Sam Altman Senate AI Hearing — Opening Statements, Key Testimony & Exchanges

Chairman Ted Cruz — verbatim remarksGood morning. The Senate Committee on Commerce, Science, and Transportation is called to order. Welcome to our witnesses. Thank you for joining us this morning.

Chairman Ted Cruz — verbatim remarksIn the last two years, AI has brought the United States and the world to a critical inflection point. AI may be a technology as transformative as the internet or even more so. It has unleashed a new global industrial revolution with the potential to unlock opportunities that improve our quality of life, create jobs, and stimulate economic growth. The country that leads in AI will shape the 21st century global order.

Chairman Ted Cruz — verbatim remarksAs a matter of economic security, as a matter of national security, America has to beat China in the AI race. China has made AI central to its national strategy. And China aims to lead the world in AI by 2030, investing heavily in AI adoption across industries like manufacturing and defense. In this race, the United States is facing a fork in the road.

Chairman Ted Cruz — verbatim remarksDo we go down the path that embraces our history of entrepreneurial freedom and technological innovation or do we adopt the command and control policies of Europe? I would suggest that Congress draw on the lessons we can learn from the dawn of the internet. In the early 1990s, Washington embraced the internet and explicitly adopted a style of regulation that was intentionally and decisively light touch. Congress chose to deregulate under the Telecommunications Act of 1996 while President Clinton pursued tariff agreements and treaties that protected America's intellectual property and technological exports.

Chairman Ted Cruz — verbatim remarksFurther, in 1998, Congress enacted a 10-year internet tax moratorium so that state laws wouldn't balkanize and stymie the promise of e-commerce. The results of these decisions were extraordinary. I made a lot of money. By 2000, the United States has recorded five straight years of historic highs in productivity gains and investment growth.

Chairman Ted Cruz — verbatim remarksHundreds of thousands of new jobs were created and the United States became a top tech exporter with massive sums of private investment pouring into the US digital economy. By contrast, EU countries pursued a series of heavy-handed regulations that proved enormously costly. In 1993, the United States and Europe had economies virtually identical in size. Today, the American economy is more than 50% larger than Europe's.

Chairman Ted Cruz — verbatim remarksThe drivers of that are tech and the shale revolution. Those two comprise virtually the entirety of that massive growth o over Europe. According to one EU Commission report, only 6% of global AI startup funding flows to EU firms. 6%. That is onetenth of the amount that is going to American companies.

Chairman Ted Cruz — verbatim remarksThe report directly blames this yawning chasm on the EU's nasty regulatory approach. And yet, the Biden administration, for inexplicable reasons, tried to align AI policy with the EU to adopt their failed policies. President Biden's sweeping AI executive order, the longest executive order in American history, cast AI as dangerous and opaque, laying the groundwork for audits, for risk assessments, and regulatory certifications. Biden's approach inspired similar efforts in state legislatores across the country, threatening to burden startups, developers, and AI users with heavy compliance costs.

Chairman Ted Cruz — verbatim remarksSome of my colleagues suggest that a friendlier version of the Biden approach makes sense. They want a testing regime to guard against AI quote discrimination and have government agencies provide quote guidance documents seemingly something out of Orwell that will usher in what they call best practices as if AI engineers lack the intelligence to res responsibly build AI without the bureaucrats. Many in the industry foolishly have supported such paternalism. Harmful regulations take many forms.

Chairman Ted Cruz — verbatim remarksBiden's misguided midnight AI diffusion rule on chips and model weights would have crippled American tech companies ability to sell AI to the world. The Biden plan would have handed over key markets to China. We should want foreign countries, particularly our our allies, to buy American. I vocally opposed this rule for for months and indeed the ranking member and I together urged the Biden administration not to adopt it and I'm very pleased that President Trump has now confirmed he plans to rescend it.

Chairman Ted Cruz — verbatim remarksAll of this busybody bureaucracy, whether Biden's industrial policy on chip exports or industry and regulator approved guidance documents, is a wolf in sheep's clothing. To lead an AI, the United States cannot allow regulation, even the supposedly benign kind, to choke innovation or adoption. American dominance in AI depends on two factors. Innovation and adoption.

Chairman Ted Cruz — verbatim remarksInnovation drives breakthroughs in global competitiveness. Adoption ensures that these tools empower American workers and businesses, enabling the United States to become the world's leading adopter and exporter of AI. Thankfully, President Trump has largely reversed Biden's misguided AI agenda. In fact, I think AI was a sleeper issue in this last elections.

