Powering America's AI Future: Assessing Policy Options to Increase Data Center Infrastructure

Is it ready?

Yes.

You good?

Yes sir.

Okay. The Subcommittee on Investigations and Oversight will come to order. Without objection, the Chair is authorized to declare recess of the subcommittee at any time. Welcome to today's hearing entitled " Powering America's AI Future: Assessing Policy Options to Increase Data Center Infrastructure." I recognize myself now for five minutes for an opening statement. Good morning and thank you for being here. Today we are discussing an issue of intersection of energy policy, innovation and national competitiveness. Artificial intelligence requires enormous computing power, and that means enormous amounts of electricity. That's not up for debate. What is important to discuss is whether our approval process to build infrastructure can meet the demand within a meaningful time frame. Right now, they're not. We need to understand why and what Congress can do about it. We must consider transmission line projects which are essential to delivering power to new data centers. Permitting for those projects currently averages about four years, in some cases, and stri- stretches up to a decade, which is a tremendous amount of delay. Meanwhile, federal environment reviews under NEPA averaged twenty-six m- months in two thousand twenty-four. more than two years before construction even begins. That's a serious concern when infrastructure must keep pace with rapidly advancing technology. Then there's litigation, even when agencies ultimately prevail, when they do, roughly just three quarters of the time, the process itself can add up to six more years in litigation delays. Imagine the amount of technology advances in six years, add it on to the other two years. Up to a decade. That's crazy amount of delay. Meanwhile, China is building data centers and power structure with incredible speed through centralized planning. I'm not suggesting we adopt their system, but we should acknowledge that our competitors are not operating under the same constraints. Schools like the College of Engineering at University of Georgia are helping respond to China's surge by developing AI-enabled transmission planning tools, designed to accelerate complex grid studies while preserving reliability. Now ironically, they're based on AI models powered by the same problematic concerns. Using AI to help plan for the infrast- infrastructure that powers AI is exactly the kind of innovation we should be encouraging. Last year, the Trump administration issued executive order one four three one eight to dec- to address some of these delays by prioritizing permitting for large data centers and improving coordination across agencies. For agencies under this committee's jurisdiction, including NIST, DOE, and NASA, the implications are significant. However, executive action has its limits. The underlying status of NEPA, the Endangered Species Act, and the Clean Water Act remain unchanged. There are also practical consequences. Several projects have encountered substantial permitting hurdles despite their strategic importance. These delays not only affect private companies but also federally supported research ecosystems for those facilities that are meant to anchor these programs. Cost allocation is another a- issue. Data centers consume vast amounts of electricity, sometimes ten to fifty times more per square foot than traditional buildings. Under traditional utility rate structures, infrastructure costs are spread across all customers. That also needs to be addressed. That raises serious questions of whether residential r- rate players could end up subsidizing upgrades that primarily serve high intensity industrial complexes. Some companies are addressing this proactively. Google has partnered with utilities to fund new generation capacity directed rather than shifting costs to the rate payers. Other companies are exploring rate uh space-based computing platforms that would operate outside the traditional terrestrial permitting framework. While still early stage, these concepts are highly and are are highlight how regulatory friction can influence where and how infrastructure is developed. We should also consider how this conversation extends beyond major met metropolitan regions. Much of rural Georgia and rural America still operates under a cooperative grid infrastructure originally built in the mid-twentieth century electrical policy. That system was not designed for compute intensive infra- infrastructure. That system, i- if the AI integration expands into agriculture, logistics, and public services, world grid capacity could become a binding constraint. So what needs to happen? We must evaluate whether a permitting framework, largely designed for the nineteen seventies, is suited for infrastructure that now is central to national security and economic competitiveness through two thousand twenty six. That includes reviewing litigation timelines and considering whether AI related infrastructure warrants dif- determines different treatment. I'm not advocating that we eliminate all environmental review. We should always be friendly to the environment. We all love to preserve. I'm not advocating that we change the rules in some sort of dramatic capacity, but we do need to change, or we will fall behind. That much is certain. This committee's responsibility to understand how regulatory frameworks affects innovation, research infrastructure, and American competitiveness is critical to these technologies. That is the purpose of today's hearing. Our witnesses can help us clarify both the scale and the challenge of potential solutions and I appreciate their willingness to be here. I look forward to a substantive discussion how Congress can ensure America maintains its technological leadership while addressing legitimate, regulatory concerns. And now I recognize the ranking member, representative from Ohio, for her opening statement.

Thank you, Chair McCormick, for holding this hearing in uh certainly a timely and relevant discussion. So when you go to watch your favorite show, send an email, shop online, or save a photo, you're using what is known as the cloud, which is really a general description for a data storage service. But this cloud is not in the sky. It's in our backyards. To manage the massive amount of data used by consumers and industry, Tech companies have constructed massive data centers full of rows and rows of computing and network com- equipment. And these very real facilities have very real consequences for the communities that they are in. Many people may be more familiar with maybe a server room in their office. Uh, think of a service, uh, uh, one of these type of server rooms at a much more large and massive scale. In Ohio alone, there are two hundred and seventeen data centers, the fifth most of any United State. States. This includes four in Stark County and seven in Summit County in my district. I wanna be clear that I'm not reflexively opposed to data centers. In my district, Ohio's thirteenth, a number of townships are already hosting these centers and will alway- and I will always support these communities to pursue job creation and added tax revenue through large infrastructure investments however that does not mean that Ohioans should pay higher electricity bills for billionaires in Silicon Valley to see higher profits. We must fully understand the trade-offs of data centers, both the potential they hold for quality of jobs and investments in our communities and the impact on electricity costs, our environment, and even our health. For instance, my district is home to a cryptocurrency company that once operated the world's largest data center in China, but is now headquartered in Akron, Ohio, now known as Seoul AI Limited, with a data center in the middle of the city. This data center can use about eighty-three megawatts power per day and has the ability to draw enough electricity to power sixty-three thousand homes. For context, there are ne- nearly eighty-five thousand homes in all of the city of Akron, and Ohioans have already begun seeing increases in electricity costs because of these data centers. Moreover, due in part to increased energy use, studies have shown that communities can see increases in air pollution from these crypto-mines increasing health risks. And of course, unfortunately, we can't talk about energy in Ohio without talking about bribery. Seoul AI, formerly known as Bit Mining, was also fined ten million dollars in a foreign bribery scheme, again, now headquartered in the Midwest of the United States. It is clear that we need to do more to protect residents near these centers, while also ensuring that they are operating using good legal business practices that do create jobs for our communities. Many communities offer generous incentives to these data centers, and we need to make sure that they are holding up their end of the bargain. especially when using taxpayer dollars. The truth is that there is much we don't know yet about these data centers, but the known impact of electricity prices, the environment, water utilities and air pollution should all give us pauses before we agree to an unprecedented build-out of data center infrastructure without any guardrails to protect the American people and our bank accounts. It is no secret that land for these centers is cheaper in some parts of my district compared with other parts of the country and I'm certainly not opposed infrastructure investments that will boost Ohio and our country's competitiveness. But while ultra-wealthy companies may find our property prices affordable, constituents in my district are facing an affordability crisis everywhere we turn. So, we are at a crossroads. We need to decide the right way to harness this explosive growth in data center infrastructure so that they increase job opportunities and investments in our communities instead of simply increasing costs for the American public. One path is the Trump administration's approach, few safeguards, rapid expansion, and a total embrace of data centers with little thought to the communities that host them. That's not the approach that I advocate for, and I hope that we don't take that method, that road. We need to choose a path that avoids mistakes before they are made, instead of dealing with them afterwards and putting strong guardrails in place to protect our communities, not only our health, our environment, as well as our bottom lines. If we can address these questions with rigorous, high quality research, our country and our constituents will be far better off. To that end, I am pleased to welcome Doctor Eric Massonet to the hearing. Doctor Massonet is a leading researcher trying to model and understand the long-term impacts of data centers and his perspective will be invaluable for the committee and apologize for the challenges getting in today. It's a little bit busy day on the Hill today. Uh, thank you for your willingness to testify, as well as all of our Uh, witnesses, and thank you. Mister Chair, yield back.

