Powering the Technology Revolution / Davos 2025

Summary

This panel discussion at the World Economic Forum focused on the role of AI in the energy transition and its impact on global energy challenges. The panelists, representing various sectors of the energy and technology industries, explored how AI can contribute to more sustainable and efficient energy systems.

The discussion highlighted that while AI adoption is driving increased electricity demand, it also presents opportunities for optimizing energy production and consumption. Panelists emphasized that AI is not the sole driver of increased energy demand, with factors like electric vehicle adoption also playing significant roles. They noted that the energy sector’s inefficiencies could be addressed through AI and other technologies, potentially reducing infrastructure costs and improving grid management.

Innovations in computing hardware, such as more energy-efficient chips and liquid cooling systems, were highlighted as ways to mitigate AI’s energy consumption. The panel also discussed the potential of AI to accelerate research in areas like fusion energy and materials science, which could lead to breakthroughs in clean energy production.

The importance of regulatory reform in the energy sector was stressed, with panelists arguing that current regulations often incentivize inefficiency. They called for changes that would encourage the adoption of new technologies and more sustainable practices. The discussion also touched on the need for a balanced approach to energy transition, combining renewable sources with existing technologies like natural gas in the short term.

Panelists emphasized the importance of focusing on immediate, practical applications of AI in energy management, rather than solely on future possibilities. They highlighted the potential for AI to improve energy efficiency in both large-scale utilities and small renewable energy plants. The discussion concluded with reflections on the broader implications of AI for the workforce, emphasizing the need for education and reskilling to adapt to the rapidly changing technological landscape.

Keypoints

Major discussion points:

– The impact of AI on energy demand and the need to balance increased electricity usage with sustainability goals

– Innovations in energy efficiency, grid management, and clean power generation to support AI infrastructure

– The role of regulation and business models in driving or hindering adoption of new energy technologies

– The importance of focusing on near-term practical solutions while also planning for future breakthroughs

– The need for workforce training and education to adapt to AI-driven changes in the energy sector

Overall purpose:

The goal of this discussion was to explore how AI and energy technologies can work together to enable sustainable growth of AI capabilities while also accelerating the transition to clean energy and addressing climate change.

Tone:

The tone was generally optimistic and forward-looking, with panelists highlighting opportunities for innovation and progress. However, there were also notes of caution about hype and unrealistic expectations. The tone became more practical towards the end when discussing near-term priorities and workforce challenges.

Speakers

– Dan Murphy: CNBC’s Middle East correspondent based in Dubai

РAndr̩s Gluski: CEO of AES, a leader in sustainable energy infrastructure

– Greg Jackson: CEO of Octopus Energy, a company transforming the global energy market with technology solutions

– Antonio Neri: CEO of Hewlett-Packard Enterprise

– Uljan Sharka: CEO of iGenius, where AI is used to drive smarter decision-making in industries including energy

– Anne Bouverot: Special Envoy of the President of France to the AI Action Summit

Additional speakers:

– Laszlo Bodo: From Shell

– Roberto Patino: YGL from Venezuela

AI and Energy: Powering Sustainable Growth

This panel discussion at the World Economic Forum explored the intersection of artificial intelligence (AI) and energy, focusing on how these technologies can work together to enable sustainable growth while addressing climate change. The diverse panel of experts from the energy and technology sectors provided insights into the challenges and opportunities presented by AI’s increasing role in the energy transition.

Impact of AI on Energy Demand and Infrastructure

A central theme of the discussion was AI’s impact on energy demand and infrastructure. Andrés Gluski, CEO of AES, noted that while AI is driving increased electricity demand, it is not the sole factor contributing to rising energy needs. He characterized AI’s energy requirements as a “bottleneck” rather than an “Achilles heel” for increased AI adoption. Other panelists, including Antonio Neri of Hewlett-Packard Enterprise and Uljan Sharka of iGenius, agreed that AI is accelerating innovation in energy capacity and forcing change in the sector.

Greg Jackson, CEO of Octopus Energy, emphasized the need for regulatory reform to enable increased electricity supply, highlighting the potential for AI to optimize existing infrastructure. He pointed out the current inefficiency of the electricity system and how AI could significantly improve it. Anne Bouverot, Special Envoy of the President of France to the AI Action Summit, stressed the importance of efficiency gains in AI and computing to reduce overall energy needs, suggesting a potential shift away from the “bigger is better” paradigm in AI models.

Innovations in Energy and AI Technologies

The panel highlighted several technological advancements shaping the future of energy and AI. Gluski discussed progress in renewable energy and battery storage, while Neri detailed innovations in liquid cooling and other efficiency improvements in computing systems. He cited the example of the Lawrence Livermore Lab supercomputer, which demonstrates significant performance improvements with reduced energy consumption. Neri also mentioned the circular economy potential in data center cooling systems.

Sharka pointed out that AI supercomputers are delivering massive performance gains with less energy consumption, citing the example of Nvidia’s Blackwell chips. He suggested that AI could increase energy efficiency by more than 50% in the next three years.

Jackson provided a concrete example of AI application, describing how his company optimizes electric vehicle charging for over 220,000 vehicles in the UK. This practical use case demonstrated the immediate potential of AI in energy management, offering charging that is three times cheaper, seven times cheaper per mile than petrol/diesel, and the potential for 20,000 km of free driving.

Challenges and Opportunities in Energy Transition

The discussion addressed the challenges facing the energy transition. Gluski and Jackson highlighted regulatory and business model barriers to grid optimization, agreeing on the need for changes to incentivize efficiency in the energy sector.

Gluski also noted the importance of addressing human factors, including pushback on climate change issues. Bouverot emphasized the need for global conversations linking AI and energy/climate issues, announcing that France would host an AI Action Summit in three weeks to facilitate such discussions and explore potential solutions.

Neri advocated for focusing on near-term efficiency gains through basic automation and digitization, rather than waiting for revolutionary technologies. This pragmatic approach was generally supported by other panelists, who agreed on the importance of leveraging AI to improve existing infrastructure efficiency.

The panel briefly discussed nuclear energy’s potential role in the energy transition, acknowledging its complexities and challenges.

Workforce and Education Needs

The panel recognized the critical role of workforce development in the AI-driven energy transition. Bouverot stressed the importance of training and reskilling workers as jobs change with AI adoption. Neri argued that “everybody in this room will have to have a minor in AI” regardless of their field, highlighting the pervasive impact of AI across all sectors.

Sharka suggested focusing on system integration skills rather than massive infrastructure investments, indicating a shift towards software-based solutions in the energy sector.

Balancing Near-Term and Long-Term Goals

A key point of discussion was the balance between focusing on immediate, practical applications of AI in energy management and planning for future breakthroughs. While some panelists, like Neri, emphasized the importance of near-term efficiency gains, others, like Sharka, highlighted AI’s potential to accelerate long-term innovation in energy infrastructure.

