《赢下AI竞赛》第4部分：Scott Bessent、Howard Lutnick、Chris Wright 与 Doug Burgum

Secretary Bessent, it's wonderful to see you. Before we deep-dive into AI, do you want to give us the high-level update on the 3-3-3 plan? How are things going? You had an incredible clip, by the way, with Maria Bartiromo, where you talked about some of the things happening economically. Maybe just level-set everybody on what's going on, just for good framing. During the campaign, you called it the 3-3-3 plan.

I had a plan that I called 3-3-3. The idea was to get the budget deficit, which was running about 6.7% of GDP under the Biden administration—the highest we'd ever had when we weren't at war or in a recession—down to 3%; achieve 3%-plus economic growth on a persistent basis; and create 3 million more barrels of oil equivalent, so oil and gas, before President Trump leaves office. And look, we're full speed ahead.

June was the first positive June for the Treasury since 2015. We actually had a surplus, and we did that in a good way: We took in more revenue, some from tariffs, and brought down spending.

When I think about what we can do here, what I'm really excited about is the idea that, with AI, we can go back to the paradigm that existed when I was younger, in the 1990s. Alan Greenspan was able to run the economy very hot in the 1990s, and because of the IT boom, we had this very powerful, non-inflationary growth. I think it's highly likely we could have that now, and that kind of growth would bring down the deficit very quickly.

There's been a lot of talk today about the amount of capex spending that needs to go into AI and all of the jobs that it creates. You posted a couple of days ago and talked about an inflection point you've seen in capex spending. As a steward of the U.S. economy, can you tell us what's happening?

It's a combination, and it's a barbell. I've been in Pittsburgh twice in the past 4 weeks. Four weeks ago, I went with President Trump when he announced the U.S. Steel–Nippon Steel deal, a substantial investment by Nippon Steel into an old, very important industry. Then, last Tuesday, there was an AI summit in Pittsburgh with all the big players. Pittsburgh is a natural location for AI: lots of cheap energy, and Carnegie Mellon and Pitt are there.

It was very interesting to see the juxtaposition there. We are seeing this incredible capex. The hyperscalers have obviously been in an arms race—the big 5, the big 7. We estimate that is approximately 1% of GDP a year.

Wow.

So, $300 billion is being spent on AI.

Wow. I mean, it's an incredible thing because, as you alluded to a little bit earlier, it does violate a lot of economic theory in the sense that it just hasn't had the negative, pernicious effects. Do you think that's a “yet” thing, or do you think we're in a structurally different kind of economy now?

You mean the AI boom?

Yeah.

We've seen throughout history that technology can drive these things. If you go back—and I'll talk about the ones I was around for; I was not around for the railroads, but I used to teach economic history—in the 1880s and 1890s, the railroads made it 10 times faster to cross the United States. We had this incredible productivity boom. It was gigantic GDP growth, and it was disinflationary. Imagine having double-digit GDP numbers while inflation was negative 2%, negative 3%, negative 4%, just because costs were coming down.

Then, in the 1980s under Reagan, we had what I would call a deregulatory boom. It's hard for everyone in this room to remember, but everything used to be regulated: the price of airline tickets, telephone bills, banking services. In the 1980s, we had a deregulatory boom. Paul Volcker brought down inflation, but it was also the deregulation.

In the 1990s, which I previously mentioned, we had an electronic buildup, and then finally it kicked in, especially in office work. That led to a big productivity boom, and we paid down the national debt, right?

We had a surplus.

We had a surplus, and it seems crazy. I found a paper the other day that said people were wondering, “Well, what are we going to do if there aren't any government bonds?”

But we fixed that. There are plenty of government bonds. I do think there's a chance now that we could have this growth acceleration. I'm shooting for 3%, but I can tell you the trajectory of the debt path really changes. If we can also have lower interest rates because it's non-inflationary—and I think the Fed is going to have to be open to this idea—

Let me ask 2 questions on that. The first is that, in the examples you gave, we didn't have some of the tariffs that, since we last talked, several of these trade deals have been negotiated further. You probably have better clarity on what the tariff rates are going to be. What do you estimate the dampening effects on the growth rate to be, if any, associated with the tariffs in those trade deals?

Second, I'd love to hear your point of view on the Chinese report of selling half of their U.S. Treasuries, and where the market for Treasuries is going to fall over time here. So, 2 parts.

