Schneider Electric Coalition Spotlight

Welcome to the stage. So,

we're going to talk a bit about data center energy demand in our conversation. So, Frederick, thank you very much. You are executive Vice President, energy Management at Schneider Electric. And oftentimes when we talk about how we're going to meet all of this energy demand coming from AI data centers, always humming, always requiring cooling.

There's this conversation about increasing baseload power and just meeting power demand as opposed to one that has to do with this distinction between peak demand at certain times of day and certain locations. So, just to set the stage for everybody, so we're all on the same page. Um, can you talk to us a little bit about why the distinction between peak demand and meeting total energy supply is an important one to make?

Yeah, for sure. and thanks for having me here. Uh the the the thing to understand is that a data center is a concentration of electric load at a given point in geography, right?

And in order to power this significant amount of energy, you need first to look at your energy availability, you need to look at your grid, eventually making sure that you have the right supply of energy available at this point of time. But this is not enough because energy exist, electricity exist and during the day, you have most of the time abundant energy, but you need as well to look at the peak, right?

Because what the problem can be is at the peak where eventually the supply of energy may not be enough for one given data center. And that's that's the two side of the things you have to look at. Reinforce the grid, make sure at this point you have enough power and then look at the curve during the day.

Now, Schneider Electric has been looking at flexible, flexible demand and energy efficiency

long before today and this conversation around AI I demand. So, can you talk about well, why flexible demand is so important to ease constraints that new energy additions cannot handle alone?

Yeah, and and that will come to the point of the peak. How you manage the peak. You have two way to decrease the peak. The first one, which is the most easy is energy efficiency.

The more you are energy efficient, and here you're looking about the whole electricity consumption of a given places, right? If you're more efficient, you reduce your peak. The second thing you can do with flexible with digital technology is managing a transfer of load out of the peak to other part during the day. and here AI is part of the solution.

So AI can help you to better manage this energy fluctuation and especially in multiple applications like homes, building, where you have natural flexibility with HVAc with heating system that are by design the load can be shifted.

And there's an element to this that is so critical in the moment because time to power and time to get things online is so important. And so doing more with the energy that we have enables us to do things more quickly. Um, and data center investors are increasingly also acknowledging this and looking at efficiency as a part of their answer. Where should they start and can you outline kind of the path forward and the steps they should be taking?

Yeah, so the first thing you need to look at is of course the ability the availability of the site you have chosen to go along the journey of how much power you need and how much compute capability you need and remember, time to power equal time to token. So that's the first point to look at. The second point is looking integrally at your compute system, your power system and your cooling system.

And the third thing that matters is simulate before you operate. And today we have tools that help you to design, simulate, optimize your design so that when you implement it and then the time become extremely short because you want to have things very fast on time, when you implement it, you have simulated everything and you can simulate your efficiency when you will operate the site. That's super important.

The simulate before you operate. I really like that thing. Um, so

what's stopping us? What are the barriers to adoption when it comes to implementing these technology changes?

Look, probably what of the one of the big barrier today is people have a tendency to look at things in silos. Like I look my power, I look my cooling, I look my compute and everybody is not much speaking to each other. What we call energy intelligence or energy technology that drive intelligence is really connecting the dots between those systems so that at any point of time, you can have the visibility of what is the real need of power for the compute, for the cooling and how you optimize that permanently.

And by doing so, digital technologies are helping a lot and AI is part of the solution. So data center are here to support AI, but AI can help data center to be much more efficient.

Oh, okay. So the data centers need AI that they're actively generating in order to to keep going. So for those leaders who are actively investing in AI, what decisions should they make right now to keep from energy really constraining them as they go forward because these timelines are moving very quickly and, you know, some of the things I hear about putting new CCGT plans online, sometimes, you know, we're talking years, not months.

Yeah, sure. So all what I explain on flexibility is is giving by the way available power today, okay? Of course, uh we need more electricity for the future because all new source of technology are using electricity. We speak about AI, we could speak about robotics, mobility, all will use more electricity. So there is a need for more power for sure, but there is energy available in short term in most of the time during the day.

So what people should look at is of course their power availability, their capacity to be flexible if at one point of time, they need to support the grid, they need to become actor of the grid and of course, simulate everything so that they can on the long run optimize their installations.

Simulate before you operate. So, uh Frederic, thank you so much for sharing your views and the long experience that Schneider Electric has in flexible demand and creating, you know, the supply required in order to keep all of this exciting technology moving forward.

And thank you for having me.

Thank you.