We didn't expect it to be so windy. As we know, wind is essentially free after you have your cost of capital with tax credits. At one time,

14:12 it was like negative 30 something dollars was in the money just because of all the different tax incentives. And so wind is free. And so if it's really windy, power is going to be really cheap. So

14:21 that's where you don't want to buy too much either. But you, as they say, long and wrong and short and fired. So - And so how did people do this before? Because I assume it's more than data. I

14:34 assume it's more than the best AI model.

14:38 Were people just guessing the bids before and the demand? Yeah, they had a really hard time because you think about it back in the day

14:48 kind of meter made shown up and your dog chased them and they were pulling one meter read every like month or two. And then, I mean, you hear about places in New York where obviously everything's

14:59 very densely populated like the door will be locked to the basement where the meter reader is and they go read it like once a year. So you've heard those horror stories and then some crazy bill comes

15:08 out. So the data just wasn't that good. So yeah, it was pretty kind of blind guess because you're dealing with what they call summary data which is like a data point every month Not that helpful,

15:19 we're doing something every, I mean, every second, but definitely every 15 minutes. And so it was definitely a lot of guessing our good friends Microsoft Excel is how they kind of have put that

15:31 out, but this is something that the AI use case is extremely helpful because as I mentioned earlier in this, this is becoming a harder, harder problem. And so with a harder problem, you need

15:43 better solutions. And so from our standpoint, we're not just saying, old-fashioned linear regression model is going to do it. We're looking at the latest and greatest deep learning all the above

15:54 and then ensembling different models together. And so saying these six models and they automatically weight itself, this is where it's nice having people much smarter than I controlling all of this.

16:06 And so our team of PhD data scientists go in and will merge the model together. And so 22 of this model and 18 of this model, etc. So any, any output that you're seeing is going to be a derivative

16:21 of four to six different models. And then we also have depending on the region, two to four different weather vendors. So this is why someone would hire you instead of trying to do it themselves.

16:33 Exactly. And this is the nice thing. I mean, A, you get to raise venture and venture capital, which I'm very passionate about. And I mean, you know all about from your career is that you got to

16:46 the best and brightest stuff that's going to come out, you've got to kind of say, hey, this is a good idea. I'm willing to put a million bucks into said company that may or may not work out.

16:54 Historically, it's not going to work out. And so we've raised30 million, have a hundred employees and have been at this for six and a half, I guess January will be seven years. So we have that

17:04 and then we get to scale that out across over a hundred companies. And so if you have a problem, it's a problem we've already seen because again, we have those six customer segments with really

17:15 good use cases and all in really good clientele in all those different customer segments. So you get to say, hey, like big retailer, we've already seen this so we can come and do this. So it's

17:28 really that first build kind of covers it and then you get to go in and from like economies of scale essentially. Do you find the models are kind of customer specific or is there lessons learned?

17:43 Not just in terms of like how you do it, but in terms of the data you collect, in terms of how it's able to create better predictions for others. So is there a network effect of having more kind of

17:54 customers on the system? Definitely a network effect of having more customers on the system combined with how long we've been in the market. And so you look at Europe, which we just launched, the

18:03 press release came out June 11th of this year Texas, we launched the kind of MVP, like minimum viable product in November of 2018. So we've learned a lot more lessons in Texas than we have in

18:17 Europe today. But however, the models are very transferable. That's how one of the jokes we have internally is it took us a year and a half to stand up PJM. So PJM is basically Chicago all across

18:28 the Mid-Atlantic region originally stood for Pennsylvania, Jersey, Maryland, but now is like a dozen or so states. and it took us a year and a half to stand at PJM, and it took us six months to

18:42 stand up 16 European countries. So we've gotten a lot better. We have a lot more people, a lot more funding, and have done it before, but this is where going back, if we make a change to the

18:48 model, it's across really all of the models, but then we'll have to hone in and be uncertain specific regions, 'cause California looks much different than New York.

18:59 So the way you're describing the models, and I guess the way your AI algorithms work, it seems very complex If someone asked you, why are you forecasting really high prices for tomorrow, can you

19:12 answer that question? Me personally? No, I'm just wondering, if someone asks you why, 'cause it's just the model that are predicting it, and there's so much going into these different models.

