>> democracy is always an unfinished creation. >> democracy is worth dying for. >> democracy belongs to us all. >> we are here in the sanctuary of democracy. >> great responsibilities fall once again to the great democracies. >> freedom and democracy must be constantly guarded and protected. >> we are still at our core a democracy. >> this is also a massive victory for democracy and freedom. >> california democratic congressman ami bera and you tell republican senator john curtis joined a discussion on u.s. competitiveness in artificial intelligence, focusing on national security and the race against chinese technology. >> my name is stephen overly.

i'm thrilled to be here with our panelists. i talk a lot on the podcast about ai and its implications. and i don't think the consequences or potential consequences are any more significant than when you talk about national security. whether that is competition with china, the war in ukraine. whoever dominates in ai is going to wield significant geopolitical power. that is the backdrop of our conversation. i want to start a conversation with some news this morning, which any of you may have seen. the biden administration signed an executive order directing the department of defense and department of energy to identify government sites where k.i.a. data centers can be built. i'm sitting -- senator curtis, i'm curious to come to you as someone who is focused on ai energy and infrastructure. what is your reaction to this latest news? sen. curtis: to put things in perspective, in utah two thirds of that state is owned by the federal government.

so if you are not looking at federal land you are taking two thirds of the map off of the equation. think about this. thing about trying to transmit energy. if you cannot go across that two thirds of the state this is good news. i'm anxious to learn more about it and i tend to think, as i was thinking about this this morning, i'm thinking about, i can picture in my mind a data ai center hooking up to a fire hydrant and no water coming out. that is what i feel it is going to be for power, right? current trajectory late -- current trajectory rate nothing is going to come out. mr. overly: chris, i want to come to you to build on that. openai is one of the supporters of ai imperative. you have also -- also argue that data centers and energy are key if the u.s. is going to be china in ai. does this morning's executive

order actually move the immediate cash needle in a meaningful way? -- needle in a meaningful way? and what does it leave out? mr. lehane: i'm always reminded of a book i may or may not have read in the late 1990's. "guns, germs, and steel." that made the case that nations succeeded when they were all top to marshall there national resources and get a competitive advantage. if you think about ai as the electricity of our generation, it's going to impact everything. to be able to prevail on ai for democrat, u.s.-led ai to lead the world that is ultimately infrastructure -- an infrastructure exercise. it is going to be chips, it is going to be data, it is going to be talent, and it is going to be the energy piece. so, to get specifically to your question, i think what the biden

administration put out today does send an interesting signal in the following way. we have identified that there is roughly $175 billion of dry capital out there globally looking to invest in ai infrastructure. there is an understanding that there is going to be a need for it. in part driven by the geopolitics we will get into. that there is an interesting return while on it. the question for this country is, as two countries that can build ai at scale. the u.s. and china. do we want that 175 billion in the u.s. to create jobs and economic activity, or the p.r.c.? the biggest obstacle. if you think of the economics of -- these data centers, i think of them as supercomputer laboratories. your parents' oldsmobile. the biggest challenge is you have to buy your energy at the front end and you are ultimately

going to sell your computer on the back end. for people investing in money, how long is it going to take, the delta between buying the electricity and selling the computer? what you get with the biden administration is, at least from a signaling perspective, at least on federal land, trying to short the timeline between when you get your project, shovels in the ground, and the project going forward. i think you know, putting an emphasis on the fact that we do need to get to a place where we go from thinking big to acting big, to ultimately tilting big. mr. overly: when we are talking about the first 100 days of the next administration, if this is a signaling exercise what is a tangible step the next administration can take in the first 100 days to advance on that and make a tangible impact? mr. lehane: the senator mentioned utah.

obviously there is federal lands across the country, but i do think, can we take this basic concept and figure out a way to work with states on non-public land, to be able to expedite the building process? i think that is going to have to be incentives moving forward. some of the challenges around data centers have been, with the states benefit from this? they get come -- some construction jobs. those are all good things, but can the states think about these data centers as being larger ai ecosystem, economic development opportunities? something we have suggested is, you look at a data center potentially needs one gigawatt of energy. that is about the size of what portland, oregon takes in energy. could you have a certain segment of that computer allocated for that states purposes?