To address the interest-rate question, I'll take the second one first. We expect that the Chinese will slowly divest, but with the passage of the GENIUS Act last week, I think we could see several trillion dollars of demand for T-bills because of the way the legislation works—it’s under 90 days.

I think that's really going to lock in the U.S. dollar in terms of individuals on the street. Whether it's Nigeria or Qatar, people are going to be using U.S.-backed stablecoins. If I think about the alternative—a central-bank digital currency from China, the euro or the ECB, or even Canada—you know, a lot of you will remember that during COVID, the Canadian government didn't like what some truckers were doing, and they seized and froze their bank accounts.

With a central-bank digital currency, you could put out a mean tweet—not that any of you are—

No one up here is known for doing that. Ever.

If you have a government-backed digital currency, they can shut you down, as opposed to this kind of unbridled choice that consumers are going to have with U.S.-dollar stablecoins.

On the first question, about growth rates being hampered by tariffs, is the revenue you're seeing effectively offsetting the rates?

We haven't seen that yet, and I think there's a good chance that we could. If we think about China, China has a high tariff rate. It's 30%. The Chinese business model is like the brooms and the water buckets from Fantasia. They just keep going. It's an employment agency.

I'm thinking of the song. I know the piece.

It's an employment agency, so they will just keep cutting costs to maintain market share. We haven't seen that thus far, and a lot of the other foreign producers have cut prices to maintain market share. A lot of U.S. companies have eaten into their margins to maintain market share.

The other thing we're seeing is that the tariffs are creating onshoring. You might have seen—can't remember whether it was yesterday or the day before—AstraZeneca said that they were going to build an incredible $50 billion plant here.

We're seeing this big onshoring move, which I think can accelerate all of that. I think there's a very good chance that, just like with AI, we're now in the construction-boom phase, and then we're going to be in the use-case phase, right?

I think we could have this massive construction boom.

And then the factories get populated. Part of President Trump's One Big Beautiful Bill—the most powerful part of that is the 100% immediate expensing—

Right.

—of equipment, and we also did it for factories. Not only are we trying to make the U.S. the best destination regulatory-wise, we're also making it the best destination tax-wise. You can immediately write off all the equipment for the next 5 years. You're going to be able to write off the factory structure. I see Secretary Burgum, right, and we're going to have cheap energy.

Which seems like a pretty good combination. Should the Fed remain independent? Should Trump replace the Fed chair? You guys seem a bit frustrated with him. What are your thoughts there?

You guys have done such a good job in terms of confidence in the markets. CPI went up a little bit in June, and it does seem like the economy is very strong and people are very confident. Polymarket is showing that no rate cut is the most likely case in September. So how do you think about the Fed?

If you look, the Fed publishes something called the Summary of Economic Projections, and it's pretty politically biased. But we're seeing that we could see 1 or 2 rate cuts this year. I think that once we see, over the next 1 or 2 months, that the tariffs haven't been inflationary—and I have breakfast with Chair Powell almost every week—I just keep saying that a 1-time price-level increase is very different from the notion of a persistent inflationary spiral.

I think we used to say TDS was Trump derangement syndrome. I now say TDS is tariff derangement syndrome.

Right?

And when you think about it, the market crashed, then it had the fastest recovery in history over a 54-day period. We’re back at a new high. So I think the market’s looking through all this to next year with the productivity boom. And to the question, I think, at a minimum, on a forward 12-month basis, we’re going to take in at least $300 billion in tariff income.

Yeah. Are they punishing you in a way because maybe the rollout of the tariffs was a little bit shock and awe or a little bit effervescent, however you want to describe it? It was pretty intense. Is the Fed sort of punishing you for that, in your mind?

No. I think they’re just stuck in an old way of thinking.

How much should they cut? How should they think?

Look, I’m only going to talk about the mistakes they made, not the mistakes they’re going to make. But I do think at a point they’re just going to have to admit that they have been wrong, because if you think about it, I don’t believe that a tariff is a consumption tax.

Right.

But if tomorrow we put on a 1% consumption tax, you would never say that’s 1% inflation.

That’s right.

Right. So I am hoping that, in their infinite wisdom, the—I can’t remember, it’s 350 Ph.D. economists, which I said on TV either yesterday or the day before—my worry is that the Fed is turning into universal basic income for Ph.D. economists.

Right?

I don’t know what they do. They’re never right.

Maybe you should double the number of Ph.D.s. If you go to 700, they might get it right.