This difference in perspective was evident in the debate around massive AI investments. Neri advocated for immediate efficiency gains through existing technologies, while Sharka critiqued announcements of large-scale AI investments, suggesting that plans may change as companies realize the speed of technological change.

Conclusion and Future Directions

The discussion concluded with a focus on actionable steps and areas for further research. Key takeaways included the need for regulatory reform, the importance of efficiency gains in both AI and energy systems, and the critical role of workforce training and education.

Audience questions addressed revolutionary AI applications in energy and workforce challenges, further highlighting the complexity of these issues.

Unresolved issues identified by the panel included balancing near-term efficiency gains with longer-term innovations, specific regulatory changes needed to incentivize grid optimization, and addressing skepticism around climate change impacts.

The panel suggested several areas for future research, including accelerating the adoption of existing technologies, exploring genuinely revolutionary AI applications in energy, addressing labor shortages in energy infrastructure deployment, and ensuring equitable access to AI-driven energy solutions globally.

Overall, the discussion painted a picture of AI as both a challenge and an opportunity for the energy sector. While AI’s energy demands present hurdles, the technology also offers significant potential to optimize energy systems, accelerate the clean energy transition, and drive innovation in sustainable practices. The panelists emphasized the need for a balanced, practical approach that leverages current technologies while preparing for future breakthroughs, all within a framework of updated regulations and a skilled, adaptable workforce.

Dan Murphy: ♫ ♫ Welcome to Red Bee Media’s Live Remote Broadcasting Service. I’m from CNBC, I’m CNBC’s Middle East correspondent based in Dubai and I’m thrilled you are all joining us for this next session here at the World Economic Forum. It is a privilege to have you sit in and learn from our expert panel of guests joining me for the next hour to talk about a topic that is of course absolutely critical in the world of energy. We’re going to talk about the role of technology and of course plenty of other issues impacting the energy transition. So let’s dive straight in. It is fair to say that we are at a pivotal moment in both the technology and energy landscapes, as cutting-edge innovations such as AI reshape the global economy. This next session is informed by the work of the World Economic Forum’s AI Transformation of Industries Initiative, which is part of the AI Governance Alliance. The white paper, released on Monday, is excellent and offers some very compelling insights, so please do check that out. We are seeing seismic shifts in the way we produce, distribute, and consume energy, and yet, as we push for a more sustainable energy future, there is a pressing question. How do we ensure that AI becomes a cornerstone of clean, efficient energy systems, rather than exacerbating global energy challenges? So this is a topic we’re going to unpack for you. Joining us are leaders from both technology and energy who are shaping how we power our world. I’m very pleased to say that we are joined by Andrzej Gluski. He’s the CEO of AES. It is a leader in sustainable energy infrastructure. We’re also joined by Greg Jackson. He’s the CEO of Octopus Energy, a company at the forefront of transforming the global energy market with its technology solutions. Antonio Neri is here. He’s the CEO of Hewlett-Packard Enterprise, which is transforming everything from cloud computing to energy management. Ulyan Shaka is the CEO of iGenius, where AI is being used to drive smarter decision-making in industries, including in energy. And Anne Boivreau is also here. She’s Special Envoy of the President of France to the AI Action Summit. My first question is this. Energy has been characterized by some as the Achilles heel of increased AI adoption. So my panel, would you agree with this? And what is the first thing that comes to mind when you think about AI’s impact on energy? Andrzej, you’re first. Sure.

Andrés Gluski: Well, I don’t like the term Achilles heel. That’s fatal. I think more of a bottleneck. So I think it’s something that needs to be addressed. So I think the technical solutions are there. It’s a question whether you have the right price signals for this occur. I think the other thing to realize is that. the impact of AI on energy demand is not uniform around the world. So people are facing different issues. I think it’s most acute in the U.S. at present, but realize that even in the U.S., if you look at the projected growth in demand for energy, the majority of that is not AI. It may be half of it. So I think you have to put that a little bit into context. So again, I think of it as a bottleneck, not an Achilles heel.

Dan Murphy: Very interesting. Greg?

Greg Jackson: First of all, I think, Andreas put it right there, that this isn’t the only driver of increased demand for electricity. I’d make a few observations. The first one, as an entrepreneur that entered the energy market from other industries, I’ve always been amazed that the energy sector doesn’t see these as huge opportunities, right? You know, if your job is to sell electrons and to transmit them, any opportunity to do more of that should be something we grasp with both hands. And I think it speaks to the need for regulatory reform. Now, we’ve got different regulations in different parts of the world, but broadly speaking, if you look at what are the barriers to deploying more electricity more cheaply, they often boil back to regulation, whether it be the inability to get permits to build stuff, or indeed the regulatory management, for example, the PUCs in the U.S. And I think a really fascinating thing here, though, is we already had this sort of background increase in demand in some countries as people get richer and buy air conditioning, and others as transport moves from petrol and diesel to, or gasoline, to electricity. But no one in the world has more ferocity than the hyperscalers. And I think that this is kind of like an unstoppable force meeting an enormous rock. And I hope that what it does do is create the sort of pressure for change in the world. the energy sector. Change in regulation and you know for good or for bad, the new administration in the US has talked about the need for this. The new government in the UK talks about it. So I think if the hyperscalers can put the pressure on that, enables that change to happen faster, not only will we be able to deal with AI, but we’ll also be able to accommodate all the other very important uses for electricity, which by the way largely contribute to decarbonisation. And the last thing I’ll add is that I think it creates the opportunity for real technology improvement, rapid technology improvement. You know in the UAE, Mazda and TACO just announced 24-7 clean power from a combination of solar and batteries, a gigawatt scale. Now you know if we can speed up, and at prices that are comparable with fossil fuels, even the Gulf where fossil fuels are cheap. If we can do that, if AI, the pressure from AI, enables us to scale this stuff faster, I think it’s gonna be good for everybody.