Thank you, ranking member Sykes. I now recognize ranking member from the full committee for statement.

Well, thank you, Chairman McCormick and ranking member Sykes for holding this hearing, and I thank the witnesses for coming here today on such a busy day. Uh, this hearing is a great opportunity for all of us to think carefully about an issue that's moved to the center of our political uh debate. Data centers loom large in the United States today and They are presented as the foundation of America's global leadership in uh artificial intelligence. They're the subject of projected investment on an astonishing level, perhaps hundreds of billions of dollars in the years to come. They're a major source of economic growth and an engine of the stock market. Everything about data centers seems outsized in scale and uh ambition. Of course, there's another side to this story. The impacts of data centers on local communities can be just as huge as their national aspirations. Data centers require massive amounts of energy to operate, and their demand for electricity, uh, threatens to strain our grid and to send energy prices surging for residents and small businesses. They consume enormous amounts of water to prevent overheating and emit air pollutants from back-up diesel generators, that can expose nearby communities to health hazards. They create noise pollution that can distress people living nearby, and their demand for power, if it comes from fossil fuels instead of clean energy, could deal a, uh, a dangerous setback to America's clean energy transition. Now this, uh, debate has a familiar ring to Californians. Our state has considerable experience in trying to balance the transformative economic growth uh, that technological innovation can bring with the impacts of that growth on the cost of living and affordability. Certainly, for those of us in the San Francisco Bay Area, this has been a crucial question for many, many years. And California will be central to this debate once again, as the state with the third largest concentration of data centers in the country. I have no doubt that different communities will reach different conclusions about what's in their best interest. But there's one thing we should all be able to agree on. More data and more transparency about data centers uh is im- is essential.

Hmm.

Communities should not have to make decisions about whether to host a data center without access to basic information uh about electrical uh usage and water consumption, air quality, noise pollution, and the potential effect on energy prices. Researchers should be able to use real-world data uh, instead of relying so heavily on estimates and assumptions, all of us, policymakers, regulators, researchers, local communities, really the American people, need consistent, reliable, transparent data to make informed decisions about data center growth. The lack of that transparency is troubling and I look forward to discussing it further today. It's worth remembering that ignoring data center impacts does not make them go away. Now the Trump administration has become America's biggest cheerleader for data center build-out. All we hear from the administration is that the best thing for America is more data centers built as quickly as possible in as many parts of the country as possible. That strikes me as foolish and short-sighted. If our country isn't smart about data centers, if we fail to act with diligence and prudence, mistakes will be made, and that will inevitably lead to backlash. And the beginnings of that are already, uh, starting to be seen. Data center operators and supporters should appreciate that their social license to operate is a valuable but also a fragile thing. It's in their own self-interest to be as transparent as possible about data center impacts so the country can make reasoned judgments about how to sustainably mitigate them. I hope to learn more in today's hearing about the f- how the federal government can do more to support this kind of transparency and, Mister Chairman, I yield back the balance of my time.

Thank you. Ranking member Lofgren, I appreciate your statement. Let me introduce our witnesses now. Our first witness today is Miss Paige Lamberton, Research Fellow at the Center for Energy and Environment for the Competitive Enterprise Institute. Our next witness is Mister Marsden Hanna, Head of uh Energy and Sustainability Policy for Google. And our last and final witness is Doctor Eric Massonet, Professor and Mellencamp Chair uh in Sustainability Science for Emerging Technologies for the University of California, at Santa Barbara. I now recognize Miss Lamberton, sorry, Lambermont, for five minutes to present her testimony.

Chairman McCormick, Ranking Member Sikes, and distinguished members of this panel committee, thank you for holding this hearing today and inviting me to testify. I'm Paige Lambermont and I'm a Research Fellow in the Center for Energy Envir- and Environment at the Competitive Enterprise Institute, a non-profit, public policy organization dedicated to free market principles with a focus on regulatory issues. I'm grateful for the opportunity to speak to you today about the importance of permitting changes and innovation to meet rising power demand. I have three main points that I want to make in my testimony. Power scarcity has been created by policy decisions, permitting reform is essential to meet this challenge, and now is the time for new and innovative solutions. Power scarcity is a policy decision. For the last several decades, energy policy in general and electricity policy in particular, has focused on making do with less, curtailing demand, producing less power, and electrifying things like cars and appliances. These policies have simultaneously created less reliable electricity, while also using government intervention to increase reliance on electricity. This was never a plan that was going to foster reliability, but now that demand has begun to rise for the first time in nearly two decades, the true folly of that approach has been revealed. According to a DOE report, one hundred and four gigawatts of firm power capacity are set to retire by twenty thirty. At the same time, demand growth by twenty thirty is projected to be between thirty-five and a hundred and eight gigawatts. The delta between retiring capacity and new demand will require new reliable capacity to fill the gap. Ideally, the permitting and policy ecosystem would favor the swift construction of new reliable power plants and the supply chains for building these plants would still be intact. In this hypothetical world, new power demand from data centers would be an opportunity to generate electricity to solve complex problems and generate economic productivity. But on our broken power grid, where it is incredibly difficult to build new power plants and the supply chain to build those power plants has broken down, this opportunity becomes a challenge. The priority should be the formulation of policy that turns that challenge back to an opportunity. We are more than capable of meeting growing power demand, but this will require removing regulatory barriers that impair the construction of new power plants and the reli- reliability of the power grid. There are many policy changes that will contribute to this, But two of the most important actions will be comprehensive, broad-based and technology-neutral permitting reform and removing obstacles to innovation. Permitting reform is essential to meet this challenge. To help solve the current supply issue, our federal permitting system must be reformed in ways that go beyond the National Environmental Policy Act. The Clean Water Act, Clean Air Act, Endangered Species Act and other laws must be amended in parallel to remove overbroad, duplicative and unnecessary requirements to secure permits. This would help create a federal permitting process that allows new development to occur in a timely, predictable, and cost-effective manner, while maintaining high environmental standards. Regulatory agencies also play a role in making it more difficult to build new power plants. So while the EPA is a primary culprit, another agency that's been especially restrictive is the Nuclear Regulatory Commission. The advance act helped to bring the mission of the NRC in line with the goal of stewarding, rather than simply overseeing the nuclear industry. But improvement must go much further if nuclear power is going to play a role in meeting rising demand. In China, the current timeline for a large reactor from start to finish is seven years, and there are currently twenty-seven reactors under construction. The last large reactor built in the US were plant vocal units three and four, which took fourteen and fifteen years respectively and began construction in two thousand and nine. The United States is not losing on quality of technology or ability to innovate in this space it's losing on regulation and that is entirely a policy choice. Ensuring that the regulatory structures that govern the development of new power plants of all types are reasonable, predictable, consistent, and scientifically-based, is essential to develop new power plants at the necessary speed to meet rising demand and remain globally competitive. Now is the time for new and innovative solutions. Another option to ease the bottleneck around getting new power on the grid is to enable innovation including off-grid utilities. Consumer-regulated electricity is an opportunity to create an alternative to the traditional grid pathway for new customers, that are willing to pay for their own electricity. CRE would allow for privately financed off-grid electric utilities to serve customers including data centers. These utilities would be physically islanded from the broader power grid. This would allow companies who want power quickly to avoid the traditional utilities queue while bearing the cost of their own infrastructure. CRE provides flexibility and speed to companies working on innovative projects, while providing cost insulation to rate payers. For this to be possible, individual states would need to pass legislation exempting islanded utilities from their definition of a public utility. This policy innovation has already become law in New Hampshire, and related legislation has been introduced in several other states. At the federal level, the Data Act S thirty-five eighty-five, introduced by Senator Tom Cotton, would enable CRE by clarifying that CRE utilities are exempt from federal regulation as public utilities. To meet growing power demand and protect rate payers, it is essential that we both reform our permitting system and remove government obstacles to allow new and innovative solutions. It is technologically possible to meet rising demand. And we can and should do so in a way that protects existing rate payers while allowing for new development. Thank you.