if it is megawatts, that state can use it to stand up its own ai ecosystem so they are not just getting the jobs from the building, they are building a sustainable economic model for themselves. you give that in return for the state expediting the building process. i think the signal today is we do need to address the timelines to get this stuff built, particularly given the imperatives. but as we think about the federal level, and we have a senator and congressperson here who can talk about this -- do get a sense that there is an emerging interest in doing big infrastructure type of legislation. whether it is one bill, several bills, but thinking about how you can create those incentives for the states. mr. overly: congressman barrack, notably one of the -- congressman bera, companies that build these data centers, they have to be built using clean energy. it is a prime example of some of

the trade-offs and considerations, right? do you rush to develop this technology, even at the expense of climate goals or the expense of labor goals? we have seen a lot from this administration trying to balance those. i'm curious, is it a fair trade if you get government land to build a data center you have to use clean energy? rep. bera: i think it is, and it think there is a bipartisan opportunity. if we think about modular nuclear energy, it's going to be easier to do that under a republican administration that it would have been under a democratic administration. is that a place where there are enough of us on the democratic side who think that technology is moving forward, relatively safe? gaetz has invested a lot in this area. could we find bipartisan agreement? the energy needs of these data centers is going to be huge. it is not going to be wind or solar, it's going to be new sources of energy. i think that is the area we

could find some bipartisan meant to move forward. another announcement the biden administration recently made was these advanced manufacturing centers. i think there is one coming to sunnyvale. i think that is a smart investment, to put federal resources into that next generation of chips. to make sure we continue to keep the lead bit. modular energy is an area we could work together on. mr. overly: it's been a big week for last-minute minute ai policy pushes. but yesterday the commerce department also announced new export controls, effectively restricting the sale of microchips used to develop artificial intelligence. that included country caps. not just on rivals like china, but even on some u.s. allies. zach, i want to come to you. this is your wheelhouse. in the first trump administration you oversaw export controls.

will these new restrictions hamper adversaries it in -- and the way they have been promised? and with this new administration do you expect them to follow through on this? ms. nikakhtar: that is a great question. i wanted to talk about the china competition generally. it is one of the reasons we are talking about ai, right? when we are evaluating china competition heisey a lot of things in the news about how the u.s. is leading china, but i have seen this story before. for the last 25 years. every time we have gotten there we thought, chips and batteries and whatever, we are so far ahead of china. china races and catches his -- catches up and beats us. we have to remain vigilant. when we think about ai in the context of the elements, the hardware, the chips, the software, the language models, data, energy, money goes into it, china is going to invest

between $1 trillion in $2 trillion. regulations, infrastructure, and workforce. we have to think about all of those elements in terms of competition, right? one or two are not enough across the board. which brings us to the chip controls, right? and i 100% think the trump administration is going to -- they will agree with the imperative, but the execution was a normal sleep flawed. right? we are putting arbitrary caps for countries on computer power. rather than using chip diplomacy, right? if we are the preeminent ai chip designers, etc., then we have to use diplomacy to get countries to work with us rather than move to china. and yes, maybe china's chips are further behind, but they can put more together to get that same compute power. but arbitrary caps make no sense. we export the chips, we give

american businesses the opportunity to get in those markets, but then the china restrictions would be end-user restrictions, right? if these data centers, ever, are going to be helping adversaries, then that is the limitation on export controls. yes, we are moving in the right direction, execution flawed. i think and am hopeful that the trump administration will fix it. mr. overly: senator curtis, is this a rush job or policy? sen. curtis: i want to answer all of their other questions. [laughter] let me just say it feels a little bit like somebody has to say the emperor has no if you say we are going to let you do these on public lands, but they have to be clean energy, it is going to take five years at the fastest -- and probably more like 10 years to do that with nuclear, right? where are the minerals going to come from? and are the solar panels going to be built in the united states?

i think we really are fooling ourselves in many respects. not only what you have just said, but the energy side of this. it is great to say we are now opening up these lands to do this. where is the energy going to come from? if you put restrictions on that it is not going to happen. i think we have to be more realistic with the volume of energy we are talking about to do what needs to be done. it is going to be natural gas. in long-term it can be nuclear and renewables can layer it on top of that, but i don't think anybody sees a path forward. let's be honest about that and talk about, how do we make natural gas clean, right? i think those are questions that are far better questions if we really think we are going to build these type of ai facilities. if we don't, they will be built overseas and use coal. ms. nikakhtar: also an important point we should tease out, the

