Power, AI, and Infrastructure: Vincent Petit on the Next 1,000 Days
Insights
- US electricity demand is set to grow by roughly 1,000 terawatt-hours over the next decade, with AI accounting for 40 to 50% of that increase.
- The grid is significantly underutilized due to peak demand patterns, and demand-side AI solutions could reduce peak load by up to 60%, buying a decade of infrastructure runway.
- Energy resilience, AI leadership, and national security were once separate policy conversations. They are now the same conversation.
The US power grid was not built for the economy AI is creating. Vincent Petit, Senior Vice President of the Schneider Electric Research Institute, sat down with Jeff Kavanaugh, Head of the Infosys Knowledge Institute, at the Semafor World Economy Summit in Washington, D.C. to discuss what he calls a generational modernization cycle, one running entirely on electricity and accelerating faster than infrastructure investment has kept pace with. Petit argues the core bottleneck is not energy supply but grid architecture, and that the next three years will determine whether the US captures or misses the AI-driven modernization wave. The conversation covers demand-side AI optimization, reindustrialization pressures, geopolitical stakes, and why leaders need a long-term infrastructure vision rather than short-term operational fixes.
Vincent Petit:
It's fair to say that the world is not moving in an incremental way and that actually we are going to and that actually we are going to witness profound transformations of our economy. And the real driver of those transformations is AI. And that will have implications on the energy system. And when you look at what this means, and this is a topic that we've researched. We can do so many things remotely now. We can work remotely, we can shop remotely, we can entertain remotely. So all those things are now possible and they will grow in terms of use.
Jeff Kavanaugh:
Welcome to a series of conversations with leaders who are turning AI's promise into real business results. I'm Jeff Kavanaugh, and today we're joined by Vincent Petit, Senior Vice President of Schneider Research Institute at Schneider Electric. Welcome.
Vincent Petit:
Thank you for having me, Jeff.
Jeff Kavanaugh:
Your research points to a generational electrification and infrastructure cycle in the US. How is AI accelerating that shift, and why are the next 1,000 days so critical?
Vincent Petit:
I would not describe it as a generational cycle of electrification. Rather, I would describe it as a generational cycle of modernization, because what is really happening out there is not so much that electricity demand is growing out of thin air. It's that electricity demand is backing up significant modernization of our economy. So this is really what's going on. And in fact, electricity demand has been growing like clockwork for the last 80 years. It's been doubling every 20 years. In the US, we've been experiencing a pause of about 15 years since, let's say, the end of the 2000s. And now that pause is over, and we're back on track with modernization. The whole consequence of that is actually that we see electricity demand now growing about 1,000 terawatt-hour in the coming 10 years or so. And AI is a big component of that because, of course, AI is central to modernization. It's probably going to represent about 40 to 50% of that number. So significant, but it's not the only thing that's going on out there.
Jeff Kavanaugh:
You argue the real story is not supply, but how energy is used. How is AI and digitalization reshaping demand across manufacturing, mobility, and buildings?
Vincent Petit:
AI is central to modernization. So what I was explaining is that wave of modernization that we're into, and which is running on electricity. When you look at people building forecasts and scenarios and stuff, they tend to have a very incremental perspective of what's going on. But the reality is that the way we live today, Jeff, and what we are able to do today together, we were not necessarily able to do the same and we were not living exactly the same way 25 years ago. So it's fair to say that the world is not moving in an incremental way and that actually we are going to witness profound transformations of our economy. And the real driver of those transformations is AI. And that will have implications on the energy system. And when you look at what this means, and this is a topic that we've researched, think of a couple of things as examples, we can do so many things remotely now. We can work remotely, we can shop remotely, we can entertain remotely. So all those things are now possible and they will grow in terms of use.
Jeff Kavanaugh:
Just to toss a number out, I was with a leading energy CEO earlier today. He said they're going from one to two megawatt hours to forty. Actually, a million megawatt hours to forty is like a 20-fold increase in a few years. That was since the ChatGPT moment in the last four years.
Vincent Petit:
That's what it was. Absolutely. Remote activities now becoming the norm. Think of reindustrialization and advanced manufacturing, which is a big boost on energy demand in the US. Think of robots everywhere. Robots at home, robots at work, robots in the office. Think of drones and autonomous mobility. All those things basically are now enabled by AI. They are leading to a modernization of the...
Jeff Kavanaugh:
Transportation.
Vincent Petit:
US economy. Yeah, exactly. And what do they need to exist? They need electricity because all this modernization runs on electricity, not on anything else. So this is why we have such a growth now of electricity demand.
Jeff Kavanaugh:
As AI scales, energy and grid capacity are becoming strategic constraints. What are the biggest bottlenecks that can rate limit that growth?
