EP 119: AI in Renewable Energy - Insights from NVIDIA

This is the Everyday AI show, the everyday podcast where we simplify AI and bring its power to your fingertips. Listen daily for practical advice to boost your career, business and everyday life.

So we talk all the time about how AI can help power our economy, but can it literally power our future? That's one of the things that we're going to talk about today on Everyday AI. This is your daily livestream, podcast and free daily newsletter helping everyday people like me and you make sense of what's going on in the world of AI. Because there's a lot going on and there's things going on that we might not even really think about in our day to day lives. Like how does AI impact our energy usage? And it's more than you might think.

So.

So stick around. We're bringing one of the leading experts from Nvidia to actually help us answer some of those questions about how AI and power relate to each other. I'm excited to talk about it, but before we do, let's get to the AI news like we do every single day, Monday through Friday. So let's start at the top. AI is here to stay in medicine, and I'd say it's in a big way. So Mayo Clinic, one of the leading medical institutions in the world, has just recently appointed a chief Artificial Intelligence officer. Y', all, this is something I've been telling companies for many months. You gotta get a chief Artificial intelligence officer. So radiologist Bhavik Patel was recently appointed to the role. So make sure you check that out in the daily newsletter because also in quoted in the article is former guest and everyday AI regular Dr. Harvey Castro. So see what he has to say about it as well. All right, could there be a new type of generative search on the way? So according to some SEO specialists who spotted this in the wild, Google could be shying away from its traditional search generative experience. So what's called sge, Google debuted this a couple months ago that really changes what search results look like and require. It relies much more on the generative experience. So specialists have spotted what they're calling SGE Light, which is essentially a stripped down version of Google's new search generative experience. So it looks much, much different. And people are wondering if this is kind of signaling from Google if they're stepping away from this generative experience. I'm not sure. I actually just think that they're trying to ease it in in a more gradual way. That's my take. All right, last. Are smarter robots coming? Probably. So Google's AI arm DeepMind is working with more than 30 other research institutions with the goal of creating a general purpose AI system that can work with different types of physical robots and perform many tasks. So essentially all these companies are getting together and they're creating a better way to train physical robots. So, so the pro the the project is called OpenX embodiment and it's already achieved superior results in comparison to the commonly used methods for training robots. Wow. Smarter robots on the way New Google Search Chief Information Chief Artificial Information Officers so, so so many things going on. So make sure if you haven't already, go to your everyday AI.com to sign up for that free daily newsletter. But you might be here to talk about AI's role in renewable energy. I'm excited for our guests. I hope you are as well. So please help me welcome to the show Mark Spieler, the senior Managing Director of Global Energy Industry at Nvidia. Mark, good morning. Thanks for joining us.

Good morning, Jordan. Good to see you.

Yeah. All right, so hey, just as a reminder everyone joining us live, thank you. If, if you have questions about AI's role in renewable energy and maybe what Nvidia is working on in that realm, get your question, get your questions in. We'd love to take questions on the show. And as a reminder, if you are listening on the podcast, check your show notes. You can always come in after the fact. Join the conversation here on LinkedIn. Mark, maybe just start us off at the top, maybe just talk a little bit about what you do in your role as the Managing director of Global Energy Industry at Nvidia.

Sure. So I joined Nvidia about four and a half years ago to basically come in and run their energy vertical. Nvidia goes to market by industry, so I have peers across healthcare and retail and financial services. I run the energy industry and what my goal is and role is is to take our horizontal platforms and software stacks and use them to solve problems in the energy industry and then build an ecosystem of partners that, that can leverage those software stacks and create solutions for end customers to leverage AI high performance computing visualization, but specifically targeting energy industry use cases.

So you know, real like let's even hit rewind because you know, I think maybe if you don't follow energy, and I know not all of us do, but it seems like there's a lot of problems or challenges in the industry, in the energy industry because more and more, you know, the everyday person companies, it seems like everyone's energy needs are increasing. Right. So maybe, maybe Mark, could you just talk A little bit about some of those challenges you face because of this increased energy consumption and demand as well.

