Is AI the New Crypto When It Comes To Energy Consumption?

This episode will put you ahead of the curve on some key conversations taking place these days in the renewables sector.

Numerous companies are striving to match their energy consumption – on a daily, intraday or evenly hourly basis – with renewable energy sources. It’s an ambitious goal, so today's guests are Gridmatic Founder and CEO Matt Wytock and VP of Business Development David Miller. Their team leverages artificial intelligence to make that matching a reality.

And speaking of AI … the rise of ChatGPT and other generative AI solutions has propelled the technology to the top of the headlines. Not a day goes by when we don’t hear about how some company is using AI in a new way … or how many jobs AI is going to make obsolete. But you know what we don’t hear much about? How much energy it takes to power the computers that are cranking thru those AI algorithms. In fact, This story arch is eerily similar to cryptocurrency. Crypto had been on the scene for a while before people started asking hard questions about how much energy all those crypto mining computers were pulling from the grid. AI is no different. 

Matt and David are going to share their insights on how AI can help solve an energy consumption conundrum created by the popularity of AI. 

We will also discuss developments in the market for Renewable Energy Certificates (RECs) and the impact the Inflation Reduction Act has had on the entire energy industry. 

Sign up for the Renewable Energy SmartBrief

Follow the show on Twitter @RenewablesPod

Transcript

(Note: This transcript was created using artificial intelligence. It has not been edited verbatim.)

Sean McMahon  0:00  

Hey what's up everyone and welcome to the Renewable Energy Smartpod. I'm your host, Sean McMahon. And we've got a show today that will put you ahead of the curve on some of the key conversations taking place these days and the renewable sector

Numerous companies are striving to match their energy consumption on a daily intraday or even hourly basis with renewable energy sources. It's an ambitious goal. And I'm going to be joined in a minute by Matt Wytock and David Miller from Gridmatic. Their team leverages artificial intelligence to make that energy matching a reality. Oh, and speaking of AI, the rise of chat GPT and other generative AI solutions has propelled the technology to the top of the headlines. Not a day goes by when we don't hear about some company using AI in some new way. Or see stories about how many jobs AI is going to make obsolete. But you know what we don't hear much about how much energy it takes to power the computers that are cranking through all those AI algorithms. In fact, this story arc is eerily similar to cryptocurrency. Crypto had been on the scene for a while, before people started asking some hard questions about how much energy all those crypto mining computers were pulling from the grid. Well, guess what? AI is no different. So Matt and David are also going to share their insights on how AI can help solve an energy consumption conundrum created by the popularity of AI. So yeah, energy matching and AI’s own energy footprint are a couple of topics that are going to be around for a while. So I hope you enjoy learning more about them today with the team from Gridmatic.

Hello, everyone, and welcome back to the show. I'm joined today by part of the brain trust from Gridmatic. First up is Matt White talk the founder and CEO. Matt, how are you doing?

Matt Wytock  2:04  

I'm doing great, Sean, great to be with you today.

Sean McMahon  2:07  

And I'm also joined by David Miller, the Vice President of Business Development. David, how you doing?

David Miller  2:12  

Great. Thanks for having me.

Sean McMahon  2:14  

All right. Well, I want to start talking a little bit about you know, what your firm does, a lot of our listeners might not be familiar with it. So why don't you kind of give us the basics? What is Gridmatic do? And where are your segments of business? 

David Miller  2:24  

Gridmatic is a clean energy power marketer, we contract with wind, solar battery storage assets to buy the power from them, and then we sell it to data centers and other clean energy buyers.

Sean McMahon  2:38  

How long have you been operating in this space?

David Miller  2:40  

We were founded in 2017.

Sean McMahon  2:42  

Okay, so among those energy suppliers and customers, and maybe even storage owners, you know, what kind of trends are you seeing these days in those markets,

David Miller  2:50  

We're seeing a lot of growth and all of these activities due to a variety of industry factors. In particular, the inflation Reduction Act has really charged forward a lot of the supply of renewable energy. And then on the demand side, we're seeing a lot of his customers that set more ESG related environmental goals for the carbon intensity of their electricity consumption. And so that's leading to higher demand as well. And when you add all that up together, get a lot more clean energy, but simultaneously, more risk in the marketplace due to the variability of that clean energy, which is a lot more volatile than the electricity grid has traditionally operated.

Sean McMahon  3:37  

How do your offerings work? You know, how do they help your customers navigate that volatility?