Chairman Ted Cruz — verbatim remarksAmericans wanted to see President Trump and Republicans and indeed all senators champion AI policies focused on innovation and adoption. The contrast has been astounding. This year, there have been over $1 trillion of new AI projects, including major investments in Texas, like the coreweave data center in Plano and the $500 billion project Stargate in Abilene by Open AI and Oracle and others. Adopting a light touch regulatory style for AI will require Congress to work alongside the president, just as Congress did with President Clinton.

Chairman Ted Cruz — verbatim remarksWe need to advance legislation that promotes long-term AI growth and innovation. That's why I will soon release a new bill that creates a regulatory sandbox for AI modeled on the approach taken by Congress and President Clinton at the dawn of the internet that will remove barriers to AI adoption, pres prevent needless state overregulation, and allow the AI supply chain to rapidly grow here in the United States. That's how we'll accelerate economic growth, secure US dominance in AI, and beat China. And with that, I turn to the rank ranking member Cantwell.

Ranking Member Maria Cantwell — verbatim remarksThank you, Mr. Chairman. Uh, thank you for this hearing, and welcome to the witnesses before us, Mr. Altman, Dr.

Ranking Member Maria Cantwell — verbatim remarksSu, Mr. Intrator, and Mr. Smith. It's a great pleasure to have all of you here, but it's especially prideful moment for the Pacific Northwest to have Mr.

Ranking Member Maria Cantwell — verbatim remarksSmith and Mr. Altman here both representing an open AI approach. By that I mean a approach where we want to win against China and a closed system by making sure that what is developed here in the United States and around the globe is an architecture where the United States wins and is open. To do that, we need to focus in winning on computing power, on algorithms, and on robust data sources.

Ranking Member Maria Cantwell — verbatim remarksAll of that will be key. Personally, I believe a continued investment in NSF helps in all of those areas as a good public private partnership with the industry that's represented here today. I'm so proud that we passed the Chips and Science Act because the Chips and Science Act also set a foundation for investing in the United States of America and bringing more of the supply chain back to the United States of America to build on a future leadership that we already have I believe in the computing power. But we also need to understand that we have to move forward on the chips act like the University of Washington $10 million grant on multi-design sets for chips.

Ranking Member Maria Cantwell — verbatim remarksThe very largecale integrated designs I'm sure that Dr. Su will tell us about today. But the fact that the United States has to continue to lead on the future designs and the implementation of that also requires us to be very smart about data centers, about sources of electricity and how we're going to build that supply that could be up to 12% of electricity demand in the very near future. So, how do we do that?

Ranking Member Maria Cantwell — verbatim remarksI've noticed in each of your testimonies, you all explained this, but I'm also very proud that Microsoft has already signed an agreement with one company, Helion Energy, in Everett, Washington for a power source supply. Maybe Mr. Altman in his testimony will talk about this, but that they hope to get very near future energy source from that. So clearly the United States leading on electricity and development.

Ranking Member Maria Cantwell — verbatim remarksSo, Mr. Smith, very much appreciate in your testimony the accentuation on the fact that the United States of America needs hundreds of thousands of new electricians. Something we should all want to get behind. The fact that having electricity and the electricians and the data source centers here in the United States and in other places will be key.

Ranking Member Maria Cantwell — verbatim remarksWhile I want to see us move forward, as the uh chairman said, we signed a letter saying we needed a broader support for export controls. I want to be clear, export controls are not a trade strategy. They are not a backpocket issue that the president of the United States whips out in trade negotiations. We are going to move fast because we are going to set standards.

Ranking Member Maria Cantwell — verbatim remarksI believe those standards should be encouraging very broad distribution of US manufactured and made AI chips and technology and that we're asking our partners overseas to comply with the rules that we establish. Things like making sure that there is no circumvention of the supply that somehow gets into China's hands. making sure that we have access and making sure that we can verify on that and also making sure that US data companies and cloud-based companies are allowed to be in that market. We should not be going to markets overseas only to have them tell us that organizations with cloud services from the US would not be allowed. This I believe would be a robust uh initiative on getting US AI chips and US AI open systems dominated around the globe.

