Fireside chat with Andreas Schierenbeck, CEO of Hitachi Energy

[0:00] Welcome back. I'm Vijay Vaitheeswaran. I'm the Global Energy and Climate Innovation Editor of The Economist. We're live from the Global Energy and AI Conference. hosted by the International Energy Agency in Paris. I'm moderating a series of conversations with leaders from government, from the energy industry, from the tech industry and civil society. We're going to explore how AI will affect the energy system and the global economy in the years ahead. With me now is Andrea Schierenbeck, who is the CEO of Hitachi Energy.

[0:33] Welcome, Andreas. I know. And thank you for having me. So there's been a robust set of dialogues in the last couple of days, both technical, but also at the policy and practical level. A lot of this is focused on the AI part of the AI and energy conversation. But you come at this, of course, from the energy side, and I'm curious to know if we Take as a given, AI is this phenomenal force, an unstoppable force that is now expanding, well-financed, perhaps with a huge potential to do good, if done properly.

[1:09] Yeah. this unstoppable force is meeting an immovable object in some ways which is the power grid which is slow moving Historically, you may disagree, not as innovative as some other industries. Keeping the lights on was more important than trying risky new technologies for most utilities. What is the impact on the grid of having these massive new data centers, these hyperscalers putting up? data centers at a scale of a gigawatt, etc. Can you give us a little bit of flavor of what challenges you see as someone that needs to be there ready to support this expansion? Yeah, I think the timing could have been better for having AI on the horizon because in our industry already we are fighting with increased demand. It was stated that the electricity demand will double until 2030 and that's an industry which has traditionally not an ecosystem which constraints and supply constraints and nets and grid and production. Now with another thing, you know, just as someone example you have in Europe 900 gigawatts of projects for a new which are waiting for grid connection. Now on top we have Jenny I and these data centers are something what we haven't seen before. They very, very energy hungry.

[2:27] so you see data centers in a range of couple of hundred megawatts that's more than a small city up to five hundred megawatts or even bigger very political behavior so they start learning and using a lot of energy and then they're stopping suddenly and they're not using anything so you put on top of the constraints that we have volatile source on the demand side. and all that comes with different time constraints. So it's a very exciting and interesting time we are living in. So you make a very important point. A lot of attention is focused on the amount or the quantum of energy, but it's also the complexity of the timing, the cycling as it were, MR.

[3:03] places pressure on a grid which was already experiencing some of these challenges as we move towards more renewable power on the grid, which tends to be intermittent, of course, solar and wind. particularly And so we're finding that the grid has a multiplicity of demands. What are the sorts of technologies that you think are needed Just to start there, then we can move on to, let's say, policies or other things. What sorts of technologies do you think should be advanced quickly and what are you working on that might help deal with some of that flexibility volatility challenge i think we are at the beginning of that journey and we are looking for new solutions and from my point of view there are multiple possibilities first of all we shouldn't see this volatility not probably as a threat or complexity but maybe as a chance because if it would be possible to synchronize the two volatilities of renewables and the data center that would probably being being a good thing on the other hand you can look in battery storage So from my point of view, we have all the basic technologies what we need, we just have to

[4:09] fine-tune it now for the requirements. Now there has been tremendous enthusiasm about batteries, not just the traditional batteries but other forms of novel energy storage have written extensively in the economist about this of late um with the prices falling with the capacity expanding globally Even so, is it fair to say that we cannot scale up batteries fast enough for the grid or even in a distributed fashion, to deal with this challenge of intermittency, and in particular, the volatility imposed by data centers. We're going to need other tools as well. Well, let me answer it in two ways. I think for data centers, especially Gen A data centers, I think batteries can be one perfect solution. Because the learning process normally takes that long. Actually, that's the biggest problem. If it just takes 10 minutes, we're living in a 50-minute world in the utilities. That's already a big thing. So to mitigate that one with batteries is fine. But long-term storage, I think, even if the prices are falling, are still an issue.