19:21 Our scientists know why things are going in, because we're not always right, we're not infallible. We're right most of the time, most places, but at the end of the day, can still be wrong. And

19:32 so we'll go back in and kind of figure out the post mortem and we'll if something is, say, next week looks really high, then we'll go in and kind of dig in and make sure to understand why it's

19:44 working that way. But it does a little bit turn into a little bit of a black box. So no, we definitely stay on top of it and understand the inputs and make sure that the different weather inputs

19:56 are going correctly, making sure the ISO didn't give a negative print or give bad data. So yeah, there's a lot going on. I mean, that's how you get 100 people and half those are engineers, like

20:07 engineers and scientists. And then it's, I guess it's also fine tuning the models again, right? Like if it is showing you something that's kind of anomaly, then you would go in there and tweak it.

20:19 Yeah. And you said you guys are working on deep learning. So that's part of where you are automating the tweaking almost, right? Exactly. So our models retrain every single hour because they run

20:29 every single hour and they've been retraining since. November of 18, and this is one of the things we actually have, it's a little bit of a late mover advantage. Our tech stack started in Google

20:41 Cloud. We've since moved over to Azure, but at the end of the day, this

20:47 isn't an Excel model that had a little bit of VBA on top of it, and then became this, and became this, and so this is something that, if a company's 20 years old, which there are multiple 20 year

20:58 old companies in this space, there was no cloud, right? And so this is where being a younger company and having that tech stack, but now being old enough and well capitalized enough, definitely

21:10 gives us a big advantage over someone who's had to do different iterations and say, Hey, we're gonna move into the cloud, and we're like, We started their day one. So - The cloud native. Yeah,

21:19 we are definitely cloud native. And that's where, I mean, AI is the hot button right now, and I will say one of the coolest things that I guess accolades that we've had one of my investors texted

21:32 me one morning and I was making oatmeal for my three-year-old. And he goes, Hey, congrats on the list. I was like, Okay, like what list? And he goes, Ah, Andreessen Horowitz put outtheir

21:45 American dynamism, top 50 AI companies in the US. There were 12 energy companies and one Texas companyand we were the Texas energy company. That was kind of the first thing we have no moneyfrom

21:56 them at all. And that was kind of the first, Whoa, okay If we end up on energizes roadmap, I'm like, I know guys, you

22:06 let our series be. And that wasn't like your goal, right? To come on the top list of AI companies. That's not how you started. No, ours was to do what we thought was best in class. And that's

22:16 what we've done. And it was just really amazing to get noted for that. Because I do feel that too many people are, everyone always chases the, shiny object. And right now everyone is just chasing

22:29 AI. You see a bunch of people changing their domain to dot AI. They're like, come on. You just like, you just saw that in blockchain before that. Like it's always something, right? There's

22:39 always a new shiny object that's like best in class. You're like, build the company that you want to build for the reason that you want to build it. If it takes care of the latest and greatest by

22:49 all means, but that was definitely something that that kind of fell into our laps. And we're like, Oh, they noticed that we've been doing AI the entire time. And this is what it's like to hire

22:59 kind of best-in-class engineers and data scientists. We've lost one data scientist in the history of the company. And that was to Google DeepMind. And I was like, Godspeed, that's who you're

23:09 competing against. That's who I want to lose to. I was like, best of luck. So his name, his name was also Sean. I was like, I guess it's just too big. It's the name. Too many Sean, too many

23:17 Sean. So he says something I'm curious about. So you switch from from Google Cloud to Azure, I assume they didn't show up or Azure Cloud credits, and that's why you switched over. Was there more

23:29 to it? There was definitely more to it. I mean, Microsoft is definitely owning the narrative around AI. And then the other thing too, is just having that opportunity to deal with them from a

23:42 partnership standpoint. And so because Alex had been at Microsoft for a couple of years and had a great relationship with that team, Microsoft has a very high percentage of US utilities and an even

23:55 higher percentage of European utilities. And so it really helped from a go-to-market standpoint. That was really the reason. And obviously they're in Cineviz 'cause more stuff's gonna run on Azure.

24:07 And we moved over and it's been a great partnership that's just getting launched. I think last month is when we put out that we are transactable in the Azure marketplace. So it was definitely some

24:19 commercial implications, but also on top of that, I mean, Microsoft is definitely owning the AI story with OpenAI and that investment. And they are been a great partner for us. But you said

24:31 something transactable on the marketplace. What does that mean? So different utilities have what's called a Mac spend, Microsoft Azure commitment to compute. And so you'll go in and say, I'm said

24:44 big utility. I agree to three year X million dollars of spend And so if you

24:55 work with Amperon because it's in Azure, then you can say, oh, we have a three year one million dollar contract. So that comes out of your Mac spend. So it's a really great, there's a bunch of

25:07 cool companies that have the same ability. And so we also get to kind of partner with those companies and work together with utilities just because again, any budgeting is one of the hardest. about