Ranking Member Maria Cantwell — verbatim remarksWhy do we need to move fast? We need to move fast because if we don't, we are looking at another Huawei, another instance where the United States is behind and also saying we should tear out this system that now we don't like for lots of reasons and backdoor policies. So, I'm all for winning. That is why we passed the Chips and Science Act.

Ranking Member Maria Cantwell — verbatim remarksI'm all for winning and that is why we have passed seven bills out of this committee last year that kind of got stuck in the lame duck. I think the chairman of the committee wasn't ready to move forward in negotiations with the House and Senate on those seven bills. But those bills, a bill between myself and Senator Young on the uh Institute for uh standards, NIS standards, which I think we still need to do, my colleague and I, Moran, on education and scholarships, small business, and uh the bill by my colleague here, Senators Clolobachar and Thun, which was also related to the Nest standards. So, we had an opportunity a year ago to move fast. we didn't do it.

Ranking Member Maria Cantwell — verbatim remarksSo, let's do this now. Let's get together and figure this out. The faster the United States moves now. I like this great uh Paul Romer quote which was about collaboration is the next phase of innovation.

Ranking Member Maria Cantwell — verbatim remarksIf we don't collaborate here, if we throw down on politics instead of getting the policy right, we won't move fast. Let's allow these people to do what they do best and let's make sure the United States has the right policies in place so that our open AI standard wins the day. Thank you, Mr. Chairman.

Thank you. Uh I'd now like to introduce our witnesses for today. Each of our witnesses and their companies represent critical parts of the AI infrastructure, hardware, and software supply chain. Our first witness is Sam Altman, the co-founder and CEO of OpenAI.

OpenAI is one of the world's most advanced AI companies known best for its chat GPT product. Our second witness is Lisa Su, the chair and CEO of Advanced Micro Devices AMD. AMD develops high performance processors, graphic chips, and AI accelerators that power artificial intelligence. And Dr.

Su is also a Texan. Our third witness today is Michael Intrator, the CEO and co-founder of CoreWeave, an AI hyperscaler. CoreWeave is the world's largest purpose-built AI cloud platform. And our final witness is Brad Smith, the vice chair and president of Microsoft.

I believe everyone is familiar with his company. Uh Mr. Altman, you are recognized for your opening statement.

Sam Altman, OpenAI — remarks paraphrasedOpening statement summary — Sam Altman, co-founder and CEO, OpenAI. Altman said that when he last testified about two years earlier, ChatGPT was "a curiosity," and that it is now used by more than 500 million people a week and ranks among the largest websites in the world, materially raising productivity for scientists and other users. He said investment in infrastructure is critical, that "the next decade will be about abundant intelligence and abundant energy," and described visiting the Abilene, Texas site he called the largest AI training facility in the world. Deviating from his written testimony, he told the story of growing up in St.

Louis as a computer nerd during the internet boom, drawing "a straight line" from that experience to founding OpenAI, and said he did not think the internet — or the AI revolution after it — could have happened anywhere but America, which is why US systems and policy must keep that pattern going.

Lisa Su, AMD — remarks paraphrasedOpening statement summary — Lisa Su, chair and CEO, AMD.

Su said AMD is a US-headquartered semiconductor company founded 56 years earlier that builds high-performance computing chips, that its work with the Department of Energy powers the two fastest supercomputers in the world (at Oak Ridge and Lawrence Livermore), and that AI is "the most transformative technology of our time" but that US leadership "is absolutely not guaranteed." She laid out five priorities: run faster by keeping compute available; protect open ecosystems as "a cornerstone of US leadership"; build a robust domestic supply chain with more US manufacturing; invest in talent so the US is "the best place … to study AI, to work in AI"; and, on export controls, pair national-security protections with widespread global adoption of US technology,

warning that if US technology is not adopted abroad "there will be other technologies that will come to play."

Michael Intrator, CoreWeave — remarks paraphrasedOpening statement summary — Michael Intrator, co-founder and CEO, CoreWeave. Intrator said CoreWeave began "in a garage" doing GPU work for cryptocurrency mining, pivoted to AI infrastructure, and now operates more than 30 data centers across 15 states with more than 250,000 GPUs drawing 360 megawatts, with revenue up roughly 12,000% over two years and a March 2025 IPO. He said modern AI needs purpose-built infrastructure — "we cannot run a 21st century economy on the 20th century's infrastructure" — and that since 2018 the compute needed for advanced models has multiplied roughly 100,000-fold.