[5:08] longer than that, I don't think that batteries are up to that. So, and I know there are companies that are working on longer-term storage, but that's a longer-term play. We're talking about challenges in the next one, two, three, five years, not a decade out. So we need some other... approaches to this and including one challenge that Fatih Birol, the executive director here, made this morning was for the data centers themselves to provide some flexibility, that is to themselves act as a flexible load The CEO of Electricité de France was here today suggesting that data centers may need to provide that given the penetration of renewables. And again, the challenge of variability. So we'll look forward to both the software and hardware solutions coming. But just to get very basic about it.

[5:51] and This... massive wave of demand is coming on top of a system that is underinvested. After years in the U.S., for example, of basically zero electricity load growth and not a lot of additional capacity investment You've talked a little bit about the challenge of grid congestion and the slow pace at which things are built out because of nimbyism perhaps, or regulation, and some of the challenges. how do we move forward more quickly to upgrade the grid with the technologies that are available, that are at the ready, given some of the challenges of, again, red tape.

[6:27] congestion, nimbyism. I think I would consider it from two different sides. On one side, the regulation, which has slowed things down, did it on purpose because you wanted to protect taxpayers' money. You build only what is needed. In a period of slow growth, that was a perfect set of constraints to do that that way. But now I think we have to act faster and we have to maybe adapt that. it takes too long and it makes us slow. So probably we can cut a lot of red tape, make it faster and of course cheaper.

[7:06] On the other hand, the capacity in the industry is constrained because after decades of no growth, you know what it means. It means that you cut capacity, which is not needed. Now we have to add capacity. But of course everybody is cautious. And even if we as Hitachi putting 1.5 billion into capacity expansion of transformer factories, this is all capacity which is already sold or promised. So we need visibility, framework agreements, standardization going forward to invest more. more capex on the ground but we want to see who is buying it well let's spend a moment on that because it's it's something most people would not understand there's a global shortage of transformers This is one of the basic building blocks of the grid. People get very excited about hyperscaling AI and all the applications we can use it for, but if you can't get the transformers in the ground to get the grid to get the power, You're not going to have the fancy AI on the advanced applications we're going to use. You're investing $1.5 billion, you just mentioned.

[8:03] How do you reassure your shareholders who may have been experiencing low growth in the past or been boom and bust cycles that you'll actually make your payback on this? That's what I said. Our investments are based on very sturdy and very conservative business cases. So the capacity we are adding is already sold or promised. We have capacity reservations, we have framework agreements, so we have customers signed it up. So market commitments from your customers. Exactly. But, of course, there is no room for more. You know, who haven't signed up is actually at the moment in a very interesting spot.

[8:34] Right, I see. Okay, so we're going to need to see more of that. Let's just spend a minute that we have left on policy. or other ways of dealing with as you talked about the way that things were regulated in a low growth or no growth environment made sense then. We're in a new era, clearly. We've heard the forecast today that electricity growth going forward may rise six-fold of the rate of energy growth overall. In other words, massive electrification is ongoing as a secular trend. What needs to change in terms of either regulation or voluntary industry accords, standardization, put forward one provocative idea that we can land on. I have a lot of provocative ideas in that. Pick one. One or two probably. I think one would be to allow the industry and the utilities to put capacity reservations and framework agreements forward even if they don't have an approved investment plan. Because at the moment they have to wait as the regulator is approving the investment plan, But if you know that you're needing five or ten transformers, make a reservation already now for what you need, probably need, in a very conservative way. At the moment, that's not allowed. Our customers are not able to do that. That would make a tremendous change in this whole thing. And on the other hand, doing permitting faster, maybe by reversing it instead of asking for a permit. And if you don't get an answer, you don't have a permit, just reverse and say, "Look, I'm not going to do it."

[10:02] where three months or six months process time. time to look as a permit if there nothing happening it's its okay to go for to summarize invest and build is what I'm hearing. Yeah. Great. Well, we're going to have to end it there as we're out of time. Thank you for joining us. This conversation, live from the IEA's Energy and AI Global Conference will be available for replay on the IEA's website. Up next, I'll be speaking with Jeffrey Olson, head of the power commercial team at Kairos Power. We hope you'll join us then. Thank you.

[10:33] ♪ ♪ you you ♪ you you you you ♪ ♪ you you you