Sure. So the energy industry is, is, is huge. Right. Most countries economies are based on energy, right? Right. Typically oil and gas, but now more and more renewables. Right. Every country requires energy to perform and that mix of energy. And you hear the industry talking a lot about the energy transition. And we originally went from burning wood to oil and gas and coal, and now to renewables. That transition has been going for as long as we've had energy. And now the goal is how do we help create the increasing demand for energy in the most environmentally friendly way as possible? And so how do we continue to develop the oil and gas requirements that we need in order to have reliable and resilient energy production? But how do we replace those with renewables in a way that we can depend on those things and store the energy that's created in the forms of either hydrogen or in batteries and others for when the wind isn't blowing or the sun isn't shining. Right. So that whole energy space is very dynamic. And some of the world's largest energy companies are looking at how do they accelerate that transition, but how are they able to meet the energy demands at the same time? And so the technology is not quite there, but AI is helping them develop new solutions fast. And then the biggest thing is probably how do they balance, how do they balance the supply and demand? It used to be pretty easy. Now it's highly complex because energy is being produced in far more places. Rooftop solar battery walls, EVs that can plug in and push back to the grid, all of those things. It's very complex. And so the industry is, is adapting the best they can, but regulations are pushing very fast as well.

Yeah. And you know, you brought up so many, so many great points there, but it spurred a question in my mind is, you know, when the average person, right. Is using more energy, you know, you brought up a great example is, you know, there's more electric cars, you know, there's more electric vehicles, you know, now than there were, you know, five or 10 years ago. And, and even things we don't really think about, right. Like a lot of our, you know, listeners and viewers use now generative AI, you know, pretty, pretty frequently. And that creates increased energy demands. So, you know, with all of these kind of, I guess, new energy suckers, right, like everything that needs the energy, how specifically is AI even used, you know, to address these things? Because I think sometimes we just say, oh, I can do this and I can do that. But specifically, you know, how can AI address kind of balancing this, this energy, this energy need.

Sure. So let me, let me provide a scenario that most people are used to. Right. I have to get across town. I have an 11:30 meeting. What time should I leave my house? Right. And what's the best route for me to go? You know, when I was a kid, my dad used to get up really early and he used to watch the news before work. He'd see where there was traffic and he would make a decision. Once he left the house, he was all in. Right. Because he didn't have real time updates. Today we put it in our navigation system. It tells us it's going to take 19 minutes to get to this location. And here's the route I need to go. Right. As we get more and more cars on the road, or if there's an accident or a downfall, we start to see that, you know, we get rerouted. Right. And that the time frame expands. Well, energy is going to be the same way. Right. You know, as, as we start to put more and more electric vehicles on the grid, as we have people producing their own solar locally, how much energy needs to be produced and where does it need to be moved in real time? Does somebody have an electric vehicle? Does somebody have a generator? Do they have solar? So the ability to monitor those things in real time and to be able to make good decisions means that we're going to have to produce less energy and less energy will be wasted if we had to put three times the cars on the road in the next seven years. And when we talk about the goals for EVs and everything else, it's far more than three times. Right. You can either build incredibly amounts of infrastructure and new roads, or you could redirect traffic using technology like Waze. What we're saying is there's no way to possibly build enough infrastructure to accommodate. So we're going to have to use AI. We're going to have to simulate what's going to happen, and we're going to have to take real time data feeds, compute it at the edge, and basically create a balanced grid using artificial intelligence.

Yeah. And, and you know, maybe let's talk about that a little bit because these are things, you know, I, I wasn't even aware of. We were briefly talking before the show, but talk about like the importance of balancing the grid because, you know, essentially in times if energy is, is not used, it's wasted. Right. So, like, talk a little bit about that and then how AI is Is can help solve some of those challenges of, of balancing it, right?

