Loading summary
Dario Gill
So good, so good, so good.
Nordstrom Rack Announcer
New markdowns are on at your Nordstrom Rack store. Save even more. Up to 70% on dresses, tops, boots and handbags to give and get cause.
Stephen Overlee
I always find something amazing.
Dario Gill
Just so many good brands.
Nordstrom Rack Announcer
I get an extra 5% off with my Nordstrom credit card total queen treatment. Join the Nordy Club at Nordstrom Rack to unlock our best deals. Big gifts, big perks. That's why you rack.
Stephen Overlee
Hey, welcome back to Politico Tech. I'm your host, Stephen Overlee, and on this show, I break down tech politics and policy with the people shaping our digital future for a tech podcast. We've spent a lot of time lately talking about energy, and that's because one of the big questions hanging over artificial intelligence and really all promising technologies of the future is how the US Is going to power them. This week, I was invited to the Aspen Cyber Summit in Washington to interview Dario Gill, the undersecretary for science at the Department of Energy. He's one of the leading officials in the Trump administration, grappling with this question and overseeing the research and development of future energy sources like fusion. On the show today, here's my onstage conversation with Undersecretary Gill.
Hello, everyone. I'm Stephen Overle, the host of the Politico Tech podcast. And we have a very important conversation ahead of us about how we're going to power all these technologies of the future, from AI to quantum. I know that's a question top of mind for our guest here today, Dariel Gill, the undersecretary of science at the Energy Department. Thank you for being here today.
Dario Gill
It's great to be here.
Stephen Overlee
It's good to connect with you again. I've spoken with you in the past, your previous role as director of research at IBM. I know you are very focused on the potential for AI and for quantum. In your new role, what do you see as the biggest energy challenge to be solved?
Dario Gill
Well, it depends on what horizon we're looking at. The way I tend to look at investments in science and technology is you have to construct a portfolio that spans different options in terms of disciplines and areas. And that's true for energy and then different horizons with different appetites for risk in the very near term. I mean, everybody knows that we're in a unique circumstance, that for many decades we had seen electricity demand either flat or slightly declining, in part driven by the computing revolution that we're witnessing right now. We see forecasts in terms of an appetite for growth, that the electricity grid may need 100 gigawatts of additional power being put into the system. Now we're seeing computing environments and data centers that are in the gigawatts and appetite to do things that are even like 10 gigawatts etc. So in the near term it is about with the existing grid projects that are also being implemented behind the grid power, how do we enable that with the technology that we have right now? So that's like immediate. And I'll tell you, the rate limiting step on the progress on AI is energy at this point. So that presents a set of parts policy constraints, regulatory expediting approvals. What can we use with existing energy technologies? If we look beyond, we are seeing for example on nuclear, a renaissance on nuclear driven by computing. So that presents yet another horizon of what we can do. And then beyond that we're very bullish and we have a lot of excitement of what we can do. Also with technologies like fusion, depending on horizons we can look and we are at the department investing in all of those horizons and at the crux of the intersection between computing and energy.
Stephen Overlee
And so where are you starting with working all these different horizons? I mean if I were to look at your to do list, what's number one?
Dario Gill
Well, for me in the, you know, if I look at it from a science perspective, I look at, you know, the horizons that are a little bit, you know, beyond. So for example, we are putting a lot of energy into fusion and we released recently a roadmap for how are we going to decide to share investments that we have to do from a federal government perspective to complement with the $10 billion of venture capital that have been raised for fusion. But at the department it is focused on like very much also near term of what do we need to do to make sure that we win in AI and we meet the needs of re industrialization in America and in AI. So I think there's a division of labor. But if you're asking me specifically, given my role in looking forward in the science agenda, I would say that from an energy perspective, fusion is an area that we're putting a lot of energy into.
Stephen Overlee
I was going to ask if you're a fusion guy, because when you talk to or you hear a lot of these tech leaders speak, Sam Altman or Jeff Bezos, there is a lot of enthusiasm for fusion and yet frankly, I sometimes struggle to differentiate hype from reality. How do you differentiate that and what's your pitch for why fusion is in fact the future?