Well, look, if you were to look at the central tendency versus how they’ve done, it’s shocking. It’s shocking. I said, if air traffic controllers did this, no one would get in an airplane.

They do seem to put a little tail on everything. Last question, maybe as we wrap this up: Secretary, as an economic historian, very briefly, tell us the lessons of these previous economic expansions and technological booms. What do we need to learn from those things, whether it was railroads, the agrarian revolution, or the Industrial Revolution, so that we don’t screw up the AI revolution? What are the few critical things we need to do right?

Well, I think the most important thing that we are doing is getting out of the way and setting the conditions for it. One of the surprises I’ve had—and I’ve had a lot of them—when I went from civilian to public servant has been that, in the U.S., we’ve made it so hard to build things.

Right. Right.

And it’s just very frustrating. I’m sure Doug and Chris will talk about it, but TSMC wants to build a gigantic fab system in Arizona, and I think it might be able to produce up to 7% of the chips that the United States needs. They’re dealing with local building inspectors, who say—and evidently, these chip-design plans are moving so quickly, you’re constantly calling an audible and saying, “Well, 3 months ago it looked like this, but in 18 months we’ve now decided it needs to look like this.” And you’ve got someone saying, “Well, you said the pipe was going to be there, not there. We’re shutting you down.”

And the level of permitting—we always talk about how Germany had deindustrialized. I think I may have even talked about it on your podcast.

Right?

We even made the decision to deindustrialize—

Through our environmental regulations. And I think the most important thing we can do is make it easy to build things again, stay out of the way, and not overregulate.

Secretary, thank you. Thank you very much.

Thank you.

Thank you guys for being here. I know it’s been a rushed afternoon. We did not expect the incredible turnout that we’ve had, but thank you both. You’re the chair and the vice chair of the National Energy Dominance Council. We’ve talked at length today about the boom underway in AI. We’ve talked about this on the podcast. The U.S. energy-production capacity—electricity-production capacity—is about 1 terawatt today, growing to an estimated 2 terawatts by 2040. China’s going from 3 to 8. They’re adding an America every 18 months. Maybe you guys could just give us an update on the National Energy Dominance Council and how that work is going to try to accelerate energy production in the United States to help enable this AI boom.

Well, happy to do that. And I just want to say again, thanks to All-In for pulling together this amazing team—

And Silicon Valley.

Yes.

And Silicon Valley, too. Thanks, Chris.

When historians look back on this day, when historians look back on the challenge of our times—which is, like the summit called, winning the AI arms race—I think one of the things they’re going to conclude is that the reason why the United States won the AI arms race was because of President Trump. I’m not saying that as a political statement. I’m saying that the policy of the Trump administration is more energy-first and has an understanding of how important it is for the AI arms race.

With that, as you’ve just outlined, we’ve got a huge challenge ahead of us. China is deploying everything. They added 94 gigawatts of coal last year. One gigawatt is Denver. Over 60% of their power is still coming from coal. They’re just pouring that on. The Wall Street Journal ran an article yesterday talking about what a great job China was doing with EVs and solar. I read the whole article; they never mentioned coal. It’s two-thirds of their electrical power.

So, just by definition, two-thirds of the EVs in China should have a bumper sticker that says “Powered by coal.” This is a race of our lifetime. They’re also doing nuclear. They’re doing hydro. They’ve got no permitting issues. I mean, they build a hydro dam—it’d be like the equivalent of us putting a dam on the Grand Canyon, what they were doing on the Yangtze. So we’ve got real competition.

We can lead in technology, but we haven’t been leading on electric production. Part of the job that Chris and I have with the National Energy Dominance Council is helping cut red tape and produce more electricity, whether it’s hydro, geothermal, or nuclear. And, of course, LNG and natural gas are key parts of this. Bringing back coal and making sure that we stop shutting down baseload in America has been a key part of what we’re doing.

Yeah. And just to riff off that, where the United States gets electricity today, in order, is natural gas by far, then nuclear, then coal. Those 3 sources are 75% of U.S. electricity and 90% of what matters, which is electricity that’s there whether the sun’s shining or the wind is blowing.

We had, in the previous administration’s plans, the removal of 3.5 gigawatts of hydropower. We’re going to stop that. There are plans between now and 2030 to close 100 gigawatts of power plants—100 gigawatts—and we’re stopping most of that. If we need to add 100, after the meetings they had this morning, I think it’s more than 100 gigawatts in the next 5 or 7 years. The first thing to do is stop subtracting 100 at the same time you want to add 100.