25:21 selling to utilities and so this comes out of normally like a CIO CTO budget and is already spoken for. So yeah, it's been a really good go-to-market motion and obviously Microsoft is bigger and has

25:32 more people than we ever will feel confident in that one. Well, and I'm glad that you went down that route because I think one of the things that was not clear to me and maybe this is a more of a

25:40 local trend is if you're planning to serve in the energy industry, there is a advantage to starting on Azure as opposed to some other maybe a newfangled technology because the kind of market trust

25:52 exists for the customer. I assume if they hear that you're on Azure

25:57 and they're going to give you their data, it's like, okay, check the box, it's SOC2 compliant. I don't even know what that means. But all these things that like an IT department would care

26:08 about, is that part of what you mean with the go to market? Absolutely. It definitely has that higher level of trust. They're already engaged there. They're already like have sellers there who

26:17 know, um, there was a large utility that we had had conversations with and Microsoft goes, would you like to speak with them? We've spoke with them like a handful of times. Would you like to

26:31 speak to the CIO and the CTO? Yes, please.

26:36 So it's just a different level of conversation just because I mean, it is one of the biggest companies in the world, right? And so it definitely gives you that. And going back to SOC 2, did not

26:46 know what that was originally when we started the company We had a large company that owns generation in town and they're like, you have to be,

26:56 have SOC 2 compliance. So I go to Abe, my co-founder, our CTO. And I was like, hey, we have to do this. He's like, how big is the contract? I go,

27:06 24, 000. He's like, no. I was like, this is one of the best logos in the space. We have to do it. And so we did it I think at the time, we had to. eight people at the company and now we have

27:20 a hundred and so doing it with very little tech debt was one of the best decisions that we made and since then now you get to walk into and basically you just say here's a copy of our last sock to

27:31 report we just did I think we just did our audit everything went great so we were on year I don't know four or five of sock to now but doing that with very little tech debt was great and I think Abe's

27:42 not mad at me anymore to

27:46 do that tell us about your startup journey I mean you started six years ago a hundred people today being able to raise 30 million series B not many companies reach the stage so tell us a little bit

27:60 about you I'm certainly not in Houston yeah not tech companies in Houston yeah so tell us about that journey the ups and downs and what do you think you were able to do differently to get to where you

28:12 are yeah

28:15 definitely something that I think about a lot so I grew up here in Sugar Land, like big Astros fan, Rockets fan, Texans are fun. Went to AM, like I'm Texas boy through and through, but I was

28:28 living in New York, moved there in like 2016, 2017. And in spring of 2017, I went to New York Energy Week and I actually volunteered in New York Energy Week and A volunteered at New York Energy

28:42 Week as well. And so it was really cool to see, I know y'all did a ton of programming with Houston Climate Week And so it's just really exciting to see that here 'cause that's how I met my

28:50 co-founder back in spring of '17. And so my thesis was I wanted to know what New York knew about energy, not a lot, but I also didn't see a vibrant tech scene in Houston. And up there, especially

29:05 in 2017, it was co-founders everywhere, data engineers, data scientists. I mean, just this level of talent I've never seen

29:15 And so, Abe's joke is. I'll build it if you can sell it. And so I was like, all right, this like young, like New York hacker, basically like sold a hacker company to Intel coming out of college

29:28 was like pre IPO at Etsy, early a planet, like stood up a data science team at McKinsey and I was just like, whoa, got Forbes 30 under 30 in energy. I was just like, this is great. And so just

29:42 us two, the odd couple, but kind of came in and said, hey, let's build this. And then I've thankfully been networking since freshman year at AM, which was a couple of years ago. And so I have a

29:55 basically 25 year old network and I was able to go around and say like, hey, what's interesting? What's helpful? How do we do this? I think so many people walk in and say, hey, you have this

30:05 problem and this is your answer. That almost never works in going in and being like, so I'm smarter than you and have more resources so you should do things this way That very, very. rarely works,

30:17 especially with the stubborn cowboys in Texas. And so having the Silicon Valley types roll up and telling someone to insert large inner like oil and gas major here that they don't know what they're

30:30 doing, doesn't go well. I worked at a company that got bought by Lehman Brothers and the culture clash between the two is pretty hilarious. And so that's where I think that we could actually kind

30:42 of figure this out. And so that's what we did. And the other thing I think, we're one of the other things we did that was really wise, is we just stayed very focused on the customer and just, I

30:53 mean, I ran sales for the first five years, basically of the company. And then really kind of, I guess we've had like proper sellers for like two, year and a half maybe, and then two years now.