Antonio Neri: Antonio. Yeah, good morning everyone. So I like the context that has been said by my colleagues. First of all, I do believe that this is an amazing opportunity to accelerate the energy transition and the innovation around that transition. For enterprises, and you know, as I look at the room, probably 99% of the people are all enterprises here. It’s not gonna be an energy problem. It’s gonna be for you, because ultimately you get these models, they’re all pre-trained, and when you deploy them, the energy consumption is actually very minimal. In fact, if you look at us as a company, we provide large infrastructure to train this model, which consume an enormous amount of energy, but when you deploy in our own environment, once that model is trained, it’s actually a very small, finite amount of infrastructure. And that is not different than what we do today. But I look at this as an opportunity to accelerate innovation across the entire ecosystem. The source of energy that Greg just talked about it, the grid, the ability to transform the infrastructure and accelerate even the transition for the next generation of cloud computing. Today, when we deploy the systems, think about it, we deploy copper, gold, there is a lot of friction and heat being generated. Now we are throwing water out of the problem. But fundamentally, the next generation systems are gonna use different type of materials. You’re gonna use silicon photonics instead of transmitting electrons. As you talk about, we’re gonna transmit light. The cost is significantly lower, the performance is higher, and the energy consumption is very, very small. So I think for us, as I think about the societal challenges around transition to net zero, decarbonization, and acceleration of the transformation of the businesses through AI, it’s an ecosystem opportunity to drive that innovation and accelerate that innovation. But in terms of the question, I think for enterprises, it’s not that problem. For hyperscalers, absolutely it is, because they are in the business to train these models. And there is insatiable amount of cloud computing capacity with accelerators. But as I think about enterprises of the future,

Dan Murphy: I think we actually can accelerate that transition to net zero while we deploy AI through this innovation acceleration. Leon.

Uljan Sharka: I agree with Antonio. I think AI is going to be instrumental to force change and increase the speed of innovation in building up energy capacity around the world. But I think it is not going to be the main product that will consume that capacity. And we can already see this, which is different from what is being forecasted in terms of energy needs for AI. For example, the Nvidia recent GPUs named Blackwell chips. They consume 25 times less energy than the previous generation, and they produce 30 times more compute. So this happened in 12 months. Imagine what will happen in another 12 or 24 months if that process continues. So basically with AI accelerating the scientific process, so on one side helping with building energy differently and in a more efficient way, and on the other side having chips that are more efficient will actually lead to less energy than we are forecasting for AI. But I think it’s great that we are heading in that direction because who knows what else we are going to build with AI. We’re just at the tip of the iceberg. We’ll build new products and services that for sure will require more energy. So it’s great that we’re building that. Indeed.

Dan Murphy: Anne. Thank you.

Anne Bouverot: A lot has been said. And I agree with all of this. I don’t want to repeat it. I just want to comment on a word, I think, unsatiable or the fact that it’s this ever growing demand for energy. Yes, this is what we see now and for electricity. But I think we need to be careful that the bigger is better paradigm in AI is maybe not what we will see forever. Yes, of course, there is a race to build the biggest large language models and to win this race, particularly with key players in the US. But we also have to keep in mind, and you as leaders in business, that what will really make AI a reality are the applications. And specific applications of AI will be in specific industry sectors. And they will be based on smaller data sets and maybe not on the biggest is better model. So they will still need energy and hopefully clean electricity, carbon free electricity. But this bigger is better is the only route is something that we have. to be mindful of is maybe not the only future.

Dan Murphy: Indeed. And I wanted to pick up on what Anne and Ulian said here, because as you’ve both noted, the rapid adoption of AI and other emerging technologies is going to drive this increase in electricity demand. But a question for you, Andres, how are these technologies also unlocking an increased demand for clean power in particular? Sure.

Andrés Gluski: I think this is a good opportunity to sort of clear a little bit the hype. So right now there’s a lot of hype about nuclear. And nuclear is wonderful. And it’s concentrated, you get a lot of zero carbon energy. But really a big build out in the West is probably about a decade away. You know, you have to test out these small modular reactors. You have to know how much it costs to build them. You have to know what’s the permit. You have to have the people to run these. So over the next decade, a lot of this is going to be basically renewable energy. Renewable energy with batteries. If you look at what’s in the interconnection queue in the States, that’s about 90%. And gas is a much smaller percentage. So that’s what’s really going to power it, I think, in the next couple of years. And I think batteries are going to get a lot cheaper. We actually, I would say, invented grid scale lithium ion batteries about 15 years ago. And at the time it looked impossible. I’m just amazed that now in the States about I would say 30% of that queue is batteries. But this is going to have to be combined with gas. Because the batteries distribute the energy. They don’t really generate it. So gas is going to play an important role. So you have to look at the total carbon footprint and really see how you can lower that. But having said that, one thing that hasn’t come up, I think, is also the efficiency of use. So there’s tremendous amount of efficiency of use. I think it was mentioned about GPUs. Most of the energy is lost in heat. So there’s tremendous efficiencies that can be done. But again… It’s not just AI. Just think of the electrification of fleets. Just think of the reduction, for example, in emissions in cities. You know, the improvement in people’s health, the lowering of health costs. So all this comes together. So I think there’s been a little bit of hype and sort of like, it’s nuclear. You know, you’re not going to build renewables. No, we’re going to build a lot of renewables. We’re going to have a lot of batteries. What I do think we should also emphasize, it’s not the same problem for everybody. And the solution is not the same for everybody. So, you know, I was in telecom before I got into electricity. Just think of something like mobile cellular and what it did, for example, for Africa, for African businesses or Southeast Asia. And it wasn’t just like massive, let’s say donations from the rich people. It was just really making the technology available. So I’m very excited with a lot of these new technologies. I do agree with Greg that a lot of this is the business model. A lot of this is the regulations. But I think we shouldn’t lose sight of, you know, myself is what we want to make people capable of, you know, having the benefits of computing, having the benefits of transportation,

Dan Murphy: more than the electrons. So how do they use those electrons most efficiently? Can I follow up on something you said as well? I thought what you mentioned about nuclear is interesting because in some corners at the consumer level, and of course, at the political level, there’s still significant pushback on nuclear. Do you think artificial intelligence means we are at an inflection point for nuclear energy? Well, you know, AI can change many things. And it also will come up with new materials as has been mentioned here. And I’m very hopeful about that.

Andrés Gluski: I think, again, nuclear is going to be part of the solution. But I think we have to realise that… People have been saying this for 40 years, though. It’s true. Well, look, 40 years ago, there was an accident at Three Mile Island. Nobody died. However, there was a movie, The China Syndrome, with Jane Fonda. It stopped all nuclear construction in the States. Since then, we’ve built one new nuclear reactor, and it was a multiple times of the original budget. To my knowledge, most of the new nuclear reactors that have gone up in the West have been a multiple of their original budget. So, you know, we really have to be able to build these on time and on budget. We can do that, but it’s not just a technical issue, it’s a regulatory issue. If you look at the permits you need to build a new reactor in the States, if you printed it out, it’d probably fill this room. It’s just amazing. And so it is, I still think, a little bit premature to say that’s gonna be the only source of energy, you know, within five years. I think that’s, you know, getting ahead of ourselves.

Dan Murphy: There’s probably a lot of people who would also say that amount of regulation is necessary as well, but I think this is a really interesting talking point. Greg, bring it home for us, because at a consumer level, what would you say are some of the existing or potential opportunities that you’re seeing this technology revolution open for the energy consumers?