semiconductor manufacturing relies on critical minerals. it relies on these industrial diamonds, and i tried to go into all of the cutting tools that china controls. the dominant share. china controls mineral processing, super abrasives, etc.. though supply chains are exclusively in china. we have no hope of beating china and manufacturing chips unless we get our act together and reassure of the components -- were friend-short, but make sure we have long-term, secure, stable supply chains if we are going to be in this race. i think everybody gets so excited about the applications rather than -- we should -- we should also think about, how do we do this? what are the inputs they go into making this? those are important and i appreciate your flagging that as well. mr. overly: whenever i talk to members of congress about their ai priorities national security

or otherwise, i start to hear this sound in my head. it all sounds expensive. candidly, we have a republican-controlled congress that has talked a lot about putting checks on spending. i guess my question for each of you from the i'll use it in, how do you convince lawmakers -- or do you need to convince lawmakers to spend on these ai priorities? sen. curtis: i would like to use ai in my office. the burden that is put on ourselves even to use it is astounding. so, first of all, i don't have a lot of confidence in the congress. but i tend to think, to answer your questions, our job is to put the guardrails on this thing, and the free market is ready to roll. i think the only thing holding things back is not money, it is the guardrails.

and we don't put guardrails out everyone is scared. if you don't know what you can expect and what the rules are going to be it is hard to make those type of investments. my guess is, if we put the guardrails out there the free market would go crazy. rep. bera: i'm going to agree. if there is a priority in the first hundred days it is, what did those guardrails look like? in california the state legislator -- legislator tried to put the guardrails up. we cannot have 50 different regulatory regimes. we should try to figure this out. how do we get that right so we don't stifle innovation but we also, you know, protect the public as well? we did not do it around social media. now we are trying to do it and it is too late. it is really hard to do. ai is moving really fast. i think there is a priority. but the spinning aside. there is a priority to figure that out in a nonpartisan way that allows us to have this innovative ecosystem but also

sets those guardrails at the right place. mr. lehane: can i connect two dots here? competition is incredibly important. for those not in ai, the prc released a model in december called deepseek. deepseek competes with the best models in the u.s. at this idea -- and you were hitting on this -- that we are way ahead on this is not the case. and the stakes cannot be higher. vladimir putin has gotten this right when he says whoever wins the ai race wins the role. the ccp has said china wants to be the dominant player and by 2030 they get to new in the last year alone, and more coming. they control all of the data. the data goes into these. i think we are ahead in the chips but they can throw massive resources at ease.

in a specific idea, if i could connect the dots here, one of the things both members here, you guys will spend time thinking about this and be smarter on, but the u.s. government is going to have to buy compute u.s. government purposes. i mentioned earlier $175 billion looking to invest. need to make sure the economics work from a return model. if the u.s. government can basically do a backstop by pre-paying or committing to buying that compute it actually unlocks that capital to come into this country to invest. into these data centers. that is how you can really get an interesting potential avenue for interesting public-private approach. government is going to have to do it anyhow. if you make those commitments, you bring the capital in to expedite the buildout, those data centers become infrastructure as destiny. they become effectively our ai arsenal of democracy as we go forward to ensure we beat china.

ms. nikakhtar: i want to stress, where america has fallen behind in the competition with china is because we have not had a serious whole of government strategy moving forward. we are completely fragmented. one agency is doing one thing, another agency is doing it. need to bring it together. and i know the term shot is over-used, but for lack of a better term and something everybody understands, we need a moonshot in chips. all of that starts with a whole of government strategy. mr. lehane: and we have done it before. it is our superpower. mr. overly: how do you bring it all together? should trump set up on -- a department of artificial intelligence? ms. nikakhtar: that is what the national security council is therefore. i don't know if you all have seen the chinese robots, but they are incredible and functioning with ai. they are going to proliferate the world and china is going to

potentially win the ai race. we are at the cusp of winning or losing everything because ai is going to be everything. it's going to be the national security council, and they need to bring technical experts from the outside, right? and agency experts to figure out , how does each agency we've together its authority into a comprehensive, executable, expedient ai strategy? mr. overly: one thing we know the president has promised to do is revoke biden's 2020 three executive order which set up the u.s. ai safety institute and had a number of directives for agencies to cultivate ai-focused policies. if that is scrapped what are the implications for national security? sen. curtis: i don't think it's a great idea. or if it is scrapped, what replaces it. you have ai that is going to be