I think America became great by big, bold people making big, bold investments. That’s where we got here. Then we just drifted off track the last bunch of years and made it so hard to build something, so easy to stop something, and developed a crazy love affair with intermittent, unreliable energy sources.

You’re talking about solar. Why are you so down on solar? This is the cheapest thing you can install. Batteries are here, and they’re being produced at an incredible rate. Why are we so anti-solar? Or why are you so anti-solar?

Oh, I’m not anti-solar. So why do you keep saying that this unreliable solar, if you put batteries on it, is totally reliable?

If you take all the batteries in the United States, you could store 5 minutes of power—5 minutes—of the entire country.

But we’ve had many days in California and Texas where solar has been the majority of it, so why are you so down on solar?

It can be the majority on a sunny day in the summertime. That’s not what matters. In PJM, where we are right now, at peak demand this year, 97% of electricity came from sources other than wind, solar, and batteries. Wind, solar, and batteries delivered 3%.

You’re cherry-picking D.C. You’re cherry-picking D.C. Let’s talk about California and Texas. These are very populous states.

Absolutely. Let’s talk Texas. The peak-demand times in Texas have been cold spells with low wind. They’re high-pressure systems in the wintertime. Wind and solar go on vacation. They’re 35% of the capacity in Texas and 8% of the delivered power at peak demand.

You’re talking about 2 weeks. I live in Texas.

Yeah, but those are the 2 weeks that matter, right?

No, the other 50 are the ones that matter, actually. But sure.

In Winter Storm Uri, when they weren’t ready, over 200 people died. We don’t want people to die. We want the lights to go on when people need them. And it’s the system cost that matters. If you’re not there at game time, all you are is a parasite on the system that is there.

Let me redirect this back to AI, because—

Good idea.

Good idea. If you actually forecast the growth of just the servers, then the robots and all of these things, we’re going to need terawatts and terawatts. That’s on one side. On the other side is this latent fear that some people have that this will somehow upset the apple cart—sustainability, the climate, et cetera.

How do we create the logical bridge so that people really understand that this is all possible, that this is not going to destroy the Earth, and that we can get this abundant energy? Especially because, as you guys have said very well, if we don't do it and somebody else has marginal-costless energy, they will de facto win.

So how do we frame the argument so that people can understand this better?

I've been writing and talking about that for 20 years, and you're 100% right. To me, it comes down to the same thing AI is focused on: data and facts. We've increased atmospheric CO₂ by 50%. It absorbs infrared radiation. It's been a force for warming. That's all true, but if you look at the trade-offs, it's not in the top 5 problems the planet faces.

The biggest source of decarbonization, not just in the United States but globally, has been market forces. Cheap natural gas displaced coal, and what's a lower-carbon energy source? Nuclear. That's on all the time. This administration is all in to get the nuclear industry moving again. Natural gas is the fastest-growing energy source on the planet. Get out of the way of that. Let natural gas grow.

It's the cheapest source of electricity in the U.S. I'm pro-solar as well; I just don't want taxpayers to pay for it. I want businesses to pay for it. But solar is going to keep growing.

When you get to nuclear expansion, I just want to talk about nuclear expansion for one second. How do we actually build these things faster and have the capability and the technical construction know-how so that these aren't 15-year projects? Also, how do we incentivize the states to basically get out of the way, or these other organizations that can launch frivolous lawsuits and slow it all down? How do we do that?

There are a lot of regulatory reform things. First, we're working on FERC, right? FERC has this inefficient queue system that just gets gummed up with mostly stuff that's never going to happen. FERC came out yesterday with a new system where you're going to prioritize things that matter. They're going to move through faster.

You saw the Supreme Court's decision on NEPA. We've got to get NEPA back to where it was: a process check on the environment, not an avenue for lawfare to stop things and kill things. So there are structural changes, and there are just common-sense reforms.

We're going to get rid of Clean Power Plan 2.0, which says you're going to have to have carbon capture and storage 15 years out on any natural-gas plant. What's going to power AI? Let's just be honest. What's going to be the main source of new electricity in the United States, by far and away? Natural gas, just because it's cheap, fast, reliable, and dependable. Solar is going to play a role. Nuclear is going to play a role. Hydro, geothermal, stop closing coal—lots of pieces—but it's dominantly going to be natural gas.