Greg Jackson: Yeah, I think, by the way, when we talk about AI, we’re all thinking about generative AI, because that’s what hit the headlines from November 23, 22, sorry. But you also got, you know, traditional AI and machine learning and, you know, a whole load of associated technologies. I cannot tell you how brutally inefficient our electricity system is. In the vast majority of countries, just before we came in, we were talking about sort of dynamic line rating. That’s basically, in most countries, houses have got smart meters, at least to some degree. And yet the vast majority of most of our grids and networks doesn’t have any real-time metering, any real-time monitoring. So you just kind of estimate how much current you’re gonna need to transmit, and then you build infrastructure designed to be able to handle the very peak. The peak might just be an hour or two per year, and it was just an estimate, and then you add on some margin for error. rather than deploying you know sensors smart meters whatever giving real-time telemetry in order that you could actually manage this system. Now the amount of data required to do that used to be considered quite a lot. Today yeah I mean it’s trivial with today’s computing capabilities especially including AI and that plays it this kind of so instead of just deploying some sensors that cost a few thousand dollars and some AI you build a billion dollars worth of transmission distribution. That’s the way our system currently works. Now we have to cut the price of energy everywhere especially in Europe but you know in order not just to be able to deliver these AI competitively but in order to be able to reduce the cost of living and to be able to electrify things like transport. And so the first thing we need to be doing is using large amounts of data to manage our systems much more effectively. Doing so you know could reduce the amount of pylons we need the amount of grid we need to build by 20 or 30 percent in some countries. But then on a device level we’re already optimizing the charging electric vehicles. We do it in multiple countries but the UK we do more of it. We’ve got about 220,000 electric vehicles in the UK where when you get home and plug in we just optimize the charging. Every car gets different schedule every day based upon forecast grid prices, forecast local load on the grid, 28 different parameters. It lets you charge a car for three times less than the average price electricity because you’re grabbing the cheapest times and because it acts as part of the energy system balancing the network. If the car, by the way, that’s seven times cheaper per mile than petrol or diesel. And then if we can discharge the car back into your household grid which is increasingly happening We give you 20,000 kilometers of free driving. This is all enabled by billions of data items per day to optimize those vehicles. But the price is enormous because you take the cost of going electric from being more expensive than petrol and diesel to dramatically cheaper. And we’re only just scratching the surface. And of course around the world many utilities worry about the intermittency, the variability of wind and solar. But through this kind of optimization, we can use as much of that power as possible at the time it’s been generated, freeing up the rest of the power system for baseload, for the industries and uses that can’t be shifted. In fact, some recent modeling I think suggested that if 20% of your system can be flexed like this, which requires huge amounts of AI, then moving to renewables is cheaper than a fossil fuel grid. So it will vary by geography. But the point being the price is enormous for consumers and for the climate.

Dan Murphy: And indeed for energy security as well, which is I think a really important point that you also just raised. Antonio, over to you for more on the technology side here. From the HPE perspective, what innovations exist today or are perhaps at the horizon here that are going to help AI infrastructure as Greg was talking about? Well, before I answer that question,

Antonio Neri: I wish what you described was available when I was in California, because I can tell you I had an electric vehicle and it cost me way more, a factor of 10 times more, to charge that vehicle than using petrol. And there was no optimization on any peak times and anything you’re talking about. So that’s also an implication, how we move to more green kind of transportation. But obviously that’s part of the problem, to make sure it’s not just clean energy, but how we optimize that clean energy. in energy usage across the grid. But on the other hand, when I think about your question about innovation, I think about innovation across the ecosystem, not just the cloud computing, right? So we at HPE focus on the entire ecosystem. We have an amazing entity, which is our HP Labs. They collaborate with governments and others. Thinking about the source of energy and how we transport that energy into these large systems. When you go to the system level, and we’re talking about efficiency of the system, HPE is focused on how we make sure we deliver the performance and the cost with lower energy to make sure it’s sustainable. So for example, in this room, we are now cooling this room with air, but I can guarantee you this glass of water is way more efficient than all the air we are pulling through this room. So when you talk about the system, think about when you burn your finger, where do you go? You blow or you go to the water and just trying to cool it down with the water. That’s what we’re doing today. Today, what we have done is moving away for throwing air, which basically blowing air to the system to cool down the system, using a lot of energies through these fans and using cold water, which is obviously more sustainable, the characteristics of the chemistry of the water is different than air, and basically reduce the energy by 90%. And one interesting fact, two weeks ago, we unveiled the largest supercomputer ever built is a HPE system, is the National Lab called Lawrence Livermore Lab in the United States. The Lawrence Livermore Lab used that system to do a lot of simulation AI for nuclear testing, obviously, because you can’t do nuclear tests anymore, obviously on the ground, and you need to manage that stockpile very well. And the other ones to do research of energy. So that system today can do 2.8 quintillion operations per second. It basically is 18 to the 10th, we call it exaflop. And that’s 27 times more performance than the previous system they had with 40% less energy consumption. So you can see the exponential growth in performance with a significant decrease of energy consumption. That innovation today is all about liquid cooling on every aspect. When you walk into the center, there is no noise, there is no fans, it’s just liquid going through the system. And that liquid also generates a circular economy because there are parts of the world we use hydropower. In fact, we are building data centers on top of the river so the water gets through the system. And then we also use the hot water to heat the buildings. So there’s a complete, in fact, in one of the locations, we use the hot water to melt the snow in the parking lot. Now, if we pipe the entire Davos here, we will not have accidents, people falling on ice and all of that. But that’s an example of how we think about the ecosystem of innovation across the environment. What excites me about the next innovations is getting rid of materials that also drive pollution, copper, zinc, obviously gold, and move into more a sustainable system using things like silicon photonics. And then also the changes in architecture. We have been living in an architecture and computing since the Turing’s day with you have a CPU and you have a memory, you have a storage, and basically collapse memory and storage and bring the CPU to the data. The most expensive part today is moving data around. When you have your iPhone, and sometimes you do a lot of things, it gets hot because it’s moving a lot of data back and forth, back and forth. Instead of moving data, let’s bring the compute to the data in a more sustainable way. So these are the things that excites us, but then also we need to bring it back to the source of where the energy is coming and how we. transport that energy. And that system in Lawrence Livermore also does another thing, which is the research on fusion energy. In fact, it’s next to an ignition facility. It’s an amazing facility. And with a hair, literally the size of a hair, we can now generate energy that’s 300 times more than the entire United States grid. 300 times. That’s a little thing that sits in a room that’s a football field. The problem is how we scale it and how we build it. And that’s where AI and supercomputers are necessary to scale these models. And so one serves the other purpose and eventually becomes an entire ecosystem that we bring