31:05 And so was able to just do it ourselves. So we raised 276 million over the course of the first four years. And then in the last two and a half years,

31:15 additional. And so that's where I think just getting that really, really strong product market fit, knowing that your customers just cannot live without you. Are you ready to lead the

31:26 decarbonization charge? Energy Technexus is your platform for growth, offering unique resources and expertise for energy and carbon tech founders. Join us at energytechnexuscom

31:37 and start building your Thunderlisset. So how did you know when you had product market fit? When there's a really cool software called mixed panel and it basically lets you go in and see how many

31:52 times people have logged into your site. And when we were realizing that people were in our app all day every day, that was kind of one of the first things. When something would be wrong, they

32:02 would catch it in like seconds or a minute and then ping us. And so you're like, we are like, we are one of your core functions. Like this is really important and a really big deal. And so I

32:12 think that's when we knew that we had product market fit. But then we also, you can never stop getting product market fit. You've always got to know there's someone on your tail. You've got to

32:23 make sure you're the most accurate. The grid is changing. The whole piece is behind the company. It's getting harder. There's other people raising money, just all of the above. Different product

32:34 sets, as I mentioned earlier, demand's not the story, it's net demand. And so just being able to continue to iterate on that, I think, is what's really important. And one of the things that's

32:45 made us successful, but hopefully we can get more of the good startups in use. So we've got a few really good ones. But again, just want to see more of that. Because I very much believe that

32:57 energy transition is going to run through Houston. Because this is where all the money is.

33:02 You spoke about how when the Winter Storm ure came, it was a difficult time for the retailers. How did that impact you? Was there a moment where you were like, I don't know if we're going to make

33:14 it through this? Oh, yeah, totally.

33:19 So I've told some variations of this, but I think I got April at this super nerdy thing on UFE, which is unaccounted for energy. And that was a number that was very, very small, historically,

33:34 like plus or minus like 01 And it went through the roof. And the charges were astronomical, had a9, 000 price cap. Ansellaries went up to22, 000 per megawatt hour. Our clients had a lot of

33:48 megawatt hours. We had very like, less than a million meters on at that point. This is one of the things that kind of made us well known. So I got a phone call and guy calls me from one of the

34:02 biggest retailers period. And he's like, you don't know who I am. But our CEO and you had breakfast a year ago. and he thinks that you may be able to help us solve this problem. And so we went

34:15 back and resettled all of these meters. And by we, I mean, A, by himself just didn't sleep and went and resettled these meters. We were charging most of our customers. So resettling means like

34:25 you're verifying the amount that came out. And the issue was that there was a bunch of energy on accounted for. Exactly. And so you're reconciled. Exactly. Okay. And from our standpoint, so my

34:36 meter was on one of our retailers and it showed that I had power I knew I did not have power. And so they were just guessing and just filling in the blanks and I was like, we did not have power for

34:46 like 72 hours. There's no chance I'm getting a bill for these hours. And there's no chance that my customer who I was on and we would use my meter for things and just look at the 22 digit number and

35:00 we're like, this is not, it's not possible. And so we went back and said, hey, there's a whole bunch of zeros in there because I mean, as you know, so many people lost power here and for very

35:09 long periods of time. And so we were able to go in and resettle that.

35:15 ERCOT resettled things out of, I guess, out of cycle. They normally have like certain like X days that they will go in and resettle and say, Hey, we'll try again, we'll try again, with the last

35:28 one being 60 days. And so we were able to go in, resettle that data, and then send it to ERCOT. And so we were doing these reports for 5, 000 bucks. People would have paid 100, 000 for it, but

35:39 we just went out there and kind of got our name out there. I was blowing up LinkedIn, telling anybody who was listening, tagging the CEO of ERCOT in my post, and just going for it. And so, I

35:51 mean, as with the Digital Wildcat or Studio, they're all for good PR stuff like that when

35:59 you're especially early stage. And so that was what really kind of did it. And then we got brought up at the ERCOT meeting. One of my good friends, she's like, You're talking about Amperon. We

36:10 have 10 employees. I was like, why are they talking about us at the meeting? And so that was kind of one of the first like, uh-huh moments of like, we're really onto something. And then it's

36:20 been a really good journey from there, but we had a bunch of retailers that got kind of, we'll call like, Aquahired, or just kind of went out of business altogether just because those were such

36:31 high prices. And you saw what, you couldn't go back and ask the end user for that But the way most of the contracts are structured, you're just locked in it at certain rate. So yeah, I think that