So if, if you think about in the past, right, we had very centralized energy generation, right? Big oil and gas turbines, coal turbines, steam turbines, those things, and basically they would push out one direction. Well, today those things still exist and they produce a lot of the energy. But you also have batteries throughout the grid. You have EVs, you have people with rooftop solar, you have solar farms, windmills. We've expanded from, I think it was 8,000 points of generation in the US to millions of points of generation and storage. If you think about how much energy they need to produce at centralized plants, they predict that based on the temperature, based on how much electricity can be produced out at the edge, and they make their best estimates. And when they have excess capacity, they have to either sell it off, less expensive, and try to store it. But if there's no place to store it, they end up having to run it to the ground. And that's expensive and it's wasteful. So, you know, and they don't want to create too little, right, because some people then won't have enough energy. And we've seen this when we have storms and other things or bad weather and parts of the grid are down. So that balance is going to become even more critical as more distributed energy resources come online. And the amount of time it takes to interconnect these things is significant because they want to study it and understand how much power is that going to take off the grid and put back on the grid. And I think AI is going to accelerate that time because as the customers or businesses want to add new energy resources onto the grid, AI can quickly analyze the capacity, the impact, and allow people to have much quicker interconnects into the grid.

You know, hey, and just, just as a reminder, if you're joining us, we have Mark Spieler, the senior Managing Director of Global Energy Industry at Nvidia. So if you have questions, make sure to get them in, you know, for our audience here on the live stream. You know, Mark, one thing you've talked about is kind of on the front end how AI can help the everyday person make smarter decisions, whether it's their commute or something else. And then on the back end, kind of in energy storage and making sure we're not wasteful with the energy and balancing it on the grid. But I'm curious, what about the actual production? Can AI help and make energy plants more efficient? And if so, how does that work?

Once again, there's a lot of ways Generative AI can help with designs, design more efficient turbines, gas turbines, coal turbines. It can help produce more efficient wind turbines. One of the things that we're doing right now is we're working with a few different wind companies on combining weather forecasting with how they design an actual wind farm. Right. Because one of the things you need to think about is a lot of people think that you want to optimize the performance on a single wind turbine, but really you want the entire wind farm to be optimized. And that may mean sub optimizing some of the turbines at the beginning so that you get better production later on. And so basically we have tools that are used for wind wake optimization. Right. Just like off the back of the boat, when you're driving the boat, you see the water wake off of the back. Well, when you have lots of different wind turbines stacked in rows, the wind that's coming off of the back of those creates an impact for the turbines behind it. And so you want to make sure that when you set up a wind farm that you're optimizing for the complete farm. And using AI and what we call physics, machine learning, where we actually teach machine learning models or neural networks physics, we can actually simulate the impact of that wind without having to do full physics simulations, which may take months. We can now do it in minutes or hours. Therefore, we can continue to run different scenarios to really get the most production. And then when we overlay weather data and predictions, we can actually determine the best location and the best design for a wind farm. Thinking out 10, 20 years, not just what the impact is today. Right.

It's fascinating. Yeah. That you can take, and I think people that use generative AI, these tasks that used to take hours or days, you know, you can be doing now in minutes. So it's fascinating to hear that, you know, kind of these, these same, you know, large gains are being made around our energy, which is just fascinating. So a question here from, from Stuart. Stuart, thanks for joining. So he's asking, love to hear anything about how green shipping, example, large tankers moving around through oceans, freshwater, rivers to shift off fossil, fossil fuel fuels. So asking anything you can point us to on innovations around green shipping.

Are you still running in circles trying to figure out how to actually grow.

Your business with AI?

Maybe your company has been tinkering with large language models for a year or more, but can't really get traction to find ROI on gen AI. Hey, this is Jordan Wilson, host of this very podcast. Companies like Adobe, Microsoft and Nvidia have partnered with us because they trust our expertise in educating the masses around generative.

AI to get ahead.

And some of the most innovative companies in the country hire us to help with their AI strategy and to train hundreds of their employees on how to use Gen AI. So whether you're looking for ChatGPT training for thousands or just need help building your front end AI strategy, you can partner with us too. Just like some of the biggest companies.