Right on time. Thank you, Miss Lambert-Mont. I now recognize Mister Hanna for five minutes to present his testimony.

Chairman McCormick, uh, Ranking Member Sykes, and members of the subcommittee, thank you for the opportunity to speak today. My name is Marcin Hanna, and I developed Google's approach to energy policies that affect the company. I appreciate the opportunity to discuss the intersection of AI and energy with the subcommittee. As the United States races to lead the world in AI, permitting and transmission reform has become a critical lever for maintaining our competitive edge. The potential for AI to accelerate US economic growth is staggering. Research estimates that AI could result in global economic uplift of more than four trillion dollars annually. However, realizing these gains, uh, as well as accelerating breakthroughs in science and fortifying national cybersecurity depends entirely on our ability to build and power the infrastructure that makes them possible. Today, the United States is facing critical challenges to meeting this opportunity. Fragmented and congested electricity grids across the country are increasingly tapped out and many can no longer support electrical load growth. Transmission lines, many built more than fifty years ago, are now running out of capacity. And adding to the challenge, new transmission infrastructure to power growth and create a more affordable and reliable energy system can take over a decade to permit and build. Many of these delays originate from our outdated and Byzantine infrastructure permitting system which is composed of a patchwork of regulations and bureaucracy at all levels of government. This leads to crippling timelines, uncertainties, and single points of failure, where litigation can delay or kill projects over procedural issues. Congress established authorities in two thousand five that were intended to speed construction of critical infrastructure transmission infrastructure to promote reliability and competitiveness. However, this over- overly bureaucratic process and legally uncertain framework has proven entirely ineffective. And over the last twenty years, this federal process has led to the construction of zero new transmission lines. U. S. electricity policies and statutory frameworks urgently need reform. This is why we've endorsed the House-passed Bipartisan Speed Act, which we believe must be paired with reforms to how we build regional and inter-regional transmission. We believe fundamentally that energy supply is a national security imperative. Winning the AI race requires winning the energy race. Currently, the United States is ahead of global competitors on AI. However, the gap is closing. The speed with which China can plan, permit, and deploy energy infrastructure is a comparative advantage. To give just two examples, China is the largest global power investor by a wide margin, with almost as much investment as the US and European Union combined. And according to a recent report, China has completed over eighty times more high-voltage transmission in the United States in than the United States in recent years. To seize the opportunity presented by AI, Congress should pass comprehensive legislation focused on three core pillars. First, cut energy permitting bureaucracy across the board. Reducing permitting timelines to accelerate project schedules can help reduce project costs and these savings can be passed on to electricity customers nationwide. Second, establish an effective federal transmission permitting framework. Congress can speed construction of national interest transmission lines that provide economic and reliability benefits to power growth by reforming authorities at the Federal Energy Regulatory Commission. Third, Congress should establish clear procedures that provide permitting certainty for new energy projects. Accelerating investment in energy infrastructure requires a consistent and durable framework, uh, upon which investors, asset owners, and purchasers, purchasers of power can rely to get projects built. At Google we believe in the promise of AI, and we are investing heavily in the infrastructure behind it, across the United States, which many of you have seen in your states. In the last six months of twenty twenty five, we have announced investments of forty billion dollars in Texas through twenty twenty seven nine billion dollars in each of South Carolina, Oklahoma, and Virginia seven billion dollars in Iowa four billion dollars in Arkansas, with more announcements to come. To meet customer demand and capitalize on growing opportunities, we uh we anticipate that twenty twenty-six investments will be in the range of a hundred and seventy-five to a hundred and eighty-five billion dollars. Realizing the potential of AI will require robust energy infrastructure, more energy efficient use, and new innovative technology solutions. I look forward to today's discussion and to working with you on these investments, They are a reflection of the opportunity that we see in America. Thank you.

And thank you for your testimony. I now recognize Doctor Massonet for five minutes to present his testimony.