Matt Wytock  3:42  

Sure, yeah. You know, traditionally, the way that, you know, folks operate large power plants in the grid is a fairly manual process, right? So you have teams of people who are looking at weather forecast, price forecast quantity forecasts, deciding how they're going to schedule, their power plant deciding how they're going to buy electricity to serve their customers. And so what we've done and what we're working on, is really trying to automate that entire process. And to do that, what you need to do is basically forecast what's happening with the weather what's happening with demand, what's happening at kind of a fundamental level on the grid, and then connect that to end to end automated decision making with optimization and risk management. And so really, the the product that we offer, you're sure something like battery storage is, you know, that end to end kind of operating platform that for the most part is fully automated. And you know, could do things like optimize when the battery charges or discharges.

Sean McMahon  4:42  

So we're hearing a lot about negative headlines I should say about artificial intelligence. So it sounds like this is AI actually doing some good that you're trying to tell me.

Matt Wytock  4:49  

Yeah, I mean, I think you know, AI really has a fundamental role to play in solving the balancing problem on the grid. And, you know, I think the traditional ways of burning fuels to meet demand, just you know, is a is a paradigm that doesn't exist in the future where most of the supply is going to be coming from wind and from solar. And, you know, we really need better ways some more sophisticated mechanisms for managing how we use electricity, and managing how we balance the grid. And that's, you know, that's really where AI can help.

Sean McMahon  5:25  

So obviously, reading trends about companies are trying to match their energy energy usage daily or into daily or even hourly. So is that one of the things that your platform offers? And then how does that work?

David Miller  5:35  

Yeah, that's right, we have a customer right now, that's a data center customer that is looking to do hourly matching of their consumption with renewable power. And so we are currently have been serving them for now the past six months doing that, where we match and record the generation mix from renewables lined up with the times that they're consuming. And there a lot of challenges to doing that. There's the upfront challenge of forecasting, and if what is the right resource mix to match this customer, because if you buy lots and lots of wind and solar, you can match it on an hourly basis. But that'll be very expensive, because you'll have a bunch of access that you won't be able to get full credit for. So our platform does upfront analysis to forecast the output and construct a portfolio that's lower cost. And then once you get into the operating phase, we can take advantage of the flexibility of resources in order to deliver on the matching targets that our customers set. So that could mean using wholesale battery storage to charge and discharge at the right times to make sure that the customers are being matched with real power, or taking advantage of the flexibility of the customers load shape itself, if they can have the ability to, for instance, adjust the set points on their air conditioning during certain times of the day or schedule loads at different time periods. And for all those things, the data that's available is not available in real time. So even with smart meters, which have been rolled out to most customers these days, it's typically at least a one day lag. And in one case that we have, it's can be up to 30 day lag. And likewise, on the generation side, there's often a lag of knowing exactly what the production is because market operators are not able to share information in real time. And so that has a significant need for forecasting services in order to actually do the matching at high levels. And that's really where we come in and can deliver that forecasting and optimization to make sure that the matching occurs.

Sean McMahon  7:37  

Okay, then when it comes to resiliency, obviously, there's a lot of demand out there for various reasons, right, you know, his load to the grid, even natural disasters, things like that. What are the conversations you're having with market participants on that topic?

David Miller  7:49  

Well, I think the way we interact with the market on resiliency is primarily as relates to battery storage, we're working with a number of large scale grid scale battery storage project owners, and for them being available during the peak periods of the grid is really critically important. And in some cases, that means thinking about when they're going to be scheduling their load. So as an example, we have a contract with a customer where we are effectively leasing a large battery on the grid. And in that contract, we can provide notice a few days in advance if we're expecting a major weather event to occur. And that notice has contracted implications for you know, the availability of the facility and gives them heads up that we really want to make sure that it's available, which is important for us to make sure that we're, you know, available during those peak periods for financial reasons. But it's also important for the grid, that it's sending the right price signals to make sure that these capacity resources, which are designed to be available during the peak periods actually are available during peak periods. And to make sure that that happens, it does require this forecasting, multi day basis.

Sean McMahon  8:59  

Okay, now, obviously, you mentioned your platform as kind of AI underpinnings, and you know, hearing a lot about AI in the news. I mean, I was joking earlier, but it's definitely a hot topic. And one of the notes I was seeing is that all these server farms, whatever computer farms that everyone's building out, those are gonna kind of be the new crypto right. A few years ago, we all focused on how crypto was straining the grid and hogging up all this energy. Are we in a place where that AI arms race, if you will, will become a pretty massive burden on the grid?