He asked policymakers to focus on four things: strategic investment stability with consistent rules across jurisdictions; energy-infrastructure development with permitting reform; global market access calibrated to national-security risk; and public-private partnerships and workforce development, citing CoreWeave's co-founding of the New Jersey AI Hub with Microsoft and Princeton.

Brad Smith, Microsoft — remarks paraphrasedOpening statement summary — Brad Smith, vice chair and president, Microsoft. Smith said the AI tech stack means "we're all in this together" — infrastructure, platforms and applications — and that US success depends on three things: innovation (more infrastructure, faster permitting, more electricians, university and federal basic research); faster adoption, or "AI diffusion," whose "number one ingredient … is skilling"; and exporting, since only about 4.5% of the world's people live in the United States.

He framed the core question as whether the industry is "trying to build machines that are better than people" or "machines that will help people become better," said it "needs to be the latter," and closed by urging investment in education and skills so that any kid "with a computer … and now the ability to put AI to work" gets the opportunity to try.

Chairman Ted Cruz — "five words to win the race" and incentivizing US investment. Cruz asked each witness for "the five words you need to see from our government to make sure we win this AI race." Smith answered "more electricians," "broader AI education," and "no using chat GPT as a phone a friend"; Altman said streamline the ability to build large things, run faster in the innovation race, allow supply chain, sensible policy.

Cruz compared the moment to past races the US won by combining a government effort with universities — nuclear energy, space, aviation — and noted AI is unusual because "a non-state actor could just as easily win this race." Altman cited the $500 billion Stargate project, the need for certainty to build the supply chain and permit electricity, keeping model creation and chip, network and server-rack production in the US, high-skilled immigration, and legal clarity for firms like OpenAI.

Su said compute is the foundation and urged simple, enforceable export rules plus domestic manufacturing and a strong talent base; asked whether firms weigh the US against China, she said countries hungry for AI "will choose what is available, and if China is available that will … be a[n] outcome that we would not like to see."

Sen. Maria Cantwell — NIST standards. Cantwell asked, "Do we need uh NIS[T] to set standards?" Going down the panel, Altman said "I don't think we need it. It can be helpful," while Su, Intrator and Smith each said "Yes." Smith clarified that "NIST is where standards go to be adopted, but it's not necessarily where they first go to be created" — they come from industry input — and that the US will need industry standards, American adoption of them, and US effort to "ensure that the world buys into these standards."

Sen. Jon Husted — child safety and AI companionship. Husted asked whether the internet did "a good job … of protecting children" (Altman: "not particularly") and how to protect them now. Altman said OpenAI wants to "treat our adult users like adults" with "a much higher level of protection" for children, that it cannot yet reliably tell whether a user is a child or an adult, and offered to have OpenAI's teams work with the senator's "to make certain that we put together the right framework early on." Altman added that people are relying on AI "more and more for life advice, sort of emotional support," a newer phenomenon "we have to … understand … and watch … very carefully," and agreed he would not want a child's best friend to be an AI bot.

Husted also asked Intrator about stablecoins, payments and AI; Intrator said stablecoins, crypto and AI "share certain DNA … attempts to build into a future where new technology will make things better," with combined potential.

Sen. Amy Klobuchar — risk-based regulation, hallucinations, the Take It Down Act, news content. Klobuchar quoted David Brooks that it is "literally unknowable whether this technology is leading us to heaven or hell," described her bipartisan bill with Sen. Thune for guardrails on the riskiest non-defense AI uses, and asked Altman whether a risk-based approach is best (Altman: "I do").

On reliability she cited a New York Times report that hallucinations are getting worse; Altman said that "on the whole … AI hallucinations are getting much better," that ChatGPT now cites sources and the models are smarter, and that "users are smart" about the systems' limits. On the Take It Down Act — headed to the president's desk after "over 20 suicides in one year of young people" tied to sextortion and fake images — Smith said the industry works to identify AI-generated and harmful content and collaborates with groups "like NCMEC" on takedowns.

On news content, Smith said newspapers should be able to "get together and negotiate collectively," that a new copyright debate will be settled "some … by Congress, some by the courts" and some by collaboration, and that the goal is to "strike a balance" between paying creators and giving AI access to data.

Chairman Ted Cruz — who is winning, and the danger of EU-style pre-approval. Cruz invoked Yale professor Grant Gilmore — "in heaven there is no law … In hell there is nothing but law and due process is meticulously observed" — and asked each witness who is winning, the US or China.