36:44 was one of the definitely the first kind of aha moments. And then we also, we just, when COVID hit, we just gave away our product for free. And so just said, we'll give it away for two months,

36:57 come back, and then we wound up having like a 90 hit rate on that. And I was like, okay, we're on to something. 'Cause people are like, I'm not gonna buy it. No, I don't have time. And then

37:05 we're like, just take it And so they took it and it worked. So there were a few moments like that where we really had to put ourselves out on a limb, but it hit pretty hard. Do you find that it's

37:21 as much about getting brain space with people? Because once they get it, they clearly, they find it valuable, right? So how do you market? I guess that's a great way to put it. You've got to

37:32 figure out the correct time for someone to do it. And it's different for everyone. Someone could say, hey, we're going through something I can't even tell you about. So I think just really good,

37:42 and this is something I wasn't good at. It's a seller. Thankfully, I don't have to sell anymore. And so going through and asking those very clear closing questions of saying, when are you

37:53 available to look at this? When do you think you'll have the bandwidth? When do you think your team is? Because for us, I can turn someone on for a trial in seconds. But if they're not going to

38:01 look at it, it's pointless. And then you just get frustrated and you're like, why don't I have an answer? Well, I haven't had time And so just setting those expectations and the sales call is

38:09 something that like Tom as our VPS sales does a phenomenal job. And so, 'cause saying, hey, we're not good right now, but come back in 45 days, because I know that X Life event is about to

38:20 happen, or X Company event is about to happen. So I'm just, give me a second. It is extremely helpful. So it's, I think just that full setting expectation, because yeah, you just need them to

38:31 have the brain space is a great way to put it to say, hey, let me actually do a proper review of the product That's also why now having a technical sales team, we're able to go in and provide a lot

38:44 of those reports for you that you don't have to do yourself. So you can go back and look at how we track historically, 'cause you're gonna put some weight into how we do for the month that you're on

38:54 trial. You're also gonna go look at the back cast and see, like, hey, how did this stack up? How did it do on this really inclement day? How did it do on this really? Well, it's whole day. Do

39:05 you find you have to give people like a 90 day trial is kind of the main thing? It's not only a 30-day trial, but we do have some, especially if it's customized data, just because I can't give you

39:15 a backcast of what your meters, I mean, you can, but it's different than us versus the Texas ISO. I can tell you what our forecast was every day at 10 am, going back to November of 2018. Great.

39:30 If you give me your specific set of meters, then it's gonna take a minute. So yeah, just different things for different people. But again, at the end of the day, if we get in there, then it's a

39:39 very good chance we're gonna stay. Yeah. Good. Yeah, talk to us about your fundraising journey, especially also being a tech company from Houston. A lot of Houston founders complain that, we

39:52 don't have a very evolved venture capital ecosystem. We don't. Did you raise capital out of Houston? Maynard Holt Veritan. Oh, of course. Yeah. Maynard's the best. He's the best, yes. Yeah,

40:03 so I mean, talk to us about that. I mean, you did have to do a lot of hustling in the beginning, you and your roots. co-founder did a lot of the work. You did the sales. He went and checked the

40:14 meters.

40:16 And then going past that and then having enough traction to show to VCs who would then come on and invest in your company, how was that journey being from Houston and then having to, I guess, go

40:29 outside of Houston to actually get the capital? Yes. Our first round, we started the company in New York and then I moved back to Houston in fall of 2019. Kind of nailed that one.

40:44 That'd be lucky to be good

40:47 and then we like properly moved everything down here like the year after. I mean, Texas is such a great state to do business. New York, not as much. And so from the raising capital standpoint,

41:01 the very first round, that was all eight. I'll give credit where credit is due. He had a relationship with notation capital. from just being in and around the New York venture scene. I just

41:12 showed up to New York and notation capital is the best. They came in early stage, never done an energy investment, said, here we go, we'll give you all a shot. So they led our first round and

41:26 then came in pretty heavy in our seed round. So that one wasn't that hard. The A was not that hard, we had multiple term sheets. The B was really, really fast And so we signed a term sheet eight

41:40 weeks after kicking off the fundraise in summer of 23, which was historically a pretty bad time to raise capital. The seed round was where we had the problem. And this is where going back to the

41:50 product market fit story, we were really good in Texas and we had a client in Australia, the Australian energy market operator, and that was it. So there was no East Coast. The East Coast

42:01 operates very differently than Texas does. And so they hadn't seen what they wanted to see from a repeatable process standpoint.