Matt Wytock  9:27  

You know, that's, that's great question. And, you know, there is an explosion happening in AI right now and a lot of excitement. I think, you know, when you look at kind of the broader story around decarbonisation, a lot of that is about greater electrification, right? So, you know, you know, returning these end uses of energy, so think about transportation, right? How do we make transportation clean? Well, we convert vehicles to electric and then we charge them with clean electricity, right. So so our thesis actually is that we're going to see big growth in electricity demand across all these industries, right. And so yes, AI is going to grow. And, you know, that's going to create more demand, which sometimes creates challenges in certain grids in certain places, but also transportation, electric vehicles are growing, we're gonna see, you know, green hydrogen, you know, big, big installations are being built now. And in places like Texas, and, you know, as we move all of these, you know, heating and cooling to electricity. So we really see a lot of growth in electricity in general. And I think that's actually going to be able to, you know, build this more dynamic grid in the future where you can, as long as you can manage, you know, when you use that, that electricity, that, you know, it's actually going to lead to a better solution down the road.

Sean McMahon  10:46  

I love that it's, I mean, there's something to be said, for having a platform that uses AI, to help tackle a challenge that's kind of being compounded by the growth of AI. So there's, there's some kind of joke in there, I'm struggling to find it. But but it's in there somewhere.

Matt Wytock  11:00  

We say that we just, you know, we work for the computers in many ways. And so you know, we're just, we're just here to keep the AI running and doing the right thing.

Sean McMahon  11:10  

I think pretty soon we'll all be working for the computer's Matt. All right. So, um, one of the questions you've been fielding from, say, Energy customers, you know, as more companies are kind of trying to shift to ESG reporting and climate disclosure practices, obviously, seems like the interest and the questions from those companies have got to be ramping up. Is that right?

David Miller  11:31  

Yeah, that's right. I think, you know, we're seeing a lot of folks concerned with the status quo of kind of climate or carbon reporting. And in particular, there are a lot of concerns around greenwashing. I'm 100%. Dream, you know, what, what does that actually mean? When you dive into it, in some cases, that can be you know, companies that are doing really direct meaningful things to improve the carbon intensity of their operations. In other cases, it's a much more tenuous link. A lot of companies use renewable energy credits or recs, to claim that their electricity usage is renewable, even though those RECs may have been produced by a resource that was located somewhere very far away at time periods that were not when the company was producing it. And so it only kind of works, that regime only kind of works, if someone else is willing to kind of balance out the grid for you and doesn't care about having actual green power. And so we're getting a lot of customers are saying, Okay, we understand that there's concerns with how things have traditionally been done. But what's next, you know, like, how are we going to move forward here. And I think this this time matching is, is one where one area where there's a lot of interest, but there's a lot of uncertainty as well, there aren't really good standards. There aren't other than a few limited locations. Right now, there aren't third party exchanges that can track and register the credits on a time basis. And really, folks are looking for kind of someone to to kind of set standards and create these programs. I think ultimately, it will come from the bodies that have set up the existing standards, which tend to go back to the States, at least in the US. But in the meantime, I think companies are looking for help in designing programs with that they think are the right way that they can show as as moving in a positive direction, because I think there isn't clear alignment on what the direction should be. But I think there is a lot of alignment that the current status quo for claiming a carbon free consumption is not really accurate.

Sean McMahon  13:39  

What's your take on the evolution of the rec markets? Do you think we're gonna get there in terms of to a place where it's, you know, there's a proper exchange set up or just, you know, the infrastructure is there to do this stuff real time?

David Miller  13:50  

We hope so, I mean, the projects that we've done so far, we've done some bespoke solutions. And we would love to tap into a more centralized exchange that increases liquidity and allows more participants in the market at a lower cost to do these kinds of time matched products. I think ultimately, it'll be driven by voluntary procurements from from large industries, and then potentially, we'll see it being driven in certain sectors. So for example, the US Treasury is currently considering the guidance for what is going to make green hydrogen eligible as green hydrogen for the production tax credit. And one of the things that a lot of folks are looking for is to make sure that that green energy is time matched and produced by new renewable resources. Because if it's not, we could be creating green hydrogen, quote, unquote, that is more dirty than creating hydrogen directly from fossil fuels. And so that may be one industry where because the specifics of the tax credit are such that it's defined as requiring more advanced environmental attributes. You may see more innovation in terms of the changes from from Rex to the next generation of Ty and matched credits that are that are tailored to that industry. So I think we'll see, you know, companies that are forward looking and in specific industries, and hopefully those will lead to more standards being created, which will then allow everyone else to participate as well.