Altman said American models, including OpenAI's and Google's, "are the best models in the world," the lead is "not a huge amount of time," and the US should "keep doing the things that have worked … and not … make a silly mistake." Su said the US is "ahead in chips today" with "the best AI accelerators," but China is "certainly catching up because there are many ways to do things." Intrator said the US is ahead but that building "very large solutions to the computing infrastructure" and delivering the power is "the Achilles heel." Smith said the US has a lead in "a close race … that will likely remain close," that the number-one factor is "whose technology is

most broadly adopted in the rest of the world" (18% of people live in China, 4% in the US, "78% live somewhere else"), and that the US must export "with the right kinds of controls," win global trust, and build financial architecture reaching Africa.

Asked how harmful an EU-style "heavy-handed prior approval" regime would be, Altman said "that would be disastrous," explaining the "three key inputs … compute … algorithms … and … data": rules that cut the US off from competitive training data, domestic chip manufacturing, or the ability to ship the products people want would make things "fall apart," and he said he is "nervous about standards being set too early."

Sen. Brian Schatz — self-regulation, AI-content labeling, paying for training data. Schatz asked whether "self-regulation is sufficient at the current moment"; Altman said "No … some policy is good," while warning "it is easy for it to go too far." Schatz asked whether an AI-content labeling regime — disclosure, not prohibition, especially for images, music and audio — would help consumers; Smith pointed to the "C2PA" standard and "content credentials" that record where and by whom something was created.

On training data, Schatz argued "data really is intellectual property … human creativity" and "we got to pay people for their knowledge"; Smith said Microsoft "create[s] intellectual property" and "respect[s]" it, that copyright protects expression but not ideas — remembering that "Shakespeare wrote a story about two teenagers who fell in love … that's fair use" — and that creators should be rewarded.

Sen. Ted Budd — AI and government service delivery. Budd asked Altman and Smith to "paint a picture" of AI reducing the friction of dealing with government. Altman described "a future where the US government offers a … AI powered service that makes it really easy to use all government services," file taxes or appeal a decision "instantly" through "an agent in your … pocket fully integrated with the United States government." Smith said "Remember when you had to stand in line to renew your driver's license?" and noted such services already exist "in Abu Dhabi" and should be brought to America.

Sen. John Curtis — permitting and energy, open-source models, data-center siting, ratepayers, small business. On permitting and power, Intrator said access to "scale power" is "one of the keys to our ability to win this race" and that once the existing grid's "elasticity" is consumed, the industry hits "a first principle … [h]ow do we get power … online now" inside a regulatory environment that makes that "really … challenging." Smith said Microsoft has "872 applications in more than 40 countries" and that the number-one US bottleneck is "the federal wetlands permit … administered by the Army Corps of Engineers" (18–24 months, versus 6–9 months for local and state permits).

Altman said US leadership matters in both open and closed models, that OpenAI would "release an open-source model … this summer" so people "build on the US stack," and that "the cost of intelligence … will converge to the cost of energy," making energy the most important long-term strategic investment. In a later round Curtis asked what makes a state attractive for a project like Stargate; Altman cited "power, cooling, fast permitting … labor force that can build these things," praising Texas. On protecting ratepayers Altman said "the best way is just much more supply," which should lower rates "for everything." On small business Altman told the story of an Uber driver running "basically his entire" laundromat business on ChatGPT — ads, legal-document review, customer support.

Sen. Andy Kim — "what is the race," and talent development.

Kim pressed Smith on his framing that "the race … [is] about adoption in the rest of the world, the 78%." Smith said it "shines light on the importance of getting it right for export controls" and urged getting rid of "the quantitative caps … for all of these tier 2 countries" that told "120 nations they couldn't necessarily count on us." Altman said the stack is "heavily interrelated" — "the chips at the bottom to the applications on the top" — and "we want the whole world on the US stack," warning that a stack the US does not trust could carry "back doors" or "data corruption." Kim asked Su to grade US talent development; Su said "the smartest

engineers are in the United States" but the country needs "more hardware developers, software developers," and the best international students, and that "high-skilled immigration is one of those areas where we want the best people in the world."