Vincent Petit:
The US is undergoing a strong wave of modernization. That strong wave of modernization is due to AI, or whatever will be the applications of AI. But for that to work, we need AI to be supplied by electricity. And all the uses of AI that we will see across the economy to be supplied by electricity. So we need a strong electric backbone. We need a power infrastructure of the 21st century to be able to supply the 21st century economy. And we don't have that today in the US. So we are lagging on this. So the biggest bottleneck that we face is our power infrastructure.
Jeff Kavanaugh:
I think that it's coming home to roost. The delay in infrastructure modernization, we can't wait anymore. Your research outlines new legacy and new frontier. What determines whether the US remains incremental or achieves coordinated large-scale modernization?
Vincent Petit:
So precisely the infrastructure. So will we get the infrastructure up and running in time or not? And for that, I think it's interesting to say a couple of things. We need to build more of that infrastructure, more of that grid. And we need to do it relatively fast. Otherwise, we will not go into the exact same route. And I think everyone is very conscious and very concerned about this. But we don't need to work only harder. We can also work smarter. And I didn't steal that line from my kids. So, don't work only harder, work also smarter. So our grid is actually completely underutilized and there are ways, and very different from, for instance, what happens in China because the load profile is very different.
Jeff Kavanaugh:
We've been industrializing so much, it's 100 flat out.
Vincent Petit:
It's much more flat, whereas in the US it's much more driven by our building stock. So of course we have a big peak problem which explains why our grid is actually underutilized. Now we can actually flatten that curve. We can actually reduce that peak.
Jeff Kavanaugh:
AI can help.
Vincent Petit:
With AI, exactly. And this is precisely what our research was talking about because one of the things we found out is that if you deploy those solutions on the demand side, so in the buildings, in the industrial buildings and so on, you can reduce your peak by up to 60%. If you manage to do that across the entire building stock, you have solved the problem of the infrastructure for at least a decade. And that gives you, that buys you time to actually build out your future grid.
Jeff Kavanaugh:
How are geopolitics, reshoring and capital flows reshaping where and how energy infrastructure is built?
Vincent Petit:
OK, a couple of thoughts on that. The first one would be that geopolitics and reshoring and so on, if anything, they are adding up to the pressure on the power infrastructure. But first of all, you need more electricity locally. So you need your infrastructure. So you have more stress on the infrastructure. So that makes the problem actually even more acute. But again, we have plenty of ways to actually address it. So that's one thing. The second thing I would say is that if we want to compete in the modern world against other economies we need to lead that AI transformation, and in order to lead that AI transformation we need to lead the power system transformation. We need to build that power system infrastructure. So I think this is making the topic even more important. Energy resilience, AI leadership, national security used to be three different topics. Maybe now they are a little bit the same topic. So I think that's one thing that we need to keep in mind.
Vincent Petit:
The other thing I would say is that we don't need to be ideologists. We're going to need a lot of different solutions in order to get there. The US is actually home to vast resources of natural gas and we need that. There is globally a solar boom because it's very cheap and very easy to deploy. And we're going to need more of that. And by the way, some people talk about dependencies on solar and so on, but dependencies on renewable are very different from dependencies on fossil fuel. I mean, 90% of the cost of a solar installation is actually a local cost. So it has nothing to do with the imports. And by the way, when you deploy such technology, you actually buy yourself 30 years of independence. So there are many things that we can do, but the fundamental thing is that geopolitically, we need to actually accelerate on the deployment of the infrastructure. Without electricity that is abundant, resilient and affordable, we are just going to miss the train.
Jeff Kavanaugh:
All three are important. If the next thousand days are decisive, what are the critical decisions leaders must make now to avoid infrastructure gridlock and enable long-term growth?
Vincent Petit:
I would say that the solutions like the ones we've discussed, they actually exist. In the next thousand days, we actually need to deploy them.
Jeff Kavanaugh:
So that's three years.
Vincent Petit:
Yeah, that's three years to build resilience. And actually there are a lot who are doing it already, that have already started a few years ago. So, because it's a good hedge against potential resiliency issues coming up. That's pretty much the short answer. Now if I wanted to make a longer answer, Jeff, or at least to open up the discussion a bit, I would say that at the end of the day, leadership is also about guiding us to the future. That's what leadership is all about. It's guiding the team.
Jeff Kavanaugh:
It's not management, it's leadership.
Vincent Petit:
It's about guiding the team to the future. What's very interesting, we've been talking about AI. I mean, AI is very transformative, and what strikes me is that too often, at least in our Western world, we tend to be too short-term focused. So we lack the vision, we lack the plan. What we really expect from a leader is a vision and a plan.
Jeff Kavanaugh:
Vincent, thank you for your time. I'm Jeff Kavanaugh from the Infosys Knowledge Institute. Until next time, keep learning and keep sharing.