Sean McMahon  15:13  

That segues to our next topic, the Inflation Reduction Act. How have you seen that impact the markets,

David Miller  15:18  

it's been a huge impact. I mean, really, in I've worked in the renewable energy field for over a decade, and it's been the biggest policy change that that I've seen in my career related to renewable energy, I think there have been a number of third party reports that have forecasted the impact just in terms of new build of renewables, and it's on the order of 20 gigawatts, incremental wind by 2030 20, gigawatts incremental solar, something not quite that large. But similar on the storage side, and we're just seeing the certainty and the values of the of the tax credits are really making the material impact on the growth. And it's as a result, it's really pushing the pressure in the development of these projects to new places. So whereas before, you know, it may have been hard on projects to make the economics work, just based on equipment costs, and the value in the market. Now, with the tax credit, the economics look good, but the projects can't get built. Because there are other constraints, like the interconnection queues are clogged or permitting is going too slowly. So it's really kind of changed the bottleneck in the overall system, which has been the, you know, fundamentally creates new challenges to deal with.

Sean McMahon  16:29  

Matt, what are your impressions of the Inflation Reduction Act? And how it's kind of changing the conversation you have with clients or potential clients?

Matt Wytock  16:35  

Yeah, I mean, I think, as David said, you know, kind of bringing forward that certainty, around, you know, what the revenue is going to be for things like storage, which are still quite difficult to plan for, you know, some merchant battery storage, it's a difficult task to try to estimate how much revenue is going to generate. So things like the inflation Reduction Act, help, you know, people who are want to build those projects, really, you know, get them over the line and get things built. And, you know, I think that's really exciting, you know, for us, because I think they, you know, that creates a need for the types of solutions to then operate those, those battery storage projects and other new types of things. And so I think, you know, in terms of moving forward, the future, so to speak, I think that it's been very positive from our perspective.

Sean McMahon  17:26  

All right. And speaking of the future, I like to ask guests to give me bold predictions about their specific niche in the marketplace. So, Matt, why don't you go first, give any bold predictions on how the marketplace for real time renewables will look and say, five years?

Matt Wytock  17:43  

Yeah, I mean, I think, you know, the prediction that I will make is that when you look at the kind of actual orchestration of real time, energy and clean energy, so be that battery storage, be that EV charging, it's all going to be fully automated. Right. And I think, maybe that sounds obvious to us here. But I think that's, it's actually quite bold in terms of the way the grid operates today. And so that's, you know, it's not at all how gas plants operate. For example, it's not at all I'd had the grid operator, you know, we're exceeding they have people looking at screens and making decisions and so so I think the next generation you know, EV, fleets battery storage, fully automated in five years.

Sean McMahon  18:30  

David, any predictions from you?

David Miller  18:33  

My prediction is that the types of AI based models that we're using as power marketers in the marketplace will be adopted by more market participants, including the grid operators. Today, most grid planning is done through optimal power flow based models, which are physics based models of the grid. They look at each individual device on the grid and try to look at how electrons will flow and are used that way to solve a least cost dispatch. And our view is that that approach to problem solving will be augmented in certain applications by AI based models, which are statistics based models used to predict outcomes for grid operations without actually modeling the underlying physics of the grid, and are better in certain applications for looking at light scenario analysis. So my prediction is that grid operators will adopt AI based models of the grid.

Sean McMahon  19:32  

So talking about your customers, right. And your conversations with them, what are the biggest challenges that they're trying to tackle? Matt? What you take effort?

Matt Wytock  19:40  

Yeah, I think the you know, the biggest challenge that we're seeing is you folks who are trying to essentially convert their operations to clean energy, right? And so 100% clean energy, what does that mean? How do we define it? How do we account for it? How do we purchase it? How do we manage it? And so I think that's a lot of companies, you know, want to take that path. And, you know, there's been big announcements from data centers like Google and Microsoft and those teams, but actually making that a reality is an open question. And I think that's a challenge that that we're all faced with right now. And so I think that's pretty an interesting one.

Sean McMahon  20:26  

And David, business development is your role. So you must be feeling all kinds of questions. What are some of the most popular?