Sen. Eric Schmitt — Europe's regulation and online speech. Schmitt, a fellow St. Louis native, asked "what specifically has gone wrong in Europe." Altman cited the running "in joke" that a new feature or model is "not available in the EU" because of "this long process … before a model can go out," leaving safe, robust models the company "will be unable to offer" and users "consistently some number of months behind." On Europeans "being arrested for things that they say online," Intrator said CoreWeave sits "kind of below" the content layer and is "really focused on the investment side," and that Europe's "bulcanization," with "each component of their … union having its own set of rules," makes large investments hard.

Altman said jailing people for online speech is "very not American," that AI is used "one-on-one" rather than as "this massive thing online" so social-media analogies are "a little bit dangerous," and that adult users should get "a lot of freedom" with guardrails "in very wide bounds."

Sen. John Hickenlooper — Copilot testing, external evaluations, chip efficiency.

Hickenlooper, a former geologist, asked Smith about Microsoft's internal processes for evaluating Copilot's accuracy before release; Smith described OpenAI's internal process, a joint "DSB, a deployment safety board," Microsoft's own applications-level deployment safety board, engineering assessment tools, and "red teams … that often go to work to … attack the features." Hickenlooper asked Altman whether "independent evaluations based on … standards performed by qualified … evaluators … done voluntarily" could help validate internal testing; Altman said "external testers and red teamers … are a critical part of that process" and "have helped us find many things in the models to improve," while saying OpenAI is "very proud of our safety record on the whole." Hickenlooper asked Su about the next

frontier of chip energy efficiency and "direct to chip cooling"; Su said the sector is innovating heavily, that AMD uses AI "extensively through our chip development … cycles" to shorten multi-year timelines and improve efficiency, and that cooling technologies are "super important … to build the large scale systems."

Sen. Tammy Duckworth — national labs and basic research.

Duckworth said Illinois is home to Fermilab and Argonne (the "Aurora Supercomputer") and criticized "Donald Trump and Elon Musk, with the support of some Republicans in Congress" for moving "to take a chainsaw to … vital research," asking whether "Doge … decades ago" would not have cut "the project that created the internet." Altman said AI's "impact … on scientific discovery" is what he is "personally most excited about," that AI tools could let science "accomplish … a decade worth of scientific progress in a year," and that the national labs are "the only partnership where we've given a copy of our model weights to another organization." Smith said the US has "17 national labs … and about 85 to

90 research universities" that have been "the fabric of much of scientific discovery … since the Manhattan project," a model other countries "seek to emulate" while Americans "run the risk of taking it for granted." Su said AMD is "also very large supporters of the public private partnerships with the national labs."

Sen. Todd Young — values embedded in models, and AI literacy. Young framed AI as "an issue of national and economic security," said US models would be "embedded with our values" on "privacy … transparency … property rights … freedom of speech and religion," unlike "the values of the Chinese Communist Party," and asked Smith whether other nations would "go ahead" and set standards "without any consultation" if the US adopts none. Smith said "it is the lesson from the evolution of privacy law": the US "didn't adopt a national privacy law.

Europe did twice," and most large American companies now apply European rules across the US because "it's just more efficient" — "you cannot be in the game if you do nothing." Young said he would introduce "the AI public awareness and education campaign act" for "a whole of government approach" to AI literacy and STEM.

Sen. Lisa Blunt Rochester — OpenAI's public-benefit restructuring, and supply-chain resilience.

Blunt Rochester asked Altman about OpenAI's decision, announced that Monday, "to apply to become a public benefit corporation … and to have the PBC govern your nonprofit arm." Altman said "we never planned to have the nonprofit convert into anything," that a PBC was "planned … from the very beginning," that "most of" the press speculation was "inaccurate," and that OpenAI hopes to run "maybe someday the best resourced nonprofit in the world" alongside a mission-aligned PBC able "to raise the capital needed to deliver these tools … people want to use." On Su's Promoting Resilient Supply Chains Act, Su said the semiconductor supply chain must be handled "end to end" — "advanced wafers … packaging … the backends and

… system test" all needing "a footprint in the United States" — with allied partners for "global resiliency." Smith said the stack's interdependence "create[s] more opportunities than vulnerabilities."

Sen. Jerry Moran — consumer data control, cybersecurity, rural broadband.