42:10 and our revenue was, someone told me and I never knew it would be so true. The hardest capital to raise is when you're between1 of annual recurring revenue and999, 000 of annual recurring revenue.

42:20 'Cause you

42:24 get measured differently. You just don't like to have, so I advise some early stage startups, some formally, some informally. And the thing I tell them is don't take revenue until you're ready to

42:37 go Because then someone has set a price on this. And you can see, no, no, no, it's a big name. It's this. And they'll say, Guess what? We only have one metric. You took, this product is

42:47 worth100. End of story, that's what it is. And so that's where it's just hard 'cause you get to tell the story and own the narrative when you have the pre-seed round of, Hey, these are two

42:57 talented peoplewho just met each other and even myself. And you can just believe in the team and say, Hey, this is gonna work.

43:07 But then once you've actually proven out what the product is worth, then that's where you're stuck. And so there's so many things I would do differently in our seed round. And there's so many

43:13 things I tell early stage companies now to do differently just because, again, once you put that dollar amount on it, then they can extrapolate that forever. Our first pitch deck, I'm pretty sure,

43:24 said got we've 'cause, hilarious is which, month per meter per think10 I, 10 million meters under management. We

43:31 were doing, unfortunately

43:35 Less than a hundred million a month. So, just a little. And so, you kind of look on that, and you look back on your first deck, we should be a unicorn already based on this 2017 deck, 'cause

43:47 everyone is by this

43:49 stage, right? It would have IPOed like three years ago, it's fine. And so, that's where you just kind of like, get to own the narrative. And that's what I think Houston doesn't have as many

43:57 early stage funds who could go in, really say like, hey, let me teach you how to run a company. Let me teach you how to do this. Let me talk to you about cap table management. Let me teach you

44:07 about board management. Just so many things that I'm definitely passionate about with this being my home that we need help on. The thing Houston does have is once you have product market fit, the

44:20 CVC space is good. So corporate venture capital, this is where you've got a bunch of like Barbara Berger at Chevron kind of invented the game and the energy space here and then so many others have

44:31 gone on and you see that the big like Shell has one or Ramco has one, Oxy has one and they're that's great, but they're not a good early stage investor because they're going to tell you that, hey,

44:42 build this for my company, build this for my, but we have had a great like we've had great success with CVC's. We work with the big utility that we have quite a bit of contracts with I mentioned D

44:55 shawl or stead has been an amazing investor. So that's kind of where I think. the early stage in Houston. They need people to be able to take that first step. But unfortunately, everyone here is

45:08 like, what's your EBITDA? And you're like, there's no EBITDA. That's not gonna be EBITDA for quite a while. It's venture. But I think that's what we're trying to help with the Nexus is to

45:17 provide some of that original, I don't know, infrastructure for the startups and the founders. Totally. Because it's so challenging. And we've been working with some founders who come through who

45:28 have everything from messy cap tables to an unclear IP portfolio. Yeah. And you gotta get them ready for financing. Yeah. And that's something that was in bits and pieces in Houston, but I don't

45:40 think there's been a lot of consolidated kind of operators and it's a big gap here, but you gotta build that kind of tacit knowledge and ecosystem and it doesn't get built overnight. So I totally,

45:53 it totally resonates with me when you say that that didn't exist here I'm gonna share an anecdote that I saw. which you might appreciate. So I was looking at some early startup, unit economics for

46:04 a while and was realizing that there's a whole group of startups that will mark list price for their total revenue and then they'll give discounts to their customers and more that is a sales expense.

46:18 Oh, geez. So that's how you get through the million dollar revenue number, you're like, you know what, this was cost to customer acquisitions. We gave them a 40 discount and it allows you to

46:28 almost double your revenue number. That's funny, you have definitely never done that. That was very much a California kind of thing. Wow. I don't think that would ever cross our minds here and he

46:38 used to do that. But when you realize it's all driven by revenue metrics and that measurement, you're like, okay. I guess we're gonna start giving everybody a 99 discount and it's still going on.

46:49 Yeah, that's how you get unicorn status. It's all sales discount right here. That's really funny. I've heard a lot of things, I've not heard that Yeah, so yeah. tricks of the trade, I guess,

47:00 but don't do that here. Definitely don't do that here. They will tear your financials apart by whatever, like, former private equity person that they're actually trying to get an adventure.

47:13 Were there any failures that you're comfortable sharing along the way? They learned along the way, other than recognizing revenue, truly. Yeah, apparently this big on this one, our own news is

47:27 gonna be at least a 50 discount on everything,

47:31 failures.