David Miller  20:32  

I think, questions that we often get are just around? How do we get meaning the customers get get comfortable with the solution? How do we trust it, you know, we can show results that have great performance in terms of improving the reliability or performance and profitability of a resource on electricity grid. And people really liked that. But anytime you're doing something that's different than the status quo, there's always concern of, well, you know, is it going to work in this case? Or is my situation different? Or is it something that doesn't apply for these examples. And in a big industry, with large capital costs, there's just a natural conservatism, which is justified, right, you can't just build $100 million piece of infrastructure and then make mistakes that damage it or or cause it to not perform the way it's intended to. And so I think AI based models, in general are a challenge to explain. Sometimes, you know, you can't tell exactly what's happening, even if you can see the results. And so I think that that's a challenge that both our customers and we face is just getting comfortable with making the transition to a future state of technology, which I think we at the company are true believers is going to happen and many of our customers are as well. But just figuring out what are the things to look for, that are going to get someone who's in charge of of a very, very expensive, very precious assets to get comfortable with changing the way they're doing things.

Sean McMahon  22:04  

You kind of mentioned the challenges you guys are facing as well, David. So internally, your own operations and your own growth plan, what are what are some of the biggest things are on your radar, in terms of things going to tackle to continue on that path to

David Miller  22:16  

growth? Sure, I think in some areas, it's different than others, I'd say certain cases, we really see that there's a large appetite for what we're doing. I think we have a lot of traction among certain types of customers, data centers that are very forward looking from an environmental perspective, or grid scale battery storage project owners, and they're it's really about achieving scale quickly. So being able to deliver all the the needs that we see, and do so effectively, and that high level of service. For other other types of customers, I think there is a bit of a timing question. Green hydrogen was an example that Matt raised earlier. And I think for us, we believe that it's going to happen, the the fundamental economic drivers look strong. But the timing of it's not totally clear to us, we're not experts in how it will come together. Because there's very complex systems that I don't think anyone has truly figured out how to permit and build, you know, a massive green hydrogen facility with new technology. It hasn't been done before. And so when we think about how are we going to invest our r&d? And what are the areas that we're going to focus on? There are a lot of questions of how quickly should we be focused? Or should we be moving on these different new areas? That's something we talk about a lot. Matt, anything to add there?

Matt Wytock  23:32  

Yeah, I think David had a had a good summary there, I think, you know, out of more kind of meta level, I would say that we're at this interesting intersection between software, and hardware, and, you know, physical systems. And so, you know, if you're building a startup in a pure software space, it's basically how quickly can you get people to use it, and, you know, get adoption and, you know, grow, grow grow. In our case, you know, we're kind of coming along alongside a bigger wave in the energy transition, as new assets are being built, and capital is being deployed, and so synchronizing the development of the technology and the business with, you know, those waves of of new asset deployments and new phases of building out the next generation infrastructure. It's just a really interesting question and one that, you know, we we think about a lot in terms of where should we be investing our time? And where should we, you know, what's the biggest focus for right now?

Sean McMahon  24:34  

Okay, and as we're all kind of talking about this, you know, real time renewables with, you know, time matching down to the hour and things like that, or minute by minute, eventually, I guess, is kind of the long term goal. Are there any regions or I should say, markets in the US, do you think are a little farther along the line or closer to getting there than others?

Matt Wytock  24:50  

Yeah, I think the biggest challenge with that is really, you know, a large amount of renewable penetration. So really the west you know, starting from Texas, you know, the middle The country that, you know, California has a lot of, you know, just renewable power. So that's number one. And then the second piece is the flexibility, right? So in order to get to, you know, really match up that clean energy with the demand flexibility is needed either from battery storage, hydro resources, geothermal. And so really, it's a combination of that resource mix. And I think at the moment that, you know, I would say, the middle of the country, including Texas, and the western United States has that resource mix a little bit, you know, further developed than, you know, for example, the Northeast markets, but, you know, I think everyone is focused on how do we get there as quickly as possible.

Sean McMahon  25:41  

Okay. Well, that's all the time we have today for our conversation. I really want to thank both of you for your time. I really appreciate it. Thanks for having us on. All right, everyone. Well, that's our show for today. Thank you all for listening. And if you haven't already, please subscribe or follow this show on Apple, Spotify, Google, or wherever you listen to your podcasts. And as always, please be sure to share it with your friends and colleagues. Have a great day.