Chairman McCormick, Ranking Member Sikes, and esteemed members of the committee, it is an honor to appear before you today to discuss the critical role of open and transparent data for understanding and shaping the growth of the US data center sector. My name is Eric Massonet and I'm a professor at the University of California, California Santa Barbara, and I have been analyzing the energy and water requirements of US data centers for nearly two decades, starting with the first US data center energy report commissioned by Congress back in two thousand seven. My remarks today are my own, and are not intended to represent the positions of the University of California, Lawrence Berkeley Lab, or the US Department of Energy. As you might imagine, for analysts like me, uh there is no hotter topic these days than the growing infrastructure requirements of US data centers. How much power and water will data centers require, both now and in the future? Which strategies can help minimize infrastructure demands, stress, and impacts on local communities? What are the most effective policy levers for achieving these outcomes? And which innovation should we be investing in today in order to achieve future goals? These are but a few of the questions that analysts uh at many organizations around the country have been racing to answer. And as you can imagine, answering them requires high quality data to produce accurate, timely and trustworthy results. And when high quality data also become publicly available, they provide critical ground truths that allow all analysts to work from a common evidence basis and they support innovation by catalyzing the development of new models new software and forecasts that can unlock a wide range of new insights for different stakeholders. For these reasons, the US has long been a global leader in the standardization, collection and open dissemination of high quality data for every major sector. including power, buildings, manufacture, and transport sectors. We've done this for every major sector that is, except for one, and that is data centers. In fact, no other major US sector, and especially one that is growing so quickly, suffers from as many public data blind spots. So what are the major blind spots, and what can we do about them? The first relates to transparency. Very few data center operators disclose the energy and water use of their specific data centers. and some companies report no data at all. As a result, analysts have to make many assumptions, which leads to a wide varie- variety in estimates about current and future infrastructure demands as you've probably all seen. And this uncertainty hampers effective decision making. The second relates to timeliness. When companies do report, they release annual data that are already out of date for this fast evolving sector by the time analysts receive them. Additionally, important real-time and the infrastructure requirements of the data centers that are being built today, are often kept secret through NDAs. This means that many analysts can be out of sync with what is happening on the ground, and this means that we can't always capture emerging trends that will affect future energy and water demands, like shifts to liquid cooling or on-site generation or energy storage or waste heat c- recovery to name a few. The third blind spot relates to the broader effects of AI. Analysts currently have a very limited understanding of whether the application of AI will ultimately add to or offset the energy and water demands of AI data centers themselves. And we also don't know where these effects will play out. Understanding the net effects of AI, what drives them, and what different stakeholders can do to minimize negative outcomes and maximize positive ones represents a critical new frontier for analysts. However, data for doing so are currently both sparse and unverified. Fortunately, spots can be readily addressed using approaches that are both well established and have precedence for other major US sectors. First, when it comes to collecting more detailed and timely data, it is important to acknowledge that US operators, utilities and local authorities already have the needed data in hand. They just need to be unlocked. Here, government surveys and data collection programs can play the decisive role. For example, the US EIA and Census Bureau have decades of experience collecting and publishing open data sets for major US sectors. And many US operators are already submitting similar data to government databases abroad. Second, when it comes to analyzing the net effects of AI, the US government has a scientific leadership opportunity. It can play a key role in developing improved data standards for measuring AI's various impacts. It could establish data collection and knowledge sharing platforms, and facilitate convenings of experts that advance the nation's scientific excellence. As the old adage goes, you can't manage what you don't measure. However, the blind spots I've described can be addressed and relatively quickly with the right commitments by government and industry. High quality open data benefit analysts across all types of organizations irrespective of their research questions. In my view, given the importance of this sector, investing in improved data foundations to unlock strategic insights should be an, an urgent national priority. Thanks for the opportunity to testify before you today, and I look forward to our discussion and to answering your questions.

Thank you, Doctor Massonet. Now we'll start hearings by uh allowing Mister Harrigan to open up for five minutes. Thank you.

Oh, thank you, Mister Chairman. And thanks to all of our witnesses for your testimony today. Uh, this really is a uh uh a fascinating and impression conversation, uh, given the demands of the future of our market economy. So thank you for your testimony today. I've got, uh, a question for you, Mister Marsden Hanna. Um, from Google's vantage point, sitting on the front lines of artificial intelligence development. Can you elaborate how the current permitting process, as difficult as it is for energy infrastructure, directly benefits our strategic competitors like China in this AI race, and what the long-term cost of our inaction with this permitting reform would look like?

Thank you, Congressman. It's a important question. Um, at - at Google, we believe it's fundamentally important that, um, we accelerate reforms to unlock our ability to build generation and transmission. Um, strategic competitors around the world, uh, do not take into account, uh, factors like cost allocation. Uh, they - they look very differently at environmental - uh, environmental impacts of infrastructure decisions. Um, and in the United States, uh, it's very challenging to build both generation and uh transmission infrastructure. When we look to site new data centers, the primary barrier that we are facing on the uh in the market is uh the challenges interconnecting to the grid. Um, we have utilities telling us it can be five years, sometimes ten years, to interconnect new facilities. And that primarily ties back to grid congestion. And so what we'd - we'd recommend is that Congress pursue uh a - a three-part package where We cut infrastructure permitting ac- uh bureaucracy across the board. We're very pleased to uh endorse the House-passed Bipartisan Speed Act, which we think has important reforms to NEPA. Um, in addition, creating an effective federal process for siting national inf- interest transmission infrastructure that can provide reliability benefits uh and power the next era of - of growth. Um, and finally, permitting certainty so that uh investors and off-takers uh know that they can rely on uh the energy infrastructure to to get built, and to have a durable permitting process to achieve it.

Well, thank you for that clarity. And look, I represent Central Western North Carolina, we have some incredible opportunities in North Carolina with respect to the future of the data economy uh we've got Interstate forty we've got Interstate seventy-seven in my district and farther out east we've got Interstate ninety-five, main arteries on the eastern seaboard. and throughout the Southeast, and that also happens to be where all the fiber is. And so North Carolina is a very natural spot for data center development. We have data centers going all up, uh, up all over our state. And while that is a magnificent economic development opportunity, I think as many of you know, it's also creates challenges. And when we have a finite amount of power, and the demand for that power goes up, the price goes up as well and starts pushing affordability concerns uh out to everyday ordinary customers on the grid. Uh, that's my constituents. And so one of the things that we've been acutely focused on is, uh, and I'm, uh, I think, uh, anybody on this committee knows I'm very pro-nuclear and wanna see nuclear energy, uh, really kind of come and see a resurgence in our country. But I'm also a realist and I also understand that nuclear energy is not our solution today or tomorrow. Uh, we've got long-term considerations of regulatory reform in order to get that into the realm uh of the next several years. But one of the things we can do immediately is on-site natural gas energy production. And talking around the industry, you've said sometimes providers have told you that it's five to ten years, uh, worth of regulatory hurdles to get across the finish line. I've been told directly by natural gas providers that if we eliminated those regulatory barriers from the time that they said go to the time that you could have on-site power generation that would suffice for a data center load is as little as eight weeks. I think that should highlight to everybody that the problems that we actually have are self-inflicted. And the problem is, is that prevents us from getting more data centers. Uh, because at the end of the day, if we're concerned about the price of of power going up, but we can't put that power in. The great thing about data centers is they don't change the culture of a place, as the place develops, because you only have about a hundred people that are gonna staff that data center. They're all making a hundred thousand dollars a year. So you get the tax benefit, you get the increase in the quality of life from a data center, without actually changing your culture, because you don't come with the congestion load that comes with normal development. So, this is something we absolutely

Thank you, Mister Harrigan, and now I yield to ranking members' thanks for five minutes.

Thank you, Chair, and thank you to the witnesses. So as I said in my opening statement, uh, many of my constituents and and I think probably the entire panel here are anxious about the risk of electricity demand and the eventual and current increase of utility costs. uh that are really struggling uh that are making our community struggle. And I appreciate the comments about permitting reform because we've heard that in a lot of different places also sit on Transportation Infrastructure Committee so uh that is a very regular topic of conversation but I do think I would be remiss if I did not mention that the worst bureaucracy here in DC is the Republican-led majority that has been one of the most unproductive in uh modern history uh while the Speed Act uh did pass Congre- the House, we haven't seen much activity from this current Republican majority in the House or the Senate. So I think our first complaint should lead and start with the leadership that is refusing to actually move forward to any of your permanent reform, or even have legitimate hearings on such issues. Uh, but I do wanna ask any member of the panel who might happen to know the answer to this. Do you h- know how many transmiss- transmission projects are currently being held up by, uh, regulatory reviews?

I don't have a figure on that at hand. But would be happy to look at that.