In the world do.

Go to your everydayai.com partner to get in contact with our team or you can just click on the partner section of our website. We'll help you stop running in those AI circles and help get your team ahead and build a straight path to ROI on gen.

That's a great question, Stuart. So there, there's a few things that AI can do in, in those situations. One is you can optimize the, the transportation routes and who's carrying what. Right? That, that to me is one of the easiest things is how do we make sure we maximize the impact of the shipping and the routes and, and you know, it's kind of like a traveling salesman type mentality is, is as you're going around and distributing things on ships, how do you optimize and reduce the route and reduce the carbon emissions and all of those required to get product from place to place? So there's a lot of tools out in the market. We've introduced a library called coopt, which is being used by quite a few companies that do a lot of distribution and transportation. And really it comes down to how do you optimize those routes? How do you put in as much information about the specific ship or truck or others to understand what features and functionalities that you're interested in and then really optimize that as far as shifting off of fossil fuels, right. It really comes down to the storage of, of, you know, energy, right. Batteries, hydrogen, other things that that ship is going to have to run off of. And really we're seeing a lot of development in those areas. Right. A lot of the large energy companies are investing significantly in hydrogen, right. And they're investing significantly in battery technologies and others. But you know, that's going to be a work in progress for a while. And so just like we talk about with data centers and other things is you can start off by, by optimizing what you have control over, right? And that's routes and packing and dropping off and picking up at the right locations. And then eventually you move to renewables as the next step in creating more green environments. And so that's what I'd suggest. I'm not sure that there's a short term solution for some of these big ships, but I tell you, the energy companies are all really focused on, you know, different ways of capturing carbon and, and then long term storage and reusing some of the carbon that's been captured off of, you know, fossil burning environments.

Yeah, no, that, that Mark, that was such a, such a great response, you know, because you, you, you laid out the kind of a two tiered approach on, on how AI can be used kind of in the shorter term and in the longer term. So great, great response there. We have one more here from Cecilia. So Cecilia, thanks for your question. So asking Mark, you know, we're saying she appreciates the need for balancing production centrally. Have you seen ways that AI is being used to begin democratizing or socializing use of renewable energy and underserved communities? So yeah, I'm interested in this, you know, as well, whether it's, you know, rural or, you know, other parts of the world, you know, maybe how can AI help in production in underserved communities?

Yeah, so it's a great question. And energy equity is a huge issue right now. Right. And obviously renewables and batteries and those things are not the cheapest things right now. Right. While they're really good and people want to adopt them, they're not necessarily affordable for everybody. And so we're working with a few different companies on getting a better understanding of how energy is being used in underserved communities. And how do we create an environment where people who have access to renewables can share that, whether they have rooftop solar or whether they have battery walls or EVs, how can that be used by people who maybe can't afford it? Right. And when you talk to regulators, there's the whole discussion around carrots and sticks. Right. And I'm not a regulatory guy or a lobby guy, but I've had to learn over the last couple years as we moved into, as I've moved into this regulated industry and really it's how do we incentivize people who have access to renewable energy and incentivize them to share it and basically create environments in which there are commercial reasons for people to make investments in these underserved communities by providing either low cost loans, by providing access so that we're in a situation where especially when the power grid does go out that you're created, you're able to create these micro grids or virtual grids and actually provide underserving communities with access to energy that's being produced locally. And so it's definitely something that every utility executive that I talk to talks about. And we've actually worked with some of them to apply for some government grants coming out of the IJA and others with a focus on creating applications and tools for underserved communities to measure and quantify how they use energy and how we can embed more distributed energy resources into their communities. Wow.

Such a, again, such a, such a deep and detailed answer. I love it. This, this is something I love about, you know, having this, this live dialogue is, is being able to ask, you know, people who are really leading the industry in these fields, you know, their, their feedback and input. So another great one because, yeah, we've, we've talked about a little of everything. But I'm wondering, Mark, and so is Monica specifically, you know, when we talk about AI and energy, maybe what's the most exciting project that you're currently working on that you're able to talk about?