Dan Murphy: together, which I think will accelerate this transition to net zero and ultimately use technologies at AI. The scaling question is the challenging question because at the distribution point, there’s still clearly a lot of issues. We’re working with archaic technology. The majority of the world’s still on poles and wires, for example. So there’s a lot of work that needs to be done there. Ulian, I want you to unpack that a little bit further. When you look at your role and some of the work that you’re doing in this space, help me understand where you see other innovations for AI in the energy

Uljan Sharka: space. And can you give me some specific examples? Absolutely. I think one example that can add on what Antonio was mentioning about efficiency is what we’re doing with our AI supercomputer. We’re building it in Italy. It’s expected to go live this summer. And it is expected to deliver 115 quintillion operations per second with just 30 megawatts of energy. With the previous generation, that would have been near to one gigawatt of energy. Wow. And that is, of course, powered by liquid cooling and all of that. So the speed of innovation is just crazy. I mean, an AI unicorn like us, which is not a hyperscaler, has basically a chance through these innovations to elevate, in this case, European capacity by 100 times. with just about a billion dollars in investment. So this is on the hardware side. Talking about use cases in data and software, it’s worth to mention that energy and utilities is among, I would say, the best use case for AI because of the nature of transactional data among other industries like financial services. So before Gen AI was popular, the most successful use cases in AI in domains like predictive maintenance, operational intelligence, have actually been taking place in energy. So now with Gen AI, I think, of course, there is a ambition to leverage that AI readiness to do more, but we need to be mindful of the state of the art of the technology because there is a reality distortion on AI. We’re living in a hype environment and to make an analogy to explain that hype, we’re assuming that these AI models, which are going to definitely change the world so we know where this will bring us, but today we’re assuming they are flying cars. This is how we think of these AI models in our imagination. Well, I have been working eight years in this space and I can tell you that are more comparable to electric cars that have less autonomy than your smartphone. So this is where we stand today. So when we think of the use cases that Gen AI will unlock, we should be mindful of the state of the art and what we’re seeing, for example, is more practical use cases like getting the control room information and putting it in the pocket of engineers because through natural language now you can have a human talk to a machine so you can get that gold mine of information to become accessible on a par to everyone working on a plant. But I will conclude with one last example that I think gives an idea of the potential that we will see in the next few years. I think the inefficiency that was mentioned here by by colleagues is not just in large utilities and large energy manufacturers, but especially after the renewable energy boom driven also by the incentives where you had this massive scale of small renewable plants, both solar and wind distributed across the world. These are not professional energy manufacturers. They don’t have even the basic operating system to make these plants efficient. So imagine what will happen as AI democratizes intelligence, not just for the big organizations, but also for the small organizations. So I think in the next three years, we can potentially increase energy efficiency by more than 50% by not producing more, but by improving and making smarter what we already have. And can I get your perspective from a state level as well?

Dan Murphy: Walk me through some of the work that you’re doing with the French government here and what is actually being done on the ground to create an enabling environment for inclusive and sustainable AI adoption that supports the energy transition and a net zero future. Also, if you could put it into perspective for us at a timeline level, what’s happening today and what work is being done to ensure that this actually accelerates into the future and what this is going to look like from let’s say today, 25 out to 2030, for example. I have four hours.

Anne Bouverot: So, well, from the perspective of governments, the conversations on AI and energy used to be completely distinct conversations and the way to plan for energy development that was across much longer cycles. To give you an example, today in France, for reasons that have nothing to do with AI, electricity production is today 95% carbon free because there’s a lot of nuclear and then there’s a strong percentage of also wind and solar. This was not at all designed for AI, but it happens to be in the lucky position because then you can install data centers. It’s also an exporter of energy that are carbon free. If you’re not in this situation for a number of different reasons, we are at a stage where the increased usage of energy can flip the trajectories that you could have had for carbon impact reduction, for example. We’re seeing that with some of the hyperscalers who had made some commitments and now are getting out of these commitments. Because of the huge efficiency gains that have been mentioned and that are real, today you can build a model similar to GBT-4, for example, for really a fraction of the cost. And you have also increased efficiency in GPUs. You have all the innovation in energy production. So this will normalize and we will get into something that I’m confident will be back on the trajectories that we’ve committed to for net zero. But we might see a bump and we need to be careful about this bump. What is important for governments to do today is first to have this conversation on AI and the AI Action Summit that France is hosting three weeks from now is an international meeting of about 100 countries, CEOs, civil society, researchers to have this global conversation on AI. This has already started a little more than a year ago. There was such a conversation hosted by the UK. It’s also taking place in a number of other places. The global conversations on energy and climate, they have taken place. Now is the time, I think for the first time when they’re joining these two conversations. And it’s important that governments get informed by experts on the energy side at global level, the international. energy agency, for example, and also by experts on the AI side, and that we get the bright minds together to try and make these two things work together. More concretely, in Paris, at the AI Action Summit, some of the things that we will discuss and announce will be pledges for a sustainable development of AI. So it is for a development of AI, but how can we make it sustainable? Also challenges to try and stimulate more innovation in terms of efficient development, and then also some work with the International Energy Agency and some of the hyperscalers and data center builders to try and plan where in the world we should plan for energy for data centers. At the level of one country, one government can have a view, but if you’re saying should there be a hub in the Middle East, where should it be, what should be the amount of energy, when this can be done by bringing together both the AI and the energy expertise.

Dan Murphy: Well I hope this has been really insightful. What I wanted to do now is turn it over to you and give you the opportunity to ask some questions to our expert panel as well. If there’s anyone in the room who would like to make a comment or perhaps ask a question, now is your opportunity to do so. So please do not hold back. It’s the World Economic Forum, this is your moment, so please raise your hand if you have a question. I think one of the other questions that I wanted to ask, while we perhaps wait for some idea generation in the room, is on the issue of climate change, do you think that perhaps we’re focusing too much on this next frontier of technology, the next wave of innovation to address issues, rather than scaling up the technology that we have today to genuinely address this challenge. We’ve spoken a lot about what’s happening in the future, all of the innovation that’s taking place around technology, but what about what we can do right now to improve… efficiency across the grid, when it comes to electricity generation, power generation, and addressing the biggest challenge of our time, which is the rising temperature and issues in our climate. Who wants to take that question?