47:34 Yeah, there's definitely been, there's been plenty, I think there's products that we've rushed to go out the door too quick that we didn't sit down with like the right end user, the right customer.

47:45 And so I think we've had something like that. I mean, our seed round was really hard not telling the full story. I caught up with a woman today this morning that was at a CBC end Um

47:59 And I, she's like, we haven't talked in like a couple of years. I'd love to hear how things are going. And we were like, you literally like, and for on has gotten bigger than your story that you

48:10 pitched me and she didn't invest, did not invest in for the seed round. And I was like, cause I was a short term trader. And so like I didn't look at enough vision. I didn't look at like what the

48:21 future could look like. And so I think that's the one piece that I definitely failed at early Um, is just not seeing like what this could actually become. And so now I spend a lot of time doing

48:32 that. And this is also where you surround yourself with the best people. I mean, you want to talk to people who have that good forward vision. And I mean, that's what I saw in energy climate. We

48:43 got to go to so many cool events with really smart people. Um, there's like a fun dinner that, uh, like Charlie Nelson through and just like a whole bunch of cool founders and all of us get to So

48:54 I think just really looking at vision is one of the things that I missed. Um, and the way you can get around this stuff of just like not knowing how to fundraise. I sat down with, um, a friend of

49:04 mine, Kieran, who's an amazing fundraiser at Arcadia and he's the one who really sat me down and helped me tighten up my story for the series A. The A and B have been pretty easy. Um, and so I

49:16 think just being around those mentors and that's what I see that you're building now is just having the, that group of the kind of like, it's a lonely job. Let's all do this together. Um, so I

49:26 think that's definitely one of the biggest mistakes, uh, nisses that we had is just not surrounding myself as much as I could with like other founders. And part of it was COVID too, right? Not my

49:37 fault. Um, but I would ping anybody who would, who would chat and just basically say like, Hey, like you want to talk about this? And so it had some like pretty big name CEOs like say, yeah,

49:46 happy to talk I'm bored out of my mind sitting in locked in my house in San Francisco. I was like, cool. Um, and so I think that that's one of the definite misses we've. We've had it in front,

49:56 so just surrounding yourself with those right people. I've got Tim Healy on my board and he was the co-founder and CEO of Interknock. I mean, he did it for 17 years. I'm speaking with him tomorrow

50:06 at the energized summit about that. And so like having someone who's been through this is just extremely helpful. So just get your advisors on correctly and then also like become a part of that

50:17 ecosystem that like you're creating right now. We had an urban future lab up in New York and

50:25 it was a woman passed the pencil, he was running it and she's unbelievable. And so it was nice to go in there and sit and look at another company and then see their100 million fund ratio. Wow, I

50:35 still got work to do. It's really how much she does by herself. I know she has a whole team behind her, but she is a tiny couple. She's moved on now. Yeah, yeah, yeah. So let's see, when we

50:46 think about the innovation ecosystem here in Houston, is there any, And you've kind of seen it of all of across the life of Ampran, is there anything you're most proud of that we've developed?

50:58 Houston, I think, I mean having climb a week was a big W. I should have been done a long time ago.

51:06 Whatever. It's here. Better wait than never. I mean just having these hubs, I mean as we're seeing Canon, I on Greentown, like all these different all these different like actual workspaces and

51:19 office bases are extremely helpful just from the collaboration piece to say like do you have a good accountant like who is really cheap and can still do the job and I only need them X amount of hours

51:30 per month and who's your lawyer just the simple things that you just don't know because there's lawyers everywhere in Houston but the ones who actually know how to do venture is not a thing. So I

51:43 think that Houston's doing a lot better of kind of like aggregating this I mean what you're putting together I think getting all the right founders together. I guess you had your launch just like

51:51 during climate week. Those are the type of things that I'm definitely proud that we're finally having from like a collaborative standpoint. And there needs to be a lot more because there were so

52:00 many cool hubs that we had in New York. So UFL was out in Brooklyn Urban Future Lab. We got in there from winning the50, 000 pitch competition for like best clean tech startup we won that in April

52:13 of 2019. That was a big like everyone It was a big deal on one of our first kind of early, like, well, we're on to something. And then we worked in a space like next to Grand Central. There was

52:25 just a whole bunch of startups like backed by the city. And so just having those ecosystems is something that I'm definitely proud that Houston's kind of like kind of putting together now. There's

52:37 still a lot of room to grow. And we're also finally seeing some good venture capital down here I mean what when you have

52:47 big names such as like Tudor, Prickering, Holt, like Dan, Bobby, and Maynard all doing this. I mean, we made room for Maynard on our cap table just 'cause it was so important to have someone

52:57 who really got it and also understood that a lot of this energy transition needs to be paid for by oil and gas dollars. And Maynard can text any of those CEOs.