Thank you. So, I asked that question because I've heard it multiple times in the testimony that these transmissions and these projects are being held up but then no one can share how many of them there are and so uh I think we heard if you can't something around, if you can't measure it, you we don't know really what the the s- scope of this issue is and - and trust me I don't think that permitting reform is uh solved by any means, but we also c- aren't ab- able to identify it. Uh, so it makes it a little bit difficult for us to say these are the problems when it could be two, it could be three, it could be three thousand. But no one seems to know, so I do hope that as we are having more discussions that people can bring forth the information that can truly help us make uh these best decisions. Uh, I also noticed that w- during my opening statement many people were shaking uh the panelists, I should say, the witnesses in agreement about not wanting to increase costs, but I didn't hear anything about it in the testimony. So I'll give you an opportunity now. Uh, we'll start with you, Miss Lambermont, about how are s- what are some of the suggestions that you have to reduce costs by also increasing capacity?

Yeah, I think that one of the options, um, that I presented in my th- sort of third point is consumer regulated electricity, the ability for new demand to be off grid in ways that, you know, it allows innovation because you can manage a grid in a different way if it's a private grid, right? You take your large capacity, it's not part of the the power demand that um regular rate payers are served in, but you can maybe manage power differently if it's in your own area. I mean, utilities historically have been very reticent to pursue innovative policies rightly because they are trying to preserve both reliability and affordability for rate payers but when you're doing your own thing off grid, you're able to do more interesting things that could lead to, you know, improvements on the broader grid if they're proven to work. Um, so I think that's one option, um, as well as the option to put power behind the meter, to, you know, pay for your own power behind the meter and build it that way. Um, I also think that an important element of this is that as you have more power somewhere, as you're using more power, the marginal cost of electricity should go down. In an economy that worked correctly, the marginal cost would go down as demand rose. Something's broken on the power grid that the marginal cost doesn't go down as there's more power. That, the fact that that math doesn't pencil, means that something else is wrong and i think we should look closer at what on the power grid right now both in the rto areas and in the vertically integrated utilities isn't working and is causing cost allocation problems that way um because i think there has been cost allocation problems both with transmission and with the way that power plants are built and mmm amortized overtime sometimes they're prematurely retired and those costs are still borne by rate payers and there's other things in the mix mmm that are making the cost problem worse that aren't just new demand Uh, so I think it's all worth considering.

Thank you. Thank you for that. But I didn't hear permitting reform in that either. Mister Hanna?

Thank you. Uh, we think that grid-connected data centers can be a helpful part of driving electricity costs down for all Americans. There was a Department of Energy study that was published, uh, several months ago that showed that between twenty nineteen and twenty twenty four, the states with the largest degree of electricity load growth saw real price declines uh for residential customers. And the reason for that is you're

Thank you, Mr. Chair. Thank you, Mr. Chair. You're back. You're back.

Thank you, ranking member. And now I recognize Representative Bigich for five minutes.

Thank you, Mr. Chair. Uh, it was great to hear in, uh, the opening remarks the importance of the Speed Act. And I'm, I'm a proud co-sponsor of the Speed Act. Was glad to see that, uh, actually pass the House and get over to the Senate. I'm encouraged, uh, to see, uh, some movement in the Senate on that bill. We all recognize that permitting reform is important. Uh, I'll also, uh, be bringing the House companion for the Data Act uh here in the next few days. So um we've got some really um important work that's happening uh in the house. Uh as the sole member of uh the the delegation from Alaska uh in the house, um it's part of my job is to plug Alaska, we talk about uh the need for uh uh instant power, making sure that we've got vast power resources available for uh AI data centers. trillions of cubic feet of natural gas and I hope uh for those that are listening and those before us today that uh you're uh looking at Alaska as a potential resource for power uh for AI da- data centers and we welcome you uh to our state. First uh question I have for the panel, should data centers supporting AI development be classified as critical infrastructure eligible for expedited permitting similar to national security facilities? We'll start with you uh, Miss Lambermont.

I think that expedited permitting rules of that kind sort of miss the point. The entire system should be as efficient and quick as we can make it. I don't think that by creating carve-outs for different things, especially for national security reasons, I think we end up in a space where the government gets to pick more winners than losers. I think that what's better is to have as efficient of a system as you can possibly have, so that every project can be permitted as quickly as it reasonably should. Um, so I think that what would be better than that would just be to do permitting reform well. Um. And that, yeah, I think we sometimes lose the point when we try to do little carve outs for specific things although I do think it is an important technology.

Thank you.

Thank you.

Mister Hanna?

Thank you. I agree with um Miss - Miss Lambermont. I think that uh uh the most important thing we can do is to unlock capacity for data centers to be able to interconnect to the grid and build with pace. Um and then let the market system uh uh uh work and let us uh connect and find a willing buyer and willing seller to be able to uh bring uh facilities to to s- to pace.

Thank you, and and Doctor Massonet.

Thank you. Uh, permitting is a bit out of my research area and uh, but I will say that when permits are issued, having public information on the types of technologies, the size of projects, uh, technology details, energy sources would be invaluable for helping analysts better understand what's happening on the ground and making better projections about what may happen in the future.

And let let me ask another question.

I think that when it comes to this issue, like the the most effective source in in the in the world for innovation is the American private sector. And anything that we do to, uh, prevent the American private sector from being able to develop this technology puts us at national security risk, and I don't think that the solution to that is government intervention to support se- uh different industries or make things work. I think that the solution is to get out of the way and allow the American private sector to do its thing because it's always innovated incredibly when given the opportunity and I think that's how we went on this.

We agree that uh it's essential for the United States to be able to compete uh with China and strategic competitors around the world that we have a permitting system that is conducive to expedient and timely construction of generation and transmission infrastructure, and that that is a very important part of unlocking, uh, energy security, unlocking energy supply to meet the AI opportunity.

I'll defer on this one again. It's a bit out of my research area, except I will say that, uh, China is investing quite heavily in data centers. We also track data center energy use in other parts of the world, and China is the number two data center hot spot in the world for investments in energy use, uh, uh, and, uh, growing rapidly.

Is there anything you'd like to, uh, to add, uh, with respect to the research that's coming out of China? Uh, what, what the, what those metrics look like in terms of investment?

I don't have numbers at hand, but I'd be happy to follow up. I generally track the size of data centers, uh, the energy use, uh, the locations where we can get data. In fact, it's a bit of a black box studying China from abroad, but I'd be happy to share what intelligence I have, uh, uh, at a later time.

That would be great. You know, I I think we can all agree that uh artificial intelligence uh i is growing dramatically. Uh it's growing at an accelerating rate. Uh we recognize that this is both a threat and an opportunity uh that we need to uh be cognizant of and we need to fully embrace in order to ensure that we win the AI race. It's incumbent on this Congress to ensure uh that we are getting government out of the way and allowing the private sector to flourish. And uh certainly that's my objective and uh I see my time's expired and with that I yield back.

Thank you, Mister Biggits. Now I recognize Miss Lofgren for five minutes.

Uh thank you, Mister Chairman, this is uh an important uh hearing and uh you know none of us, I think, believes that just having a lengthy permitting process is a valuable thing. I mean it's getting the data, getting the decisions in a prompt way and that is fully informed is kind of what we want to have, but that gets to uh the issues that you've raised, Doctor Massonet, because uh as you've uh mentioned in your statement um we don't have good data and you indicate that the only major sector of the US economy that lacks detailed high quality energy and resource utilization data is the data centers. Um, why is the data center uh, subject to such unique data gaps? How did this situation arise?