Sure. So probably, you know, and I just mentioned it right now, regarding the energy equity. We're, we're working with a small, I won't call them a startup, but they're a small company. It's called Utilidata, and they've been in the industry for, for about 14 years. And we're working with them to create what, what we're calling a smart grid chip that will go actually out to the edge and sit on people's houses. Right. In the form of a meter. And think about your iPhone today and all of the capabilities that an iPhone has. Well, that's significantly different than today's smart meters. Today's smart meters are really good at two things. Making sure your bill's accurate and turning off your power if you don't pay your bill. Right. And so those are what we would call the metering aspects. But really the future of energy is how do we understand what's going on behind the meter and in front of the meter to create an environment where the consumer actually has more control of how they manage their power. And even to the point where they can load E contracts, they can see meter to meter communication so that if a branch is hitting a wire between two houses, we'd be able to identify a vegetation management issue and go get it addressed. Right. To understand who has power, who doesn't, and understand how do we share that power by incentivizing people to do things that they might not be able to do today because there's not enough technology at the edge. And it really requires that Real time capability. So that electrons move very fast. And you and I talked about, we've talked about Waze, right, and the ability to, you know, redirect traffic if there's an accident, but electrons move a lot quicker. And so if, if we have a failure, how do we redirect? How do we get energy back online? How do we give people control over how they consume their electricity? That, that, that's going to be great. And, and it's going to create an environment where we're going to have an open ecosystem similar to an Android store. Right. That basically allows everybody to create apps that can run on this and utilities can build their own apps, there will be third parties, but really it's going to be an exciting time to democratize the way in which people consume and supply energy moving forward.

Yeah, it's fascinating. That's a fascinating response because it does seem now, now that you say it out loud, that you know, hey, these smart meters, and I think we hear about these, but doesn't seem like they're at least now very smart. Right, right. If, if there's so little that they can actually do. But it seems that the potential impact of, of what you just talked about there, Mark, is, is pretty, is pretty monumental. You know, I have another question. So if, you know, when you kind of just talked about like one of the most exciting things that, that you're working on, but maybe, you know, specifically in the last, you know, year or two, you know, I know Nvidia is actually helping power, you know, this whole generative AI movement, you know, producing, you know, these chips that, you know, we all kind of need in these, you know, generative AI systems that we use, but specifically when it comes to energy, I mean, are there any, you know, I guess recent generative AI advancements that maybe have happened in the last year or two that are maybe going to change how we deal with energy in the future, you know, so I'm wondering if some of those developments that, you know, us everyday people experience in our day to day lives and in our jobs, you know, might be changing kind of how the way that we deal with energy.