Andrés Gluski: I’ll take it. Please do. I think it’s an excellent question, because what we haven’t talked about is human adoption. So you can have a wonderful technology, but people don’t take it. You know, how are they going to use it? And I also like the analogy about the flying cars. You know, we have to make that electric car work efficiently. Well, first, we have a lot of technologies available. You know, Greg alluded to it, but basically, you hear a lot of talk about transmission. Transmission grids are probably used at 50% of capacity. Why? Because you build for the peak. So that means there’s a lot of copper sitting around most of the day. We have the technology between batteries and dynamic line rating and also energy efficiency programs to make use of that copper that’s sitting around, but much more and save billions. And then you say, technically, it’s almost a trivial issue, then why don’t we do it? Because the business model is that transmission companies make more as they invest more. They make a return on investment. So they don’t have an incentive to do this. So, I mean, that is very addressable, but it’s a human problem. You see, and when you talk about climate change, there’s a lot of people, let’s say, who are pushing back on climate change. It’s not clear what’s happening. We’re having more natural disasters. But that’s not really hitting, I think, the media in a big way. But if you really want to have the most clear-eyed view of this, look at the insurers. They’re stopping insurance in certain areas of the world. And the rates are going up. And they’re not doing that for ideological reasons. These are not tree huggers. These are very cool people who do actuarial studies. So I think that human element has to be part of it. And the technologists have to get them on board. And how this will get to that flying car future, we’ll see, as the famous U.S. baseball coach Yogi Berra said. He’s famous for malappropriatisms. He said, predictions are hard, especially about the future. So whatever we say here tonight, it’s not gonna exactly happen that way. Very interesting. But there are questions. Let’s take a question from the floor. In the front, if we could get a microphone. Also over here, so we’ll come over to you.

Dan Murphy: Brilliant, thank you.

Audience: So, my name is Laszlo Bodo from Shell, and my question is that the energy sector has been using digital technology to improve efficiency since the 1950s. Hewlett-Packard has been there from the beginning. So what I would like to ask you is that what are the things, like dynamic lighting, which is by all means useful, but it’s a continuous ongoing technological progress rather than a sudden tipping point? And what are the genuinely revolutionary new opportunities that AI unlocks in the energy space? Because I think there’s an important distinction between the two.

Greg Jackson: I think as humans, we have this tendency to excessively focus on the things that look like they’ll be revolutionary. It’s worth remembering that when the iPhone launched, its wireless connectivity was edge. It was the letter E. They didn’t wait for 5G. Now, in theory, some people knew that one day we would have hyper-fast, incredible connectivity, but we didn’t wait for that. And in fact, if we tried to go straight to that, we would never have got there because it would have been too expensive. No one would have believed that people would use the phone. No one would know how people would use this technology. So I think the history of technology really is that the first thing, the most progress comes from working with the bare minimum, but that starts revealing what’s possible next. And then you can get an investment for that. And you can go from edge to 3G to 4G and so on. And remember today, if you open your phone and get an edge, you literally turn it off. It’s faster to write a letter. And yet that was enough to change the system. we learnt at that time was around the world, I remember the hype about, some of you probably weren’t even born then, but around the year 2003 or something, about WAP, the Wireless Access Protocol I think it was, it was the mobile internet and it was so astonishingly bad that if you ever used it you would go well obviously the internet is never going to work on wireless right? And as a result all of the engineers in the world involved in building out mobile networks planned like mobile data growth would be like this and then the iPhone launched and mobile data growth went like this. They had to deliver in six months what they’d planned to do, well they’d never planned that much but ten years or more and you need to put pressure in a system to cause this kind of dislocation and force change. If things like the demand for electricity from AI forces change, that makes us use the existing boring technologies well, the greatest contribution of AI to the energy system for a while may actually just be that you’ve finally had someone who can push through rudimentary changes. Dynamic lane rating is not massively technologically complex and yet I don’t think it’s used at all in the UK. So the UK is currently embarking on building up what 50 billion, 70 billion of grid instead of being able to improve the capacity of some of our transmission lines by 40% just by sticking some sensors on and maybe software. Honestly it’s so unbelievably straightforward and so I think very often what we find is that the desire to have the perfect, the incredible solutions stops us to focus on what we need right now and what we need right now and back to your point actually is regulatory change. It is kind of crazy that the vast majority of the people that run our energy system, the companies, Their revenue is based on how much money they spend rather than save. It literally pays them to be inefficient, to build more copper rather than to use what we’ve got wisely. Changing that is hard. Those companies, understandably, won’t want to change that business model. Their investors will resist it. The energy system sector is exceptionally good at lobbying, but we need to force that change. Fortunately, big tech is also good at lobbying. When the two sets of lobbyists start colliding to get more out of our energy system, I hope that consumers will be the beneficiaries.

Andrés Gluski: Let me just give you a very specific example. We jointly developed with Google AI simulations of a whole ISO, a grid operator. For example, for Chile, it’s called Tapestry. This is a quantum leap in terms of how you would manage a grid and how you would build out a grid. That’s the one product I can say that’s completely different from what exists today. It’s not machine learning.

Antonio Neri: It really is AI and simulating potential futures. On the technology side of the house, if you think for a moment, there is a lot of discussion around AI, but through basic automation and digitization, you will gain enormous benefits before you even think about AI. On the other hand, you have to be aware of what this AI can and will do. If you think about it historically, when the telephone was invented, it took decades to reach 50 million people. I’m not talking about the mobile phone. I’m talking about the regular landline. We reached 50 million users in one month with ChargePT. Understanding there is a lot of tremendous gain here and now with the right regulation and legislation and incentives to gain massive efficiency improvement with conventional technologies that are available today, while at the same time start adopting these breakthrough technologies where it makes sense. in a sustainable, equitable way. And another example of this is quantum computing. People talk about quantum. The reality quantum is many, many years out. We are at best 100 qubits. We need at least 10,000 qubits to do some of the things that we need to do. And the manufacturing process and the scaling is not there. And so you have a lot of opportunity here now. He talked about what he can go do in the next few years versus thinking about, oh, let me wait for the next thing. But I think the regulation legislation has to be an incentive in managing the balance between CapEx and OpEx to get the best out of them.

Dan Murphy: I think we have time for one more question in the front here, sir.

Audience: My name is Roberto Patino. I’m a YGL from Venezuela. We’ve been hearing for the past weeks great announcements of massive investments on AI. Just three days ago, 500 billion in infrastructure in the US. Microsoft said 80 billion. So there’s abundance of capital. However, there’s scarcity of workers. We’re seeing labor shortages everywhere and a crackdown on migrants and undocumented workers. So I would love to hear your reflections in terms of how to square the circle. How are you, where you’re sitting, how are you going to be able to deploy all these new energy infrastructure and where are the bottlenecks with respect to the workers?