I think there are two parts to my answer. Th- thank you, Congresswoman, for the question. Um, partly, uh, you know, data centers pre-AI boom, they've always been important, uh, all around the world,

Mm-hmm.

but their national energy use and water use, uh, were - were relatively small compared to other sectors. But that certainly changed post-AI boom. The data centers that we're building now are larger in size, much larger in power capacity than previous data centers. So I would argue that data centers, given the amount of investments, the size of the data centers, uh the amount of power they're projected to require, has really catapulted them into major status. The second part to the answer is uh because data centers have been relatively small, in terms of those those resource u uses in the past, there hasn't been a lot of pressure until recently for data centers to be more transparent and the companies who operate them. Uh, this is a sector that also has a lot of proprietary concerns when revealing data, so in my view, having studied them for twenty years, they tend to release few data and only when, when uh they really need to release data, they, they release as few as possible. That said, there are some companies that are better than others that do release detailed data on their data sets.

Uh, uh, uh, let me ask you this. Um, what federal agencies would be best equipped to play a leadership role in collecting standardizing and publishing data sector information.

As I mentioned in my testimony, uh, there are two major agencies, the first is the Energy Information Administration, which is probably the one I would recommend, just because that administration has been collecting energy and resource utilida- utilization data for every other major energy-consuming sector, power, manufacturing, transport, buildings, for many years. They know how to design surveys, they know how to protect confidential information,

Okay.

and they know how to to get it done quickly.

Let me um You know, without disclosing a private conversation, I uh recently did have a conversation with the CEO of someone who was uh playing a major role in AI. And what we discussed was that we're gonna hit a wall in terms of energy supply. If you take a look at the growth of AI, we got an energy supply problem. Mister Han, I know that it's been reported that Google has made a major investment in commonwealth fusion and they are actually buying up spots for their fusion energy uh companies. I know that hellion uh has their less um prone to send out press releases, but they are also acquiring uh footprints to put it on the grid. But obviously if there can be a co-location of a fusion plant, with these data centers, it solves a lot of the bothersome problems for communities, the supply as well, um and it it addresses the issue which is the elephant in the room, which is we don't have an adequate energy supply to really meet the growth in this sector. What can you tell us about those investments?

So we are committed to partnering with innovative energy companies to scale up the next generation of energy technologies. As you mentioned, Congresswoman, we have a partnership with Commonwealth We also have a partnership with Kairos Power to accelerate, uh, order book of small modular nuclear reactors. Uh, we have a partnership with Fervo Energy to bring on enhanced geothermal. We think it's very important that as we grow, we bring on new firm capacity to meet firm load needs. With regard to co-location, uh, we think this is an important speed to power solution, and that as we look to lead the world in AI, we do need the ability to pursue uh, structures where we are citing a generation resource next to a data center. In those cases, our preference is that that resource be front of meter, so it can be a a contributive part uh, of providing energy to the grid as well as to the data center. But ultimately, this is a higher cost solution for American industry, um, and it uh, it doesn't allow us to help be a constructive part of lowering costs of the grid by paying for uh trans- a transmission system if we are grid connected.

I see my time has expired. I yield back, Mr. Chairman.

Thank you, ranking member. And now I recognize one of two people that I really uh think are leading in in uh knowledge. Well, and I'll give credit to both sides on this, by the way, uh, Mister Olbermulte, who's one of the more uh interesting people in education. And Mister Foster on the other side I wou- I think are are two of the experts in this, in arena. Uh, but I open that up uh for five minutes for Mister Olbermulte.

Oh, thank you, Mister Chair, and thank you to our witnesses. for being here to talk about this critically important topic. I'd like to continue the questioning narrative that ranking member Lofgren had started there on co-location. So, uh, Mister Hanna, uh, you're probably the expert on this. W- there are two competing narratives. You're hearing over and over again in this hearing the, uh, concern that consumers have that data center and the electric load that it brings onto the grid, uh, w- will drive up the cost of energy for them. Yeah, and that's a reasonable fear. Um, the competing narrative is the one that we're discussing on co-location where if you're going to construct a new data center, the number one thing you need access to is, uh, high quality, reliable base load. If you don't have that, you can't run your data center and that's a huge investment. So you're either going to co-locate generation when you build your data center, or you need to go to a local provider and contract to have them build some new generation so that you have access to that energy. Uh, and you're going to size that generation, whatever you do, to meet the maximum amount of energy that you might need. So, uh, in practice you're probably over-generating and then you sell the excess energy back onto the grid which actually ironically drives down costs for consumers. So those are the two narratives. Where is reality in between those narratives?

So our our belief is that data centers can play a constructive part in lowering costs by being grid-connected. But it is important that we achieve speed to power. We are in a race uh with with strategic competitors to build capacity, um and we need to be able to bring on power as quickly as possible. Um we announced earlier today uh an investment in Texas. We'll be building a new uh data center there which will have uh co-located energy solutions. Again our our hope is that uh these structures are front of meter so we can dispatch into the grid and help reduce costs for everyone but over the long term uh you know the the challenge with co-location is you require a significant uh extra back-up capacity so a fully islanded system requires a hundred percent back-up capacity by comparison a grid connected system requires only twenty to ten to twenty percent back-up reserve capacity uh so for american industry do to be competitive over the long term for American industry to help, uh, uh, pay for a transmission system that is, uh, increases reliability for all Americans. We need the ability to connect to the grid and that requires transmission permitting reform.

Right. Uh, yeah, I mean, I, to me that's a separate issue. Right, but, uh, you know, we're talking about transmission, we're talking about generation. I mean, I guess you can argue that to the extent you're not co-locating, you need the transmission to move energy from one place to another. But, um, you know, I, I think generation is gonna be a key component of that and Unfortunately, I'm I'm I'm a co-chair of the Fusion Energy Caucus. I'm very optimistic about that. It's not going to happen, however, soon enough to, uh, to save us, uh, in this instance. Uh, Miss Lambermont, so, uh, you know, talking about permitting reform, uh, and I, I'm complete agreement that this is something that we, uh, we need to be pursuing. Uh, you mentioned kind of in a passing way in your testimony, the burden that overlapping federal and state regulations play in the difficulty in getting new transmission done. And I have an interesting perspective on that because bef- I serve currently, in addition to this committee, I I serve on the Energy and Commerce Committee here in Congress and when I served in, uh, the state legislature I served on the Utilities and Commerce Committee in California. So, uh, you know, I ca- I can vouch for the fact that we have, you know, this this maze of competing regulations. Can you talk a little bit more about how those overlapping layers of regulation extend timelines and whether or not that's actually a good thing for environmental protection?

Yes, of course. Um, so, when it comes to NEPA, the - the running joke is always the process is punishment. Like, it - essentially, the process is drawn out as long on purpose to make it hard to build something. But at the same time -

I'm not sure if your mike is on.

Oh. I'm so sorry.

There we go.