It's the fastest-growing energy source, not just in the U.S. but on the whole planet. There's a reason for it: it's cheap, it's massively abundant, it burns clean, and the machinery lasts longer than machinery burning oil or coal or something else. But it's about letting businesses decide. Doug and I are not here to tell anyone what to build and what not to build. We're here to get roadblocks out of the way so capitalism, consumers, and investors can decide where—

I mean, that's the good news. Solar is cheaper than coal plants, right? So, okay, some of the scalability of nuclear—

As simple as that. Yeah, the scalability of nuclear, I think, is unbounded. What we've seen in China in the past couple of years is these Generation 4 nuclear reactors. This pebble-bed reactor is probably the most elegant, beautiful energy system designed in human history. It's incredible what it can do: the scalability, the cleanliness of it, and how it works.

We have no effort in this country today to build and deploy Gen 4 reactors because there's no economic incentive. The path to get there is so far, and the cost is so high. What can the National Energy Dominance Council—what are you guys doing in your roles to make Gen 4 reactors happen? Because everyone's saying, "Go back to the AP1000, these old Westinghouse designs from 50 years ago, and build that for nuclear." Why can't we build for the future, and what can we do to create the incentive to make this work?

The one thing that's already happened, if people are interested in nuclear—which doesn't help us in the near-term race that we're in—is that the near-term race, as Chris said, is going to be won by us getting natural-gas power online and stopping the shutdown of existing facilities. But President Trump signed 4 executive orders on nuclear about 6 weeks ago, and there's been a flood of fresh capital coming in. We've got a bunch of venture capital going toward close to a dozen different SMR startups. There's a lot of interest going on in that field. Chris's work with the national labs is redirecting that.

Nuclear has a future, but it's not the thing we need in the next 24 months. That's got to keep moving ahead. President Trump's executive order has helped that, but we've got to get focused on getting more power right now.

Nuclear is the single biggest issue I work on. We will have 3 next-generation Gen 4 reactors go critical at Idaho National Laboratory next summer. We're supplying HALEU, the fuel for these next-generation reactors. We've already committed to 5, and we'll give it to a dozen of these next-generation reactor companies.

We worked into the One Big Beautiful Bill a nudge to keep the tax credits for nuclear, because the government smothered the industry and killed it for 3 decades. Even a free-market guy like me thinks we need to get a little help to get it started.

How far away are we from it running on the free market?

Probably 10 years, because it's just a learning curve. With the small modular reactors, you've got to build up the supply chain and build them in volume. The cost can come down dramatically, but the first ones—

As you look at your energy-demand curves, do you account for this revolution happening in physical AI? Every time I look at it, it's data centers this and buildings that, but no one talks about physical AI, which is batteries in robots. Some people are estimating hundreds of millions or billions of these things being built—trillions. Is this part of the energy calculus as you think about demand?

It is a meaningful part of it, and yes, the more you look at that, the more you see increased consumption of energy there, and the more excited I get. The more we can build things at scale, the better we can get the economics.

One other data point we put out at the Department of Energy: we got 16 locations to build data centers. We said, "Who wants to come build one? We'll permit them right away. We'll help you build power generation right next to it." We got 300 responses. We will announce tomorrow the first 4 of those sites that will be developed, and then you'll hear many more coming behind that.

How do we solve the supply-chain issues around the turbines and the other enabling technologies that we need for things like natural gas? I agree with you. I have a data-center project in Arizona. It's a gigawatt, and it'll be $25 billion of capital.

But we're stuck in this weird situation where onshoring the natural-gas turbines is extremely difficult. Then you see certain people will just buy entire natural-gas plants and ship them over. So how do we solve the supply-chain constraints to generating the energy we need?

Again, back to the immediate need right now: we need more power, and we need power for factories that are producing AI, using Jensen's term, which I think everybody should stop saying data centers. A data center, if you have one the way America thinks about them, is processing a shopping transaction. It helps the seller, the buyer, and maybe a third party. If you're processing a healthcare claim, it's a provider, a payer, and a patient.

But in AI, it's general-purpose technology. We're literally manufacturing, every day, over and over, more intelligence. So that's different. It's not data centers; it's AI factories.

We've taken a look at the supply chain. If any of you are trying to build an AI factory and you need power, and you haven't talked to Chris and me and our team inside the White House at the National Energy Dominance Council, you need to come and talk to us. We're mapping things out and talking to everybody in the industry.