Dario Gill
Yeah, well, I think that that is common by the way, not just on fusion, right? In AI, in quantum and so on. We experience these portfolio of opinions, expectations and so on. But I am on fusion. I am optimistic having met with a lot of the leaders that are driving these new ventures, obviously with my own team that is driving the core investments in fusion science that I do think we are in a different phase. Obviously there's been decades of work on supporting the basic science of fusion, but something different is happening in the sense that facilities are being constructed and test environments are being constructed and. And then we have multiple teams in the United States in pursuit of q greater than 1 demonstrations in terms of sustaining a burning plasma. We've seen very clear advancements, for example on magnetic confinement, fusion of novel materials, of superconducting tape that allows us to do confinement with a much higher magnetic fields. There's evidence that there is both investment team composition and demonstration events that are going to unfold in the coming years that I think are going to change the perception of the rate of progress that we can do. Also for the first time now, we can debate on the dates, but people are building towards objective of having a fusion pilot power plant. And that's where it depends on the level of ambition people have. But we're talking about, you know, late, later this decade or you know, 2033, 2035. So we're looking at horizons now where the rubber is going to be the road, right? And we're going to be able to see whether after decades we can actually start producing electrons, you know, with control fusion plans. And if that happen, I mean, we should underestimate how amazing that is, right? The fact that we are able to sort of, you know, use a basic mechanism that powers the stars and be able to build a technology here on Earth to do that, I think is one of the pinnacles of like scientific and technological achievement. And we are going to witness this. This is not something that we're going to talk about like generations from now. So I am bullish in that sense, but realistic of how difficult it is technically.
Stephen Overlee
I mean, talking about these long term horizons, you know, I can only think we talk so much these days about the energy demands now, right? What we're going to see in the next couple of years because of AI and then also because of quantum. And you hear folks like Elon Musk say in the media term renewable energy is future and we're going to need solar and wind to meet our energy needs. Is that how you see it? Especially in light? Obviously we've seen cuts, you know, to renewable energy projects and Renewable energy staffing. Is that how you see it? Is renewable part of the mix in your opinion?
Dario Gill
I mean, renewables have a, you know, they have a role and a place, you know, as part of the overall energy mix. But when we're talking about these kind of computing centers and so on, one of the things that is very price is density, right? Because you are having these power parks and the ability to do, I mean, go back to nuclear. Why, why is there an appetite to build nuclear plants again in America? It is because of the density that it can provide and the reliability, you know, and the cost structure and the long term sort of reliability that you can enable for these computing data centers is why that is important. So I think everything has trade offs and you just have to optimize depending on what kind of system that you're building. But in this giant computing data centers and so on, the role that very energy dense solutions that are very reliable and work 24 7, it is what the market is pricing.
Stephen Overlee
You know, with all things tech and cyber, China looms very large. Right. And I'm curious, you know, the US is competing with China on AI and on quantum, you know, the development of the technology on chips, on talent in the energy space. I mean, what do you see as sort of the, you know, main point of competition with China and in your view, is the U.S. ahead?
Dario Gill
Well, that's a complex question. I mean, without a doubt it is our other pacing adversary on these technologies. Let me just parse the question on a couple of fronts. On the technology front itself, it is an absolute imperative that the United States remains in the lead and actually accelerate our competitiveness on AI and quantum and high performance computing technologies. And let me just motivate why. We are in the midst of a computing revolution and this computing revolution is going to fundamentally transform how we do science and technology in R and D. We spend a trillion dollars a year in R and D as a country annually. And that investment is the underpinning of our economic competitiveness and our national security, our innovation capacity. And I am convinced that we should set ourselves a goal to double the productivity and impact of America's R and D driven by this computing revolution within a decade. If we do not succeed, like I'll put it now in the negative, if we do not succeed and we're able to problem solve at a multiple of the speed that we can do today, and China does that because they've adopted and changed the methodology with which science and technology and R and D is conducted, we're going to be Very sorry for ourselves. We're not going to let that happen and we have to do that. So the stakes are very high because this is not just business as usual. A fundamental methodological approach or how the workflow of science and technology is being conducted is underway. The energy part of the equation is that it goes hand in hand. A world in which we see energy addition and not subtraction, a world in which we demand that we need to be able to create more energy to actually solve our problems faster than before and drive these competitive advantages is a world that China has a lot of clarity on too. They're not envisioning a world with less energy for themselves. Both of them. The administration's energy dominance agenda and and our science and technology agenda, driven of harnessing and advancing the computing revolution to transform how we do R and D in science and technology are the two companion pieces of the strategy.