Uljan Sharka: I’ll try to take this one. So I think that type of announcements are going to change their plan in about six months once they understand that the future is old and that the speed of change that you will have to borrow in the meanwhile, because when you build such a big thing as a 500 billion supercomputer, it takes many years. So when you realize that buying so many chips and building so much energy, which will potentially get more efficient in the meanwhile, does not sound like a good plan. So I think it’s, of course, it makes a good media effect, but the reality is that we are going to get the main benefits from innovation in simple things like system integration. Because even if you invest 500 billion in a data center, but then software cannot leverage it because the data is fragmented, that’s useless. So we need to really remove the hype and the noise and focus on the things that can deliver value today. And I will conclude also with something on the use cases. I think the recurrent use case for innovation is going to be system integration and improving getting data on the same place.

Dan Murphy: And the revolutionary thing will be as simple as AI replacing that with a native AI operating system for data. And we will not need 500 billion to reach that goal. And I’ll give you the final word.

Anne Bouverot: Yeah, on the workers, maybe that’s a more general comment on AI and work. There’s also a fear that AI will displace all the jobs, except maybe the physical work. This is actually not something we’re seeing. We’re not seeing an increase in world unemployment. However, we’re seeing that AI will impact a lot of jobs. So some of them will maybe gradually disappear. A lot of them are new jobs being created, but the majority is that jobs are changing. So in terms of the arrival of AI and not just in the energy sector, what’s super important is to focus on training, skilling, re-skilling, thinking about the reorganization of jobs in energy, in health, in media, in every sector. Because this is happening very, very quickly, as you reminded, and will change the way we work. But in terms of jobs disappearing or. or not having enough workers, this is probably not something that will happen with AI.

Dan Murphy: It’s more, how do we adapt? And the pace of change is really, really fast. Really, really, really fascinating. To add to that, just to make sure. I know we’re out of time, so we’ll have to make it really quick.

Antonio Neri: No, very quickly. I mean, it’s the education system. I argue everybody in this room will have to have a minor in AI. Independently, if you wanna be a physicist, a doctor, or an accountant,

Dan Murphy: you’ve got to learn this technology. Absolutely, fair point, and an excellent way to end our conversation. Ladies and gentlemen, thank you so much for taking the time to be with us this morning. Thank you for your questions, and please thank my panel again. Thank you. Thank you. Thank you. Thank you.

Agreement Points

AI’s potential to improve energy efficiency

speakers

РAndr̩s Gluski
– Antonio Neri
– Greg Jackson
– Uljan Sharka

arguments

Advancements in renewable energy and battery storage

Liquid cooling and other efficiency improvements in computing systems

Potential for AI to improve efficiency of existing energy infrastructure

AI supercomputers delivering massive performance gains with less energy

summary

The speakers agree that AI has significant potential to improve energy efficiency across various aspects of the energy sector, from renewable energy and battery storage to computing systems and existing infrastructure.

Need for regulatory reform and business model changes

speakers

РAndr̩s Gluski
– Greg Jackson

arguments

Regulatory and business model barriers to grid optimization

Need for regulatory reform to enable increased electricity supply

summary

Both speakers emphasize the need for regulatory reform and changes to business models in the energy sector to enable more efficient use of resources and increased electricity supply.

Similar Viewpoints

Both speakers argue for focusing on immediate, practical improvements in efficiency and integration rather than waiting for or investing heavily in future revolutionary technologies.

speakers

– Antonio Neri
– Uljan Sharka

arguments

Focus on near-term efficiency gains vs. waiting for revolutionary technologies

Focus on system integration skills vs. massive infrastructure investments

Unexpected Consensus

Importance of education and workforce development for AI in energy

speakers

– Antonio Neri
– Anne Bouverot

arguments

Need for widespread AI education across disciplines

Importance of training and reskilling workers as jobs change with AI

explanation

Despite coming from different sectors (technology and policy), both speakers emphasize the critical need for education and workforce development to prepare for AI’s impact on the energy sector and beyond.

Overall Assessment

Summary

The main areas of agreement among speakers include the potential for AI to improve energy efficiency, the need for regulatory reform, and the importance of focusing on practical, near-term improvements. There is also consensus on the need for education and workforce development to adapt to AI’s impact.

Consensus level

The level of consensus among the speakers is moderately high, particularly on the potential benefits of AI in improving energy efficiency and the need for regulatory and business model changes. This consensus suggests a shared understanding of the challenges and opportunities in integrating AI into the energy sector, which could lead to more coordinated efforts in addressing these issues. However, there are some differences in emphasis and approach, indicating that while there is general agreement on the direction, the specific strategies may vary.

Different Viewpoints

Role of nuclear energy in addressing AI-driven energy demand

speakers

РAndr̩s Gluski
– Greg Jackson

arguments

Gluski argues that over the next decade, much of the energy build-out will be based on renewable energy with batteries.

Jackson suggests that the pressure from AI enables us to scale this stuff faster, I think it’s gonna be good for everybody.

summary

While Gluski emphasizes the near-term importance of renewable energy and batteries, Jackson sees AI as potentially accelerating the scaling of various energy sources, including nuclear.

Focus on revolutionary technologies vs. incremental improvements

speakers

– Antonio Neri
– Uljan Sharka

arguments

Neri emphasizes the importance of focusing on near-term efficiency gains through basic automation and digitization, rather than waiting for revolutionary technologies.

Sharka argues that AI is forcing change and increasing the speed of innovation in building energy capacity around the world. He suggests that AI will be instrumental in accelerating the development of energy infrastructure.

summary

Neri advocates for focusing on immediate efficiency gains through existing technologies, while Sharka emphasizes the transformative potential of AI in accelerating innovation in energy infrastructure.

Unexpected Differences

Approach to massive AI investments

speakers

– Antonio Neri
– Uljan Sharka

arguments

Neri emphasizes the importance of focusing on near-term efficiency gains through basic automation and digitization, rather than waiting for revolutionary technologies.

Sharka critiques announcements of massive AI investments, suggesting that these plans may change once companies realize the speed of technological change and the importance of software integration.

explanation

While both speakers are involved in technology and AI, their perspectives on large-scale AI investments differ unexpectedly. Neri advocates for immediate efficiency gains, while Sharka is more skeptical of massive infrastructure investments, emphasizing the importance of software integration.

Overall Assessment

summary

The main areas of disagreement revolve around the role of different energy sources in meeting AI-driven demand, the focus on revolutionary vs. incremental technological improvements, and the approach to large-scale AI investments.

difference_level

The level of disagreement among the speakers is moderate. While there are differing perspectives on specific strategies and priorities, there is a general consensus on the importance of improving energy efficiency and leveraging AI to address energy challenges. These differences in approach could lead to varied policy recommendations and investment strategies in the energy and AI sectors.

Partial Agreements

Both Jackson and Gluski agree on the need to improve efficiency in the energy sector, but they differ on the primary obstacles. Jackson focuses on the potential of AI and data to optimize existing infrastructure, while Gluski emphasizes the need for regulatory and business model changes to incentivize efficiency.

speakers

– Greg Jackson
РAndr̩s Gluski

arguments

Jackson argues that AI has the potential to significantly improve the efficiency of existing energy infrastructure. He suggests that using large amounts of data to manage energy systems more effectively could reduce the need for new infrastructure.