We're a neutral party, but we're saying, "Here's where the shortages are." We've talked about things like the Defense Production Act. We've talked to companies that are producing turbines. Everything we're doing, we say, "Hey, you've got to ramp up," because some of these people are sleeping on the sidelines. They don't think there's going to be real demand. We're saying, if anything, the demand is underestimated.

So we're trying to ramp up supply into the supply chain, but please contact us. We're there. Think of us—we're not a group that writes papers. We're a group that helps people build. We help people build projects. That's what we do.

Can't wait to visit.

Just build a data center. Then you'll get an invite.

So, AI factory.

AI factory. Somebody else can build data centers. We've got enough data.

Secretary Burgum, right, can you, as we finish up, hit on the point we were talking about a little bit earlier, which is that you take a step back? The focus here upstream on these prioritizations, from energy to critical minerals, is not just about having a new market.

Obviously, on the AI side, there's huge demand, and this buildout is important for national security.

This buildout is important for winning the AI race. But the derivative impact is what's most interesting, right? These are thousands of jobs—tens, hundreds of thousands of jobs. And then, on any of these manufacturing buildouts, particularly in factories and nuclear capabilities, they're usually going to lead to 10× the amount of indirect jobs as well.

Back to Chris's point on the supply chain for these things, can you talk a little about the job impact we're seeing now? And then, if we're successful here in building out capacity, how many jobs are we talking about? How much can we actually help the middle class here?

Well, it's a fabulous question, Christian, and I'm so bullish on the U.S. economy because, as our friend Scott, who was just on here before us, said, you take the combination of lower taxes, dramatically lower regulation, accelerated permitting time—just accelerating permitting—there could be $1 trillion to $1.5 trillion stuck in this 2- to 4-year federal government permitting thing. We accelerate that expenditure of capital, the onshoring—the greatest economic developer in history, bringing foreign direct investment back to the United States—and President Trump with these tariffs. You know what we announced in Pittsburgh? It's $15 trillion that's coming back.

With AI, software has always been the one thing that extended human capability more than any other in our lifetimes. And now, with AI, it's just a massive multiplier of that. But to make the factory happen, we're going to have an explosion in jobs in the trades. I mean, you're going to be able to skip college, go directly into developing a trade, make $150,000—yeah, $150,000; $120,000 to start in my home state.

And again, for people who are spending money on site selection, I'll tell you one thing: You want to build it faster, go to where the stranded gas is, build your power plant there, and build the AI factory next to it. You don't have to permit a transmission line. You don't have to permit a pipeline. Those are the 2 things. Linear infrastructure has been weaponized by the people who are opposed to energy development in this country. They weaponized the blocking of those things. I say pipeline, you say protest.

So go to the same place and co-locate. President Trump himself has said in speeches, we're going to let you operate off the grid. We can build all this stuff and keep rates for electricity for small businesses and consumers down because we've got to add to the supply. But if you're going to go to where the gas is, there are sweet places to go: the Marcellus, the Permian, or the Bakken. And you can save tens of millions hiring site-selection guys. Go find the people with stranded gas and get going.

Great. All right. Well, Secretary Wright, Secretary Burgum, thank you for being with us. That was great.

Well done. Thank you. Really great.

Nice to see you, brother.

Great to see you guys. Welcome back to doing it. How are you?

Nice to see you.

Howard, I noticed you had that incredibly smooth, refined tequila at your birthday. How was it? Take us through it.

14 minutes. Just let's get to it. Let's get to it.

Yeah. Smooth. Yeah. Well, thanks for being here. I'm going to kick us off.

So, the White House just rolled out a massive deal with Japan, which obviously plays a critical part in the semiconductor supply chain. Could you tell us a little bit about the nexus between this new, exciting trade deal with Japan and how it fits with the current debate around winning the race on artificial intelligence?

So, it was fundamental for Japan to lower its tariff because its car industry and its manufacturing industry are fundamental to its economy. And they paid $550 billion, what the president likes to call a signing bonus—right, the greatest signing bonus of all time. So they've committed $550 billion to finance projects in America that are important to the president and to American infrastructure.

So we can build power, which means we could build 10 nuclear power plants. We could build fabs, right? We could do critical minerals. We could do shipbuilding, power—we could do anything, and they will finance it. We split the profits of the project: 90% for America and 10% for Japan. And I don't think people can actually understand how powerful that is. This is the national security sovereign wealth fund of the United States of America, funded by President Trump's tariff policy.

Right.

That produced that kind of money committed to America.

Congratulations. Will that actually go into a sovereign wealth fund that you've been talking about and the president has been talking about?