Spanish Advertiser
Una silla de masajes puede pares er extravagante ocho configuraciones differentes intensidad justable. Como los acientos de la parese extravagante.
Stephen Overlee
What's the balance then between safety and speed? Because it does seem like right now when it comes especially to AI, we've really taken our foot off the brake. It's really go as fast as you can on development again under this idea that we're in a race to win. There are a lot of concerns people have about cybersecurity, about privacy and the potential negative effects of this technology. How do you find the balance?
Dario Gill
We spend a lot of time dealing with this and I know everybody in the audience does as well. I mean cyber is an example, but I mean I could name the example of non proliferation too and the implications of advanced AI technology for bioweapon design and proliferation or nuclear proliferation and so on. So these are areas that we take incredibly serious and there has to be a very, very focused effort on equipping our teams that deal with these problems with the expertise and the tools to be able to keep pace with the threat looking. Cybersecurity is a good example around that the rate at which AI driven cybersecurity attacks and things can happen can only be countered through technology itself. There's many other dimensions in addition, I'm not touching on the panel was properly addressing all the cultural and human capital and education and ele. But like for me as a technologist, just looking at the technology, it's like it goes hand in hand. And I just am a realist on the rate of progress, of what happens and for Good or bad, we are in an era where the progress of what is happening in computing and these areas is accelerating. It's not decelerating. And therefore we need to have the commitment to make sure we're leading in that context, but then create enough institutional capacity and talent to divide the nature of problems. Who are the people focused on accelerating it for science? Who are the people focused on doing this for defense? Who are the people who are doing this to protect us against the wrong use? I don't think we can afford to say, well, let's all sit down and wait until we figure everything out. I think what I try to encourage everybody is to raise their knowledge and their abilities and their institutional capacity for all of us to do our jobs. Thankfully, in the United States we have a lot of institutional diversity, a lot of people that care, different stakeholders around that. But it's not appropriate to say, well, until I get up to speed and come up with what's going on, I want everybody to stop making progress. So it's that balance around demanding every institution to race to the challenge and do their job.
Stephen Overlee
And so practically speaking, in your role now, when you're again, as you were saying, kind of looking at these longer term horizons, right. Are you thinking about cybersecurity in the early stages of R and D? Is that a problem that will sort of industry can deal with down the road or some other agency can deal with down the road?
Dario Gill
Oh, we absolutely focus on day one. Look, the Department of Energy has three core missions. We have the mission of advancing energy technologies and improving the quality of the energy system in the United States. We have the mission of discovery science. I mean, the 17 national laboratories of the United States under the Department of Energy and under the Parliamentary of Energy, you have the National Nuclear Security Administration that is the ultimate guarantor of our sovereignty with our nuclear weapons. So as you can imagine, with a portfolio like that, particularly the latter, security and a managed security environment is an integral way in which we operate our missions. The Department of Energy has its own intelligence agency and we take incredible amount of pride and focus on making sure we're protecting some of the most important secrets that the nation has. And by the way, that's not just in the realm of the world of nuclear weapons. In our core science and technology mission, we know that those are very sensitive and important technologies and a managed security environment is incredibly important. So I can guarantee you that every single initiative that we launch from day one, the research, security and the protection of, from a personnel perspective, from an infrastructure, from a Data perspective. It is baked in the design of everything that we do.