Um, at the same time - the - the NEPA process overlaps with all of these other processes. So once you've already done the NEPA review, you come upon an environmental species act. or an endangered species act review that has to happen concurrently. And then there's something with the clean water act that has to happen concurrently, and something with the clean air act that has to happen concurrently. And at a certain point you've studied four separate things on overlapping timelines that have triggered additional reviews and additional reviews rather than looking at the problem holistically and studying it quickly you're looking at all of these different branching pathways that require incredible amounts of money and man hours and delay, add delay upon delay. And then when you're in a state that has its own NEPA, the NEPA process has to happen again in this other way, and it just becomes this sort of overburdensome process that layers upon itself. And I think that we could just do it in such a more efficient way by looking at the problem holistically and saying " OK what do we actually need to cover here?"

Right, I agree. Well, thank you very much for your testimony.

Thank you.

Uh, you're back.

Thank you, Mister Obnolte, and now I recognize Miss Bonamici for five minutes.

Well, thank you to the Chairman and RANKI member. Thank you to the witnesses. Uh, a couple comments uh at the outset.

I heard the word permitting a lot in the testimony, in the opening statements, and I'll note that this committee does not have jurisdiction over permitting. That happens in energy and commerce, transportation and infrastructure. I'm glad we had to have some overlap. It did remind me that a former chairman of this committee once said you could get anything in the jurisdiction of the science committee by putting research about in front of it so uh that reminded me that uh you know in hearing some of these sort of anti-government get the government out of the way comments, I just want to note for the record that uh robust investment in basic research by the federal government does help private industry, so I just want to put that on the record. Uh the the other thing I I want to note is that I sure wish we saw as much enthusiasm for grid modernization for integrating renewable energy, as we're seeing for data centers. I I I want to note that this is something that we've been talking about for a long time, uh the importance of grid modernization. So I I do understand that we have a goal. for America to lead the world in AI. But I hear from working families right now that they can't afford to pay their energy bills. They're especially frustrated if they think they're going to be seeing steep increases uh to let profitable companies exploit the already strained grid. We know data centers are now driving one of the fastest surges in energy demand in modern history, and many of our constituents are feeling the pain. It was interesting to hear uh Mister Obernalte talk- talking about how this could actually lower costs. Uh, I think that remains to be seen. A recent analysis by Bloomberg showed that near some major data center clusters, electricity prices have skyrocketed as much as two hundred and sixty-seven percent. So in response, many states are modeling solutions to this unsustainable problem. So in my home state of Oregon, we passed the Power Act, we, the state legislature, uh, where I used to serve, passed the Power Act. Uh, and this creates a new classification for large load energy users like data centers. It requires them to pay the cost of transmission generation and grid, uh, upgrades, uh, that they require for their operations and it also requires a ten year contract. There's some serious concerns about, uh, in making these huge investments and then a couple years down the road, uh, if, if, if they're abandoned, that's, that's a huge, uh, uh, barrier in cost. So if companies wanna build hyper-scale facilities that consume massive amounts of power they need to take responsibility for the cost. So I wanna ask you, Mister Hanna, As I mentioned, Oregon's Power Act requires large load cons- customers including data centers to fund the grid upgrade, uh, costs that they trigger. So why is requiring these hyper-scale operators to pay for transmission generation and storage infrastructure good policy? And what other state policy examples should Congress look at and consider for a framework to address AI driven energy demands?

Thank you, Congresswoman. Uh, I wanna be very clear, this is this is a very important issue and we know that, uh, uh, the - the electricity prices that Americans are facing is - is very important, that, uh, as - as a company like Google we ta- we address this topic. And I wanna be very clear, we pay our way. At Google we pay our way. And that means we pay for a hundred percent of the electricity that we use to power our data centers. And it means we structure our contracts so that we are paying for, uh, the incremental, uh, infrastructure that is needed to power, um, our facility further up the grid. That's something we're very committed to. uh, no families or small businesses should see, uh, should be bearing the costs of, uh, data center, uh, growth and infrastructure. Um, and we're committed to - to, uh, doing this. So we've now - we have structured, uh, contracts through what we're calling our capacity commitment framework, in a number of states where we're paying, uh, upfront minimum charges, long-term contracts, uh, upfront collateral to ensure that as we grow, we are sure that, uh, that there is capital there uh to protect other rate payers from rate from rate increases. That's something we're very committed to.

Thank you for modeling that kind of uh r- responsibility. Professor, is Manusnet, did I say, say it correctly? Close, was I close?

Manusnet, thank you.

Manusnet. Uh, you know, utilities have to plan years in advance, but data center demand projections are swinging wildly, depending on AI growth assumptions and opaque sort of industry disclosures. So what financial and reliability risks do rate payers face if utilities overbuild? uh, generation and transmission based on inflatable forecast, inflated forecast. And what safeguard should regulators adopt now to prevent stranded cost, uh, cost and also grid instability?

Uh, thank you, Congresswoman. I'll just briefly say that better data, uh, better visibility on, you know, the justifications for new capacity additions and transmission upgrades, those should be based on a common view of, of the data center sector. So a lot of the estimates that we see are based on sort of bespoke data sets, that different analysts have to come up with on their own and we we lack a common view of what's happening on the ground. I think having that data foundation would reduce uncertainties and inject a lot more uh transparency into those forecasts.

Thank you. And as I yield back, I do wanna note, I had a bipartisan bill in the last Congress to deal with security risks. Um, it wou- to the grid, and I hope we can take that up again, uh, Mister Chairman and Ranking Member. And I yield back. Thank you.

Thank you, Miss Bonam- Meachie. And now I, uh, recognize myself for five minutes. Uh, first of all, it's rare when I ever could say that I'm disappointed as a whole with our our esteemed panel, but uh you were asked a question, a very simple question I thought by the ranking member of how many programs were delayed, by regulatory burden, to which there was no answer. And I was like, it's an easy answer, a hundred percent. All of them. In my opening statement, if you were listening, I kind of hurt, my feelings are a little bit hurt right now. I stated, a quarter of them are denied, all of them are delayed by an average of about twenty-six months, by review processes, and then a hundred percent of them are challenged in court, which delays them another two to ten years. There's your answer in my opening statement. So, I just wanna, I wanted to kinda clarify that for, for the, for the record. Uh, Miss Lamberton, Lamberfot, I'm sorry, I did it twice. Uh, the Biden administration issued numerous executive orders related to climate, environmental justice, equity, infrastructure, decisions on all capacities. In your assessment, did these policies improve environmental outcomes? Or do they just primarily add to procedural complexities?

I think that many of the executive orders, as well as a lot of the in- environmental laws passed during that period of time, rather than aiding in, you know, bringing more power to the fore in an efficient way that didn't, you know, harm harm consumers or the environment, instead just slowed things down. Um, I mean, especially some of the rulemaking processes that we saw. Uh, I mean, the reason that we don't have a supply chain for gas turbines is because everyone thought the power plant rule was going into effect until earlier this year.

And there you go.

And

I gotta cut you off because I gotta ask everybody to answer. We have a three mile road in our three quarter of a mile road in my district, in my my neighborhood, it was delayed for eighteen months to cost three quarters of million dollars to determine if it's gonna be repaved or not. That's the burden of regulation. Didn't help anybody, didn't make it cleaner, didn't make it safer, just delayed and made more expensive. Typical. Uh, Mister Hanna. Has Google delayed, realloc reallocated or relocated or cancelled data center projects due to uncertainty in energy infrastructure permitting timelines?