No, I think this is separate. What the president says about the sovereign wealth fund is, we do a sovereign wealth fund that invests when we've done paying off our deficit. Right? First, we've got to pay off our deficit before we're trying to make money.

So what this is is the Japanese government says, "I will pay for it." You want to build a nuclear facility? Build it. You want to build 10 nuclear facilities? Go build them. You want to build a pipeline? Go build it. You want to build fabs? Go build it. Whatever you think is necessary, you build it. We'll pay for it. You net-lease it to an operator, and we'll split the lease payments: 90% for you, 10% for Japan. It's a blockbuster if there ever was one.

But it's an incredible deal structure. How do you get to that?

Well, I got to that. I came up with this idea in January, and then I kept restructuring it to try to figure out how to do it because I met with some Japanese senior executives before the election, before Inauguration Day. And they said, "You know, I understand your tariff policy, but Japan's never going to open, right? They're just never going to open. I mean, in 1853, Perry took an armada and tried to break it open, but he couldn't open the Japanese market. So come up with another idea."

And the other idea was they buy it down. So, the structure we used—how they did it—they originally started offering us loans or loan guarantees, and the president was like, "I don't need someone else to loan me. I don't need to borrow money from someone else." And then finally we figured out that it really just needed to be committed capital to back projects that we want.

So it was 5 months in the making, with me talking to the president about doing different structures. Eventually, in the middle of last week, we came to the structure. The president said, "Okay, I like it. Let's bring them in and talk," and then the president made the deal better.

Are you going to replicate this? Is this like a new blueprint, or is it unique to Japan, which is protectionist and has its own unique culture?

Well, I mean, the problem that Korea has is they're staring at it. You know, they view themselves as deeply competitive to Japan. They both produce huge amounts of cars. They both produce huge amounts of electronics. They both do these things, and now they're looking at the price.

Right?

And they're thinking, "Ouch." So, how quickly did they come to see me? Let's say when we announced the deal. They were in my office today.

Oh, yeah. Fantastic.

How much have you prioritized market access for American businesses into some of these countries versus some of the other trade considerations? Where does it rank? We've talked about this a lot, particularly as it relates to AI. And I think that part of this is in the action plan and in the EOs being signed later today, but this is a broader question for American businesses.

I work in agriculture. It's very hard to access overseas markets, and there's not a lot of parity. Has that become key to some of these conversations, and where does it sit on the priority rank?

That's the priority rule. The rule is: You must open your market. Open, open, open. And let's be clear: These markets have never been open. We have Stockholm syndrome in America. These markets have never been open. There are tariffs and non-tariff trade barriers, like you can't sell an American car in these locations. Whether you want to or not, you're not allowed, or they won't buy them because the seat belt is like this or this is like that—they make these rules.

So we are demanding that the markets are open. And the issue with Japan was they were never going to open it.

So what are we going to do?

And the answer was, all right, that's where he came up with this signing bonus, right?

So, reciprocity or something interesting?

If you want something more bespoke, Vietnam is completely open, Indonesia is completely open, the Philippines is mostly open, with a small deficit and a relatively higher tariff. Right? So, there are a lot of levers, and you pull those levers.

When are you going to wrap all this up? This has been really shocking, and now I think kind of a more mundane, methodical approach. So when does it all wrap up and we can put the tariff issue behind us?

Okay. So, on August 1, whatever hasn't been settled will be settled, and the tariffs go into effect. All these 10% tariffs—they'll all just pop up to some higher number. He sent a letter to a lot of people, right? And now nothing stops them from negotiating the next day, but they're paying on that day.

So that's next Friday. I mean, that's not that far away. So we're very busy because a lot of people are now coming to the table with their best offer. But the price has gone very high.

And let's be clear about what that price is: you will open your market to America. You will open it to ranchers, farmers, and fishermen. You will open it. You couldn't sell lobster to all these places. For instance, Indonesia is completely open except for 2 products. It's a Muslim country: no pork, no alcohol, right? We're talking India, obviously no beef, right?

I mean, you do things like that. You say, “But we need it open.” If they don't want it open, there's your tariff. It's 26%, 27%, 31%, 19%, whatever it is. And then, if you decide to open it later, come on, but that's what we're doing.

You'll find out over time what other kinds of regulatory processes they have put in place. This is always the issue when any of us work in foreign markets. You've worked in foreign markets, and you go in and then you find out, well, there's this thing I have to do, and this thing takes 18 months or 36 months, and they make it hard to get the permit or whatever you need.