Stephen Overlee
I want to ask you a research question because I know you're a research guy. I mean the federal government for decades has sort of been this engine of research, right? Especially when it comes to foundational research. Lately under this administration we've seen a lot of cuts to research, whether that's National Science foundation, nih, even at the DOE Office of Science. What do you see as the role going forward for government? Is it going to continue to be a driver of innovation as we've seen in past technological waves, or do we need to just look to industry for that?
Dario Gill
Yeah, for sure. The federal government provides an indispensable and irreplaceable foundation in the science and technology ecosystem of the country. But let's analyze the current situation and how much it contrasts to the architecture that was laid out by Banner Bar Bush following World War II when Banner Ver Bush published a very famous science the Endless Frontier memo. At the end of the war, the country had mobilized the best scientists and engineers of the nation in the context of the Manhattan Project and what happened at the MIT Radiation lab, et cetera and Bannon Bush asked a very fundamental question, which is how should we design the institutional framework of supporting science and technology in times of peace? We won the war. What do we do now? And there was a piece of legislation that led to the creation of the National Science foundation. And through a series of crises there were systematic creation of new science related science and technology related institutions during, you know, after Sputnik, right. The creation of DARPA and you know, of course NASA had been created. We saw in the 1970s the creation and the energy embargo, the Department of Energy, et cetera. So fast forward to today. So to today. And so at the time it was all mostly federally driven. Right. Sense and technology. Fast forward to today. We have built, thanks to the success of that ecosystem, a trillion dollar R and D ecosystem. Just think about the scale of it. A trillion dollars. 70% of that funding comes from the private sector. Today, today, to answer the question of our capacity as a nation to innovate and to do science and technology, we got to sum what the federal government can do, a couple hundred billion dollars a year that we invest in science and technology, what states do, what universities and philanthropy do and what the private sector does. When I was at the National Science Board where I was the chair before I took on this new role, one of the things we published and we look very carefully is to analyze this change ecosystem and this Change landscape. The core thesis that we're putting forth is to do anything of great ambition in the nation in the critical technology areas that we care the most about, semiconductors and AI and quantum and biotechnology and space. It is impossible for us to succeed if we do not take the approach of what the different sectors have to do and how we coordinate and make investments together. I think it's very important that we look at the technology roadmaps, the outcomes that we want to achieve in this area and honestly ask the question, are we in the lead as not generically, are we investing? I mean, I give you an example. To answer your question very precisely on Quantum, when I was going through my confirmation hearings and I remember having discussions with senators where I was holding myself accountable to say, we are going to deliver fault tolerant quantum computers and quantum supercomputers ahead of anybody else in the world. We're going to deliver that outcome. That's an outcome. If I went and I said, look, the way we're going to lead in Quantum is instead of five national quantum outcomes centers, we're going to have 10. Instead of spending 2 billion, let's spend 4 billion. Instead of having 5,000 students, let's have 10,000. Those are all good things. They may be good things to do or not, but that's not an outcome. An outcome is delivering the capability that the nation needs to make sure that we are successful in our economic and national security fronts and so on. Because if five years from now we doubled the number of centers, we did all of that stuff, but we didn't deliver the good, I don't think anybody is going to sit around here and says, you guys did a great job. So I think that the clarity and the accountability and what are the stakeholders that need to come together to deliver on that outcome is what we're focused on. And there are reforms that need to be done, there are noble partnerships that need to be struck. But it's not as simple as saying, oh, it's just the private sector will do it or the thing. No, everybody's going to have a role, but we need greater clarity on what are the must win science and technology objectives that we have and how are we going to work together to deliver those outcomes.
Stephen Overlee
Well, Undersecretary, we are at time, but thank you so much for joining me here today and thank all of you for being here.
That's all for this week's Politico Tech. A quick programming note. Politico Tech is off next week for the Thanksgiving holiday, but I'll be back in your ears on Thursday, December 4th. Until then, go ahead and subscribe and catch up on past episodes. And for more tech news, subscribe to our newsletters, Digital Future Daily and Morning Tech. Our producer is Normal. Moleiko Pranbandy made our theme music. I'm Stephen Overle. See you back here in December, Sam.