Gluski points out that there are regulatory and business model barriers to optimizing the power grid. He argues that current incentives for transmission companies don’t encourage efficiency.

Similar Viewpoints

Both speakers argue for focusing on immediate, practical improvements in efficiency and integration rather than waiting for or investing heavily in future revolutionary technologies.

speakers

– Antonio Neri
– Uljan Sharka

arguments

Focus on near-term efficiency gains vs. waiting for revolutionary technologies

Focus on system integration skills vs. massive infrastructure investments

Key Takeaways

AI is driving increased electricity demand, but is not the only factor contributing to rising energy needs

There are significant opportunities for AI to accelerate the energy transition and drive innovation in the sector

Regulatory reform and new business models are needed to enable increased electricity supply and grid optimization

Efficiency gains in AI and computing technologies are helping to reduce the energy intensity of these systems

Near-term focus should be on leveraging AI to improve efficiency of existing energy infrastructure, rather than waiting for revolutionary new technologies

Workforce training and education in AI skills across disciplines will be critical for the energy transition

Resolutions and Action Items

France to host an AI Action Summit in 3 weeks to facilitate global conversations linking AI and energy/climate issues

Develop pledges for sustainable development of AI at the upcoming AI Action Summit

Work with International Energy Agency and hyperscalers to plan global energy needs for data centers

Unresolved Issues

How to balance near-term efficiency gains from existing technologies with longer-term revolutionary innovations in energy and AI

Specific regulatory changes needed to incentivize grid optimization and efficient energy use

How to address pushback and skepticism around climate change impacts

Optimal approaches for integrating intermittent renewable energy sources into grids at large scale

Suggested Compromises

Focus on incremental improvements and adoption of existing technologies while simultaneously developing more advanced AI and energy solutions

Balance investments between large-scale infrastructure projects and smaller, targeted efficiency improvements enabled by AI and data

I don’t like the term Achilles heel. That’s fatal. I think more of a bottleneck. So I think it’s something that needs to be addressed.

speaker

Andrés Gluski

reason

This reframing shifted the perspective on AI’s energy challenges from an insurmountable problem to a solvable issue.

impact

It set a more optimistic tone for the discussion and encouraged participants to focus on solutions rather than obstacles.

If you look at what’s in the interconnection queue in the States, that’s about 90%. And gas is a much smaller percentage. So that’s what’s really going to power it, I think, in the next couple of years. And I think batteries are going to get a lot cheaper.

speaker

Andrés Gluski

reason

This insight provided a realistic view of the near-term energy landscape, challenging assumptions about nuclear power’s immediate role.

impact

It grounded the discussion in current realities and prompted more focus on renewable energy and battery technology.

We’re already optimizing the charging electric vehicles. We do it in multiple countries but the UK we do more of it. We’ve got about 220,000 electric vehicles in the UK where when you get home and plug in we just optimize the charging.

speaker

Greg Jackson

reason

This concrete example demonstrated how AI is already being applied to optimize energy use in real-world scenarios.

impact

It shifted the conversation from theoretical possibilities to practical applications, inspiring discussion of other potential use cases.

Today, what we have done is moving away for throwing air, which basically blowing air to the system to cool down the system, using a lot of energies through these fans and using cold water, which is obviously more sustainable, the characteristics of the chemistry of the water is different than air, and basically reduce the energy by 90%.

speaker

Antonio Neri

reason

This example highlighted tangible innovations in energy efficiency for computing systems.

impact

It broadened the discussion to include hardware innovations alongside software and AI applications, providing a more comprehensive view of energy efficiency efforts.

We’re assuming that these AI models, which are going to definitely change the world so we know where this will bring us, but today we’re assuming they are flying cars. This is how we think of these AI models in our imagination. Well, I have been working eight years in this space and I can tell you that are more comparable to electric cars that have less autonomy than your smartphone.

speaker

Uljan Sharka

reason

This analogy provided a reality check on the current state of AI technology, tempering expectations and hype.

impact

It encouraged a more grounded discussion of AI’s current capabilities and limitations, leading to more realistic assessments of its near-term impact on energy systems.

I argue everybody in this room will have to have a minor in AI. Independently, if you wanna be a physicist, a doctor, or an accountant, you’ve got to learn this technology.

speaker

Antonio Neri

reason

This statement emphasized the pervasive impact of AI across all sectors and the need for widespread education.

impact

It broadened the discussion beyond just energy and technology sectors, highlighting the societal implications of AI adoption.

Overall Assessment

These key comments shaped the discussion by grounding it in current realities while still exploring future possibilities. They shifted the conversation from theoretical concerns to practical applications and challenges, emphasizing the need for a balanced approach that considers both technological innovations and human factors. The discussion evolved from focusing solely on energy and AI to considering broader implications for workforce development, regulation, and societal adaptation to AI technologies.

How can we accelerate the adoption of existing technologies to address climate change more effectively?

speaker

Dan Murphy

explanation

This question addresses the need to focus on scaling up current technologies rather than solely relying on future innovations to tackle climate challenges.

What are the genuinely revolutionary new opportunities that AI unlocks in the energy space?

speaker

Laszlo Bodo (audience member)

explanation

This question seeks to distinguish between incremental technological progress and truly transformative AI applications in the energy sector.

How can we address the labor shortages and worker scarcity issues while deploying new energy infrastructure?

speaker

Roberto Patino (audience member)

explanation

This question highlights the need to balance massive AI investments with the practical challenges of finding skilled workers to implement these technologies.

How can we change regulatory frameworks to incentivize efficiency in energy systems?

speaker

Greg Jackson

explanation

This area for research focuses on aligning business models and regulations to promote more efficient use of existing energy infrastructure.

How can we accelerate the development and implementation of small modular nuclear reactors?

speaker

Andrés Gluski

explanation

This research area addresses the potential of nuclear energy in the clean energy transition and the challenges of bringing new nuclear technologies to market.

How can AI be used to optimize existing energy grids and infrastructure?

speaker

Greg Jackson

explanation

This research area explores the potential of AI to improve efficiency in current energy systems without requiring massive new infrastructure investments.

How can we better integrate AI and energy policy discussions at a global level?

speaker

Anne Bouverot

explanation

This research area focuses on the need for coordinated international efforts to address the intersection of AI development and energy challenges.

How can we ensure equitable access to AI-driven energy solutions across different regions and economies?

speaker

Andrés Gluski

explanation

This research area addresses the importance of making AI and energy innovations accessible to diverse global markets, not just developed economies.