The barriers to bringing on new energy generation and transmission have been a barrier have have delayed our ability to bring on new capacity, yes.

And and I will say this, look at what's happened in Ireland right now, which is the headquarters for Google in Europe, they're facing those same burdens over there as they are here. And Europe's falling way behind right now in the AI race because of the regulatory burden. It's a canary in the coal mine moment for us as, as far as I look at this. Um. Doctor Massonet, I'll save you for last. One of the interesting things that I'm, I'm saying, you're gonna have increased, as you said, huge amount of data centers in California, Silicon Valley, leading in, in industry and technology and all that stuff like that. I learned to surf in San Olefrey. As a Marine, I lived there for seven years. Back in the mid nineties, we had the same challenges in the grid there. They're already way behind and energy consumption is only parabolically shifted to increasing right now. If you look at nine of the ten most unaffordable states in the union are blue states because of regulatory burden because high energy costs, because we're not building the infrastructure. Nine out of ten most affordable states, red states. Why? Because they're not as burdened by by government. When when we look at solutions, at your esteemed California university. If we're gonna have a real discussion about how we produce enough energy for something that is going to consume it, whether you want it or not, the people are demanding it. It is a popular item and it's going to help us leap into the future, to be competitive with the most competitive nations out there, or conversely, fall way behind. I would say we've already put ourselves at a disadvantage. Do you think that nuclear energy is harmful to the environment, or a major solution for the future of this country.

That's a bit out of my area of expertise because I study the evolution of data d- centered demand, uh, so how we use energy, not necessarily the supplies, but I would go back to my central argument that understanding what the options are, having data on those options would allow us to make better decisions about paths forward.

All right. Well, I I'm gonna keep my comments very short, but I just wanna make the the point that a lot of times government is not the solution or the problem. And I don't think this is any different, that we all want a clean environment. We all wanna be responsible. We all wanna pay our bills, make sure that we also, that we're not a burden to society, cuz when we become a burden to society, we are no longer going to be viewed favorably, and that's bad business. As Calvin Coolidge said a hundred years ago, the business of America is business. Cheap, affordable energy, a good grid. That is the future of America and that's what we have to support. And with that, I'm gonna yield early. Imagine that. To Mister Foster, one of my favorite members who's extremely well thought of, and and I hope you don't embarrass me by saying that by saying something, you know, on an inappropriate note.

Yeah, something I exceptionally, yeah.

With that I yield for five minutes.

Uh well, thank you Mister Chair. And then to our witnesses, my apologies for being in and out of this, it's a crazy day in case you haven't intuited that. Um let's uh uh Doctor Manson, um you talked about the importance of getting data collection. Now it um Roughly four percent of electrical demand is AI data centers. Uh, roughly one to two percent is Bitcoin mining. And in on February first, the Trump administration ceased collection of information on the energy consumption by Bitcoin mining. Were you aware of that? Oh, no, no, it's the EIA has a deep dive on discussion of that. Um, and and so it seems like that's a huge step backward. I've been frustrated for a while on this because, you know, it's it's of order. all of the data center um consumption is bitcoin mining. And when Ethereum switched from proof of work to proof of stake, it's um it's power consumption went from very significant to near zero. And there is no technical reason why that couldn't be done with bitcoin. If we did that, that would provide you know today half the energy that you're consuming in the whole. And and I was just wondering is that something that uh there have been serious efforts at trying to make that happen? Does that depend on a a nudge from Congress? Um.

Thank you, Congressman. Uh, crypto currency mining is something that folks like I do study in looking at data center energies and you're right proof of stake uh uh sorry um uh uh proof of work is a very uh energy intensive Bitcoin uh process,

Yep.

but uh that shift is really something that is a bit beyond my my expertise, how to how to facilitate a shift.

How to make the effort.

Yeah. Well, I mean, certainly we have

I can say technically yes, a shift would reduce energy use.

Yeah, we have the, in many ways we have you know, there's a voting mechanism where where Bitcoin and all crypto, they can change their underlying code. In fact, that's was done uh with Ethereum back in roughly twenty twenty two.

Mm-hmm.

And it it worked and it's fine, it's reliable. There's no problem with it, no cyber security problems that I'm aware of. And so it seems like it works and it's a not a small thing. And that might be something where with my other head on in financial services, maybe we can find ways to incentivize the collective voters of Bitcoin to actually change their um, you know, from proof of proof of it work to proof of stake, it would be a huge um, anyway, just a thought. So if there's, if you dig up uh research on you know, how close that i- I believe there were technical proposals that says here's exactly how we will switch Bitcoin. And it just seems like if we could do that, that would at least take a lot of pressure off of of the grid. Um, yeah, um, uh, so, mister mister. Um, the You know, I've I've read various summaries of various, uh, analyses of what's the cheapest way to do this, uh, to provide the energy. And, you know, r- recent experience with US nuclear, it's about ten dollars a watt. The spot price for for Chinese solar cells is eight cents a watt. And, you know, you gotta divide by six for the capacity and and the storage you need, the the battery storage is, um, it's going down really rapidly and it's gonna go down by another factor of roughly two with when sodium batteries replace lithium batteries. And so this is um already the low-cost way of of doing this. I think uh Goldman Sachs had a big report on this not too long ago, where they said the cheapest thing way to do this is just go out in the desert where land is effectively free, put out a bunch of solar cells, get a bunch of um you know, get a bunch of of batteries, even lithium batteries, which is what they were looking at at the time. And then um so the thing is self-contained. Um, in the rare instance when you're in El Paso and there's a snowstorm and, and you just sit there, OK, the, the data center's off the air, you know, I'm sure, like Google you regularly deal with having one of your major data centers off the air. You have fiber to shift the load around. And so it seems like that's a completely workable solution that doesn't involve, you know, jumping on, on grids, you know, putting these in cities where people get angry with them. And so what, um, and Why isn't the market just completely shifting that way? And what what fraction of of what you're installing is in that model? And um where do you see it's gonna be going?

So, uh, thank you for the question. Uh, we know that many different resources are gonna be a very important part of powering, uh, new incremental load growth to meet AI. We are pursuing, uh, solutions like advanced nuclear, advanced geothermal. Um, what's important about those solutions is that they bring twenty four seven, uh, available capacity. Um, we know we're we're focused on

But the, what's the reliability spec that you actually need for a data center? Because, you know, any data center can shift its load elsewhere. You know, just and the fiber, you know, you don't have to negotiate with a lot of angry neighbors to put to put buried fiber in their backyard, the way you do if you put a power line in their backyard. And so, and so, you know, you always design data centers so you can shift the load around, if one goes off for reasons you can't control. And so it So it it really, you know, so I've I my questions are about why the market isn't a hundred percent moving to what appears to be the low cost point.

H- historically we have a very, we require very high levels of reliability, and we do have the ability to shift capacity in in time and space, and that is something we are increasingly doing, particularly as a speed to power solution, to bring on capacity more efficiently.

OK.

And with that, that's the only person I allowed to go over, because I wasn't paying attention. Uh, I thank the witness for their valuable testimony and the members for their questions. The record will remain open for ten days for additional comments and written questions from the members. This hearing is adjourned.