Date: November 20, 2025
Host: Stephen Overlee
Guest: Dario Gil, Undersecretary for Science, U.S. Department of Energy
In this episode, Stephen Overlee sits down with Dario Gil at the Aspen Cyber Summit to discuss how the Trump administration’s energy agenda is being shaped by the rapidly increasing demands of artificial intelligence (AI) and other advanced technologies. The conversation explores how U.S. energy policy is adapting to power-hungry tech like AI and quantum computing, the resurgence of nuclear power (especially fusion), competition with China, the essential balance between innovation speed and safety/security, and the changing role of government in research and development.
Electricity Demand is Soaring:
“For many decades we had seen electricity demand either flat or slightly declining … Now we're seeing computing environments and data centers that are in the gigawatts … So in the near term it is about … enabling that with the technology that we have right now.” (Dario Gil, 02:16)
The Key Bottleneck:
Short Term vs. Long Term Horizons:
“From an energy perspective, fusion is an area that we're putting a lot of energy into.” (Dario Gil, 04:57)
Fusion: From Hype to Reality
“We have multiple teams in the United States in pursuit of q greater than 1 … For the first time now, we're building towards having a fusion pilot power plant … We're looking at horizons now where the rubber is going to meet the road.” (Dario Gil, 06:14 and 06:51)
“We are going to witness this. This is not something that we're going to talk about generations from now. So I am bullish in that sense, but realistic of how difficult it is technically.” (Dario Gil, 07:43)
“When we're talking about these kind of computing centers and so on, one of the things that is very prized is density ... the ability to do, I mean, go back to nuclear ... It is because of the density that it can provide and the reliability.” (Dario Gil, 08:48)
“It is an absolute imperative that the United States remains in the lead and actually accelerate our competitiveness on AI and quantum and high performance computing technologies.” (Dario Gil, 10:19)
“If we do not succeed ... and China does that ... we're going to be very sorry for ourselves. We're not going to let that happen.” (Dario Gil, 11:07)
Safety vs. Racing Ahead:
Building Security From the Start:
“We spend a lot of time dealing with this ... there has to be a very focused effort on equipping our teams ... The rate at which AI-driven cybersecurity attacks ... can only be countered through technology itself.” (Dario Gil, 13:58–14:35) “Every single initiative that we launch from day one, the research, security and the protection ... is baked in the design of everything that we do.” (Dario Gil, 16:31)
Institutional Diversity is a Strength:
“Thankfully, in the United States we have a lot of institutional diversity, a lot of people that care, different stakeholders around that. But it's not appropriate to say, ‘until I get up to speed ... I want everybody to stop making progress.’” (Dario Gil, 15:38)
Shifting Landscape:
“The federal government provides an indispensable and irreplaceable foundation ... But ... today, 70% of funding comes from the private sector.” (Dario Gil, 18:37–19:59)
Outcomes Over Inputs:
“An outcome is delivering the capability that the nation needs to make sure that we are successful in our economic and national security fronts ... If five years from now we doubled the number of centers ... but we didn't deliver the good, I don't think anybody is going to sit around here and say, ‘You guys did a great job.’” (Dario Gil, 21:36)
Coordinated Effort is Key:
On Fusion’s Promise:
“If that happens, I mean, we should underestimate how amazing that is, right? The fact that we are able to … use a basic mechanism that powers the stars and be able to build a technology here on Earth to do that, I think is one of the pinnacles of scientific and technological achievement.”
— Dario Gil [07:29]
On AI’s Energy Bottleneck:
“The rate limiting step on the progress on AI is energy at this point.”
— Dario Gil [02:56]
On U.S.–China Competition:
“If we do not succeed … and China does that because they've adopted and changed the methodology with which science and technology and R&D is conducted, we're going to be very sorry for ourselves. We're not going to let that happen and we have to do that.”
— Dario Gil [11:07]
On Balancing Progress and Security:
“I don't think we can afford to say, well, let's all sit down and wait until we figure everything out. ... I want everybody to stop making progress. So it's that balance around demanding every institution to race to the challenge and do their job.”
— Dario Gil [15:38]