That's, that's a, that's a complex question. Obviously, you know, I think the amount of energy that's going to be needed to solve a lot of the work that we're doing with generative AI is obviously going to increase. And I think people are going to be in a situation where they need to figure out the most cost effective way of consuming that energy, producing it, consuming it. And of Course, everybody has in the back of their mind sustainability and efficiency. So as we look to at all of those things. Right. You know, similar to the question about shipping. Right. It's a multi tiered approach. How do I become as efficient as possible on the energy that I do use? And then from there, how do I make sure that energy comes from renewable sources where possible, but when not possible, how do I make sure that the energy I'm consuming is developed in the most environmentally friendly way as possible? So our team is very diverse here at Nvidia and we go to market with a lot of third party partners everywhere from the oil and gas companies who are trying to drill less wells to be able to capture carbon at the point of drilling and look for methane leaks and methane detection and all of those things as they move natural gas through pipelines and use drones to identify corrosion and stuff to make sure that there's no environmental impacts, they're all using AI for that. And then the renewables companies that are trying to do everything they can to make renewables more cost effective, to make sure that they're reliable and resilient and tied into that, but then the consumers have to think about how do they consume power as effectively as possible. You and I talked briefly about data centers. Obviously generative AI and the amount of compute that's going to be required in the world moving forward is significant. We just participated in a summit a couple weeks ago. My CTO Ken talked about data center growth. And a lot of these companies are looking at 8% power growth year on year in the data center. And some are seeing spikes of 25%. Right. And you know, those data centers are working. Those data center companies work closely with us to potentially simulate, create digital twins of data centers. We work with a lot of companies that produce components for those data centers to make sure that they're as energy efficient as possible. But the biggest thing that draws energy within those data centers are the compute resources. And really it's going to come down to compute per watt. Right. And how much performance or outcomes do you get per watt of energy? And you know, that's a measure that Jensen, our CEO, takes very seriously and he continuously do focuses on how do we increase application performance to reduce the amount of watts required to solve a problem. It's not just about, okay, this CPU is 100 watts and this GPU is 200 watts or whatever. That's not the comparison. Because if you need 10 of those CPUs to do the work of one GPU, all of a sudden the Problem costs a lot more energy, right? And so how do we accelerate the adoption of accelerated computing? How do we create lower power GPUs, CPUs, data processing units, the whole impact of what's going in the data center, how do we make it as efficient as possible? And that's where software defined infrastructure is very important because as you see with some of the autonomous car companies, right, they come up with a better piece of software, they do an over the air update and you get 20 more miles to charge per charge. How do you create software stacks that when you upload them onto the systems, you can actually get 20% more output for the same infrastructure that you've already deployed at the same power envelope? So we continually look at ways in which we can increase the output per watt and we're committed to continuing down that track, both from a hardware perspective, silicon or hardware perspective, as well as software which tremendously impacts the performance per watt.

You know, Mark, we've, we've talked a little bit about everything which I love. We've talked about how AI can be used to help reduce the environmental impacts of oil and gas. We've talked about how AI can be used to increase production from renewables. But maybe as, as we wrap up the everyday AI show here, what is maybe that one take takeaway message that you want people to, to understand or to kind of glean from this conversation about how AI is kind of being used to, to power our future? What's that one message that you really want people to understand on, on how it is or how it could be used?

I would encourage everybody, especially those working in the energy industry, to start fast, right? Don't, you know, find a project, find, find a solution that you believe can be impacted by AI and try it, right? You know, I think there's a whole wait and see mentality, especially in the energy space. We tend to be a very conservative industry, right? People, nobody thinks an energy industry when their power's been running for, for, you know, six months with no interruptions, but the second it goes down, everybody's frustrated, right? You know, and what I would tell you is that it's an industry that's extremely conservative, but there's opportunities and I think there's things that we envision are going to happen. Start now, engage with the cloud providers, the Microsofts and the aws and the Googles, engage with the hardware providers, engage with Nvidia, right? Call me, reach out to me and let's talk about a big problem that you anticipate facing. And let's figure out if AI can solve that problem and let's try it and if it works, great. If it doesn't, let's move to another problem. But I think that hesitancy is we don't have time for it with the acceleration that we want with electric vehicles and everything else. We need to start now. And that's what I would encourage people to do. Start a project, try it.

Yeah, I love it. And you know, woozy, saying super interesting, great stuff. And you know, and, and we'll, we'll wrap with this. You know, Mark saying so or sorry, Brian Kennedy saying everyone is talking about AI being the end of the world. Thanks, Mark, for showing how it can also save it. So we talked about so much. If you missed a little bit, don't worry. Just make sure to go to your everydayai.com. you know, Mark, Mark dropped a lot of great resources, different programs that Nvidia is working on. We'll be sharing all that in the newsletter. So if you're not already signed up, make sure to do so. Mark, thank you so much for joining us on the show.

Thank you so much, Jordan.

All right, we hope to see you back again tomorrow and every day for more Everyday AI. Thanks, y'.

All. Thanks.

And that's a wrap for today's edition of Everyday AI. Thanks for joining us. If you enjoyed this episode, please subscribe and leave us a rating. It helps keep us going for a little more AI magic. Visit your everydayai.com and sign up to our daily newsletter so you don't get left behind. Go break some barriers and we'll see you next time.