Moran asked Altman how to give consumers "more control over how their data is used by AI companies while preserving the utility." Altman flagged "a new area": people "are sharing more information with AI systems than … with previous generations of technology," and the systems are most useful "when the model can get very personalized to you," so guaranteeing privacy when someone tells an AI "whatever's happened in their lives" — and controlling what is shared "when one system connects to another" — will "become one of the most important issues with AI." On cybersecurity, Smith said AI is "both an offensive weapon and a defensive shield," that Microsoft has "intercepted attacks against Ukraine faster than a human could detect

them," and that "the United States government must remain at the forefront," with "our intelligence agencies and especially the NSA … well funded." On rural broadband, Altman said AI can "offload so much of the processing to the cloud" and "ship a relatively small amount of data," so low-connectivity areas can be "support[ed] … quite well."

Sen. Ben Luján — national labs, the Test AI Act, and community impact of data centers.

Luján thanked OpenAI, Microsoft and AMD for their work with "the NIST US AI Safety Institute" and the national labs, and asked how the partnerships support research; Altman said models "like 03" are good at "scientific reasoning" and that scientists at the national labs are, "for the first time," calling the new models "a legitimate game-cher[ger] to their research output." Luján said he had reintroduced "the Test AI Act … to … improve the federal government's capacity to test and evaluate," and criticized "President Trump's annihilating budgets for basic research." On community impact — data centers "put a strain on energy and water resources" without "many long-term jobs" — Su said Microsoft's data centers run on "closed loop" liquid cooling

that recycles "almost all" of the water, that Microsoft aims "to be water positive," and cited "more than 90 water replenishment projects, including one … [on] the San Juan River" in New Mexico.

Intrator said extracting "more computational power out of a given megawatt is of paramount importance."

Sen. Cynthia Lummis — patchwork state regulation, permitting, small modular nuclear. Lummis said "the EU has chosen to regulate first and ask questions later" while "China appears to be fast-tracking AI," and that "our lead against China is shrinking to about only six months." Altman said complying with "50 different sets of regulation" would be "quite bad" and that "one federal framework that is light touch … seems important and … fine." No one on the panel disagreed with his assessment of a patchwork. Intrator said an investment could become "trapped in a jurisdiction" whose regulation "would not allow you to make full use of it." Smith said the US "need[s] consistency across administrations" and that Microsoft is "exploring small modular nuclear … including … with people in Wyoming."

Sen. Jacky Rosen — adversarial AI, AI and antisemitism, data-center security.

Rosen cited her bipartisan bills to bar DeepSeek from "government devices" and "federal contractors" and asked Smith about models "developed in or by adversarial countries like the PRC." Smith said Microsoft bars employees from using "the Deep Seek app" and kept it out of its app store over "data going back to China" and "Chinese propaganda," but that because the model is open source Microsoft could "analyze it, and change the code … to remove the harmful side effects." On an ADL report that "major generative AI models have perpetuated … anti-semitic stereot[ypes]," Altman said OpenAI "do[es] collaborate with civil society," wants user freedom but not products "damaging to … the fabric of society," and is "not here to …

make horrible anti-semitic products." Rosen also cited her "federal data center enhancement act" and asked Su about making chips "cooler" and data centers "more secure"; Su cited "30 times improvement" in efficiency "over the last few years" and said chip security is "also very high priorit[y]."

Sen. Dan Sullivan — a tentative US lead, key asks of government, energy, US capital funding Chinese AI. Sullivan asked whether the panel agreed the race is "a huge issue of national security, economic security relative to China" (all agreed) and whether the US lead is "tentative." Altman said "we are leading the world right now" and "will continue," and repeated the key asks: infrastructure and "as much of" the supply chain "as possible in the United States" (Stargate, chip design and manufacturing, "permitting power quickly"); "certainty on our ability to train" and a "light touch regulatory framework"; winning "at the product level"; and bringing "the most talented researchers" to the US.

Smith said he would "connect three things — energy, people, and access to capital," warned "never underestimate the ability of China to build a lot of electrical power plants … more and faster than any other country," and said that if the US keeps "bringing" and "educating the best people … the money will be here." Asked whether "American Venture Capital Funds funding Chinese AI" is in the national interest, Smith said "there's a really good question about whether it is."

Sen. Ed Markey — environmental impact, and algorithmic bias. Markey asked whether "the government" should "conduct a comprehensive study on environmental impact" (Smith and Altman: generally yes), noting data-center electricity demand, water for cooling, and "on-site backup diesel generators." Altman said the environmental conversation "has gotten somewhat out of whack" and that AI "can help" address climate challenges.