There's always a way. Does this become a continuous policing exercise for your department? How does this become part of American trade? Is this an ongoing, iterative process here?

They've bought their tariff rate down by opening the market. So if they mess with that, they're messing with the president. And I don't know if you guys have seen him on TV, but that doesn't really work well. Yeah.

Okay, so the idea is he's making the deal; he's closing the deal. The way we talk about it together is, I set the table, right? He closes the deal. And he is the best negotiator because he's done this his whole life, and he's the president of the United States. So that's an amazing power, and he wields it to get the best deals.

Well, let's talk about the big issue: China. Where are we going to wind up with China—reciprocity, TikTok, the whole shebang? Is this going to be one big grand bargain? Taiwan, TikTok—there are so many issues. Is there any way to thread the needle on this?

I think the way I think about China is, draw a line. Okay, there's below the line. They sell us baby clothes, and we sell them soybeans. That stuff, we need to do more of it. We want to buy more of that. They want to buy more of ours. We need to open that. Get this day-to-day stuff flowing below the line.

Got it.

Above the line would be our best chips: Blackwell chips, H200s and H100s, right? We don't want to sell them our best stuff. They don't want to sell us hypersonic missiles either, right? We would say, if it was open, “Well, let's take a couple of your hypersonics. Let's see what you got.” Right? So that's not happening.

So that's above the line. And then the question is, what's the line? That's the proper negotiation, right? Below the line, let's get it on. It's good for both economies. Above the line, we're competitors. Let's just call it what it is and stick with it. And then what we can really negotiate when we're together is the line.

Where's TikTok in all this? Jacob and I are both pretty adamant this is spyware. This is something that should not be on 100 million Americans' phones. It is way too dangerous. They've proven themselves to use it to spy on journalists already. And the fact that they won't divest from it, I think, tells you everything you need to know. They see this as a critical weapon against the United States. What do you think? What does the administration think?

Well, the president is reasonably positive about TikTok, provided it goes into American hands and it's controlled by American technology. Right? I think his view is that they've got to be out of it. It's got to be on an American technology stack, and it's got to be owned by Americans, period. And then how we work it through from there, we'll figure it out. Right now, it's sort of in that—

Straddling the line.

Yeah. You're sort of staring at each other, but eventually that'll get sorted out. I think that deal will happen, and America will buy TikTok, because the alternative is just shutting it off, and that just seems illogical.

Can I go back to the above-the-line, below-the-line idea? I love that saying, by the way. How do you think about the export controls to various countries and various regions? What's your risk calculus about where those things should be? And if I could just add a question that builds on top of that: you've talked about creating AI economic zones where trusted partners could get preferential access to American technology. So could you describe a little bit what your vision is for that?

I think what we're wrestling with—and this is literally the intellectual wrestle we're going through now—is the idea that we are comfortable with allies buying significant numbers of chips, right, and having a large cluster, provided that cluster is operated by an American, a trusted American operator, and the cloud is a trusted American operator, so that we know that giant cluster is surrounded by us.

Right.

Right. As you go down from there, right, that's where we go. Okay, if they want a smaller cluster, would you expand the number of people who are trusted, right? And the answer would be probably yes, right? And then when you go down from there to a smaller and smaller cluster, right, how do you deal with that?

So I think cluster size is sort of the way of thinking, rather than saying—because I went to Poland, and I was in Poland on a mission for the government, and the prime minister of Poland chases me down and says, “What did I do to America to be Tier 3?” And I was like, “I thought you were part of Europe. I didn't understand what the issue could possibly be.” So I think the answer is: ally or not, cluster size, and who controls it or not.

I think once you sort of wrestle with those ideas—and anybody who has ideas along those lines, you want to come and talk to us about it—because this is really the thinking right now, and we're debating that right now.

Howard, I just want to say thank you. It's so great to have a sharp negotiator and such a creative mind representing America. It makes me feel really great about the—

90% carry there. There may be fun.

How do we get 90% carry?

I love it. Yeah, yeah. I love that you said that you're a New Yorker.

Yeah, I am. But the—

Negotiator in chief.

New Yorker. I grew up on Long Island. My kids have grown up in Manhattan, but the negotiator in chief is Donald Trump.

That's nice that he's got you, right? He's amazing. Well, thanks for coming. We're going to make some room for the president. He's going to get ready. Howard, thank you for joining us. That was great. Thank you.