part of our everyday life. transportation systems. i focus a lot on ai and health care. that stuff is already moving. how do we make sure that is safe and efficacious and folks are protected? so if it does get replaced i would be curious what the trump administration -- because safety is a big concern of those guardrails. the other interesting thing we are going to be watching is, the president is going to be surrounded by a lot of tech talent. if you look at some of the nominees, not sure what elon musk is going to do here. and they don't all agree on how we should approach ai. i think there is going to be some -- hopefully it is creative tension within the administration that comes out with good ideas, but i think there is going to be a lot of tech talent in this administration. mr. overly: do you want to weigh in? sen. curtis: i'm hearing ronald reagan in my head. i'm from the government, and i'm

here to help. i think government has to be careful here. i will give utah a shadow. there have been some creative legislation on this. mr. overly: leading the country in a lot of this stuff. sen. curtis: they have the sandbox where you can try and experiment on ideas and i've heard the term -- we don't want to stifle this -- and it is so true. if we are not careful in our good intentions we will stifle this. another thing is, if you look at how we have regulated technology in the past we tend to -- we tend to pick a point in time and pass legislation based on what we know. now we are trying to run the internet based on legislation dickens and decades ago. i think also we have to somehow legislate in a way that moves with this and is adaptive to what we learn as we go forward with ai. that is not easy to do and that is the reason we have not done it yet, but i think it is important we figure this out.

mr. overly: let me ask you a follow-up question that has been on my mind. as much as ai is a race it is not one the u.s. is running alone during i think in many ways a big difference between the trump administration and biden administration has been their rhetoric toward u.s. allies. we have even seen, obviously president trump talk about sweeping tariffs. we have seen threats to annex canada. is there a risk that his style, his approach undermines, kind of, the global relationships that the u.s. needs here when it comes to ai and national security? sen. curtis: all i can tell you is, we have four years of prior experience and there is lots of fears of that, but i don't think it played out. in my opinion as long as we don't ruin it this is going to happen here in the united states. the most important thing is that

the house and senate just don't get on the way of this. we put the appropriate guardrails up and i will bet on the u.s. leadership and innovation every day, all day long. as long as we don't ruin it. mr. overly: we are running out of time but i want to get one more question in. i will start with you, chris, because i want the industry expectation -- industry perspective. there is an expectation that the administration will defend tech companies on the global stage more than the biden administration has. whether that is regulations coming out of europe, for instance, on digital markets, or artificial intelligence. i'm curious if that is your expectation, and as you look at the first 100 days is there some sort of stand you would like to take -- you would like to see this administration take for big tech? mr. lehane: from the tech perspective and specifically ai, it is something i think has been touched on almost by everyone on

this panel. which is an overarching what i would call coherent or holistic theory of the case strategy. i will touch on something i said earlier. that deepseek model is the equivalent of a sputnik moment in ai. this country reacted to smart -- sputnik in the 1950's by putting its foot on the gas pedal and winning. we absolutely have to play to win. i think this is an administration coming in that has made very clear that when it thinks about ai there is two things that is looking at. one, ashley security. two, economic competitiveness. two sides of the same coin. my hope or aspiration is that you're going to get an overarching strategy that is going to come at it to deliver on those two goals, those two outputs. and with the urgency of this moment demands.

because this is a zero-sum game. the u.s. has to win this. because winning is going to dictate whether the world is built on free, democratic ai rails, or on authoritarian, autocratic ai rails. and we have to win that. mr. overly: that is unfortunately all the time we have. i could go on, but i want to thank our panelists for joining us in preventing such great insights, and i would like to thank all of you for being here. [applause] [captions copyright national cable satellite corp. 2025] [captioning performed by the national captioning institute, which is responsible for its caption content and accuracy. visit ncicap.org] ♪ >> nonfiction book lovers, c-span has a number of podcasts for you. listen to nonfiction authors and influential interviewers on the afterwards podcast. and on q&a, hear wide-ranging conversations with nonfiction authors and others making things happen, and book notes plus

episodes are conversations that regularly feature fascinating authors of nonfiction books on a wide variety of topics. find all of our podcasts by downloading the free c-span now app, or wherever you get your podcasts. and on our website, c-span.org/podcasts. >> c-spanshop.org is c-span's online store. browse through our collection of products, apparel, books, home to court, and accessories. there is something for every c-span fan, and every purchase helps support our nonprofit operations. shop now or any time at c-spanshop.org. democracy, it's a process shaped by leaders elected to the highest offices. and entrusted to a select few with guarding its basic principles.