53:10 So he definitely had a big, as I was negotiating this, he had a really, really big CEO in town call my cell phone and say, I should take money for Maynard. And I was like, I've never felt

53:24 more out of my league than having Maynard have this big name, like Fortune 100 CEO call me. I'm like, I gotta make this work. So it's fun having them being Dan Pickering's conference, my VP of

53:42 finances there right now So just a lot of this stuff is going to be - Those are the people who really need to lead it. Yeah, I was gonna say we're gonna ask who the Hidden Gem was, but I think it's

53:52 Maynard Holt and Burton is one of the Hidden Gems in Houston. We don't talk about that much. Yeah. Yeah, they're definitely a Hidden Gem. I mean, you've got some other early stage, or like

54:03 they're kind of like mid stage, like kind of seed AB. And so, no, there's definitely a bunch more Hidden Gems kind of popping up certain and do that early, that early round stuff So, this is

54:18 what I may want to come back as. Yeah. Whenever this wraps up in quite a while, but going back and doing the venture thing and just helping out those early stage founders. Powerhouses here now,

54:30 next Deras, had us a really good venture arm all here in Houston. So, yeah, making some progress. Good. We ask most of our guests to share their climate impact story. So, we were wondering if

54:43 you have one that you could share with us. From a climate impact standpoint,

54:52 we've got four of the biggest renewable companies on the cap table, and that's where I think it really makes a difference, especially now as we're having asset level solar, asset level wind. I

55:03 think this is my personal thesis that things are gonna start moving away from PPAs just because you are seeing so much volatility in the market. If you're a battery, you don't want the whole thing

55:14 told You probably didn't secure a really high price for April and early May, but there were two events in April and one in early May that paid for a really big chunk of your year out of the blue that,

55:25 and it was very predictable. The sun was going down and the wind was joining it. And so, and there was 25, 000 megawatts on outage. It was pretty obvious. So I think that you're gonna

55:39 continue to see that. So I think us just helping renewables and helping batteries make money. 'Cause at the end of the day, you want them to get rich so that they can build more, right? And so

55:52 that's the whole thesis behind it. I mean, in this game, you can do tax credits for a certain amount, but at the end of the time, it still goes back to capitalistic. Like somebody has to stand

56:03 on its own two feet and then kind of go that way. That's what's so amazing now about what we see in like the EV space is I bought a Tesla because it was really fast and had a ton of storage and was

56:14 like great car, good price, all of the above. And so now you're seeing that and not just buying it because it's raw rods and electric vehicles, the best car at the time when I was purchasing a car

56:25 last year. So I think that that's the type of thing from an impact that Amprons helped just continue on that path of those companies make more money than they put more steel in the ground. And so

56:38 that's why it's great to have again all those different customer bases that some are obviously more capitalistic like traders, but then you also have IPPs. and utilities and others that are just

56:49 trying to do what's best for the environment and. Yeah. Yeah. No, it's really important because, I mean, we're gonna have more different types of energy mix going forward and we're gonna need

57:01 that reliability of having power all the time. Sustainable and reliable and cost-effective. Yeah, you're helping us do that, so thank you.

57:15 It's a, thank you, it's a hard problem You're gonna solve. Yeah. Is there anything our audience can do to support you and your goals? Why my stuff?

57:26 Tell a friend about it, yeah. Yeah, to tell a friend, I mean, yeah, it's so much that this is word of mouth. I mean, that's how you get to know people, that's how you trust other products.

57:36 And so nothing makes me happier than walking into an office or having one of our sellers walk in and being like, this huge company, this huge company, this huge company all said that we - should

57:46 have already done this. And so that happened with me. And I was like, that's really fun. That's a good year. You're going to listen to these multi-billion dollar companies. So yeah, definitely

57:55 telefront, we're continuing to get the word out there. Everything's grown so rapidly. We had 28 employees at the beginning of last year and now at 100. So things are going up, spending more

58:08 marketing dollars and then just doing more of this. This is the fun stuff, just kind of getting the story out there, not only for AMPRON, but also just the Houston ecosystem in general. So yeah,

58:19 telefriends, definitely in whatever you only can do. All right. Yeah. No, thank you so much for helping us put Houston on the map when it comes to innovation, tech, and energy. So. And AI.

58:29 And AI, yes. All of the above. So, and thank you for coming on to our podcast. My pleasure. This was blast.