Markey said "94% of all new installed electrical generation capacity in the United States" last year "was wind, solar, and battery" and criticized the administration for planning to "destroy all incentives." On bias, Markey asked whether "algorithms … be biased and cause discrimination"; Smith said "They can, which is why we test to avoid that outcome," and Altman said "Of course," adding OpenAI works "to guard against" it. Markey described his "AI Civil Rights Act," which would require companies to "review and eliminate bias … before … deploying," and said "all of the protections we have in the real world should be moved to the virtual world."

Sen. Gary Peters — AI workforce and job displacement.

Peters cited his "AI scholarship for service act" and "AI training act" (signed in 2022) and his "AI and critical technology workforce … framework act," and recalled Altman telling him "that … upwards of 70% of jobs could be eliminated … by AI," with "possible social disruption." Altman said what is "different this time" is "the potential speed": revolutions have always changed jobs — "some jobs go away … some new jobs get created" — but "it feels like it could be pretty fast," and "the most important thing … is to put tools in the hands of people early," a principle OpenAI calls "iterative deployment." The goal, he said, is for AI to be "enhancing work" rather than "displacing" it, noting it has changed "what it means to be a programmer" without ending demand for software engineers.

On open standards "at the hardware level," Su said an open ecosystem lets "the best solution at every level" win because "there's no one organization … that has all the good ideas," contrasting AMD's "open model" with NVIDIA's "closed model."

Sen. John Fetterman — ratepayer protection and nuclear, and "the singularity." Fetterman ("energy security is national security") cited a Washington Post report that data-center demand could raise residential power bills "perhaps … as much [as] … 20%," and recalled evacuating with his hamster during the 1979 Three Mile Island meltdown while still being "very supportive of nuclear" as "an important part of the stack" for addressing climate change.

Smith said Microsoft follows "two principles": invest to bring onto the grid "an amount of electricity that equals the amount … we will use," and "manage all of this in a way that ensures that our activity does not raise the price of electricity to the community," including proposing "a change in the rate that we are charged" to pay for grid improvements. On "the singularity," Altman said he is "incredibly excited about the rate of progress" but "cautious," that this may "turn … out to be the biggest technological revolution[] humanity will have ever produced," and that it must be approached "with humility and some caution."

Chairman Ted Cruz — wrap-up: surprising uses, ChatGPT versus Google, DeepSeek, the AI diffusion rule, and a 10-year state-regulation moratorium. Altman said ChatGPT has been "a real lifesaver" with a newborn, and Cruz recounted his teenage daughter using ChatGPT to write an emotional text to him.

Asked whether "Chat GPT [will] replace Google as the primary search engine," Altman said "Probably not" — "Google is like a ferocious competitor" making "great progress putting AI into their search." On "How big a deal was deepseek," Altman said "not a huge deal": a good open-source model and a consumer app that "briefly surpassed ChatGPT" but does not currently look like it will beat US apps as the default; Su said "somewhere in between," valuable for showing "there are other ways of doing things" under constraints; Intrator noted DeepSeek deliberately hired employees "four years or less out of university"; Smith said DeepSeek "raised the spectre of China's AI capability" for a much broader audience and "was a starting gun."

On what should replace the rescinded AI diffusion rule, all four favored simplifying: Altman said the model should be "winning diffusion instead of stopping diffusion," with OpenAI's "biggest and best data centers," training and "core research" in the US and inference centers built "with our partners around the world"; Su said "some restrictions are necessary … a matter of national security"; Smith said eliminate "tier 2 quantitative restrictions" but require exports only to data centers "run by a trusted provider" meeting "physical and cyber security standards," with "protection against diversion of the chips" and controls so "the PLA … isn't accessing … advanced models or advanced chips." On a "10-year learning period" or federal preemption of state AI regulation, Altman

said "one federal approach focused on light touch and an even playing field sounds great"; Su and Intrator agreed; Smith tied it to Cruz's own op-ed "that you and Senator Graham published" and to "giving the federal government the ability to lead … [on] product safety and … pre-release … reviews."

Well, I want to thank each of the witnesses. This was a very interesting hearing. It was informative. These issues matter.

Uh you saw a great deal of interest on both sides of the aisle in this topic. Uh, and so I appreciate each of you are very busy and doing a lot of things and I appreciate you being here today. Uh, senators will have until the close of business on Thursday, May 15th to submit questions for the record and the witnesses will have until the end of the day on Thursday, May 29th to respond to those questions. And with that, uh, that concludes today's hearing.

The committee stands adjourned.