Trump斡旋加沙和平协议、芝加哥国民警卫队、OpenAI/AMD、AI循环交易与金价上涨

Chamath Palihapitiya is here—our chairman, our dictator. How are you doing, brother?

Great. Fantastic. You know, you put the Uber stock price and the Robinhood stock price together: $150, I'm annoying; $200, I'm insufferable; and $250—we just hit it. I am officially off the reservation, hopping the fence. So, yeah, life's good.

With us again, David Sacks, the czar.

By the way, what's the reason your Robinhood stock is through the roof? Because of the crypto policies of the Trump administration.

Yeah, that's fair. That's half of it.

That's why it's gone from, what is it, like $3 to $130?

$2. I paid less than a penny for my shares, but yes, it went from $13.

Well, this is a typical liberal. You benefit from our policies and then you trash them.

I'm a moderate—an independent moderate.

Here we go. It's starting already. It's going to be a great week. Everybody buckle up. And he puts the namaste in your payday. Our fifth bestie is back. Brad, welcome home.

It's great to be here. I went out and got a haircut.

Welcome back into the fold.

Welcome back. I got a haircut for the occasion. I mean, Jason—Jason, do you know how long it's been since I was last on this pod?

I don't know. I've invited you like 12 times, but I think there's something going on with email. I didn't get any of my White House invitations; you didn't get any All-In invitations.

It's been so long since I was on this pod, I went back and listened. The last show I was on, we were still talking about how great Chamath's SPACs were before they crashed the entire market and it was already back. I mean, it was before Saxy Poo saved the entire crypto industry from prison and certain bankruptcy. And, Jason, it was before you got so tilted with Democrats that you went completely off-grid, moved to Texas, and bought a bunch of guns.

There it is. It's been a minute. So much changed in a fortnight.

My God. Wow.

Here we are. We're all back and we're having some fun. We'll have a little fun with our fifth bestie who's here.

Only a $5 billion drawdown for me between those memories. Who's complaining?

Who's counting, for a certain number?

We got to start off with the Israel-Hamas ceasefire deal, thanks to President Trump, who announced it just yesterday. We tape on Thursdays; he announced it on Wednesday. You'll listen to this on Friday, hopefully. He announced the first phase of a multistage peace deal.

Notably, it was on the day after the 2-year anniversary of the horrific October 7 attacks, which led, obviously, to the invasion of Gaza. This war has been particularly devastating for Gaza. Most of the region has been destroyed, 2.3 million residents have been displaced, and according to the Gaza Health Ministry—which I know some people will debate—over 67,000 Palestinians have been killed.

The deal is based on Trump's 20-point peace plan, which the White House published last week, and the Israeli government will vote on the deal today, maybe even before the show comes out. Even with all this tragedy, something has gone tremendously right when it comes to these peace negotiations and the peace talks. So, Sacks, tell us what went so right here in those negotiations.

Right now, it looks like both sides have agreed to this deal. The first phase of the deal is for there to be a ceasefire to end active fighting. It's going to allow unrestricted aid into Gaza, the release of all remaining Israeli hostages, and the release of 2,000 Palestinian prisoners by Israel, as well as the start of an Israeli troop withdrawal.

That is what apparently has been agreed to. The Middle East has a way of disappointing you. Sometimes these deals don't stick, but based on everything we understand to be the case right now, this appears to be a big breakthrough: the fighting will stop and all the hostages will be released.

I do think it is a big accomplishment by the president. The way this happened is that the president was willing to cajole, coerce, and use pressure with respect to both sides of the conflict. Longtime diplomat Aaron David Miller, who advised both Republican and Democratic administrations on Middle East peace negotiations for the last 3 decades, said—and I'll quote here—“Donald Trump has demonstrated a degree of will unlike any other president, Republican or Democrat. He has pressed an Israeli prime minister in a way that none of his predecessors have ever done on an issue that that prime minister considers vital to his political survival and the way he would define Israeli security requirements.”

That was Aaron David Miller, who again was the U.S. negotiator in Middle East peace negotiations for a quarter century. In fact, he wrote a very interesting article in 2005 called “Israel's Lawyer,” in which he was a little bit self-critical of the U.S.'s previous efforts and his own efforts, saying that in the past, the U.S. saw its role as being to represent Israel as opposed to being an objective negotiator.

He critiqued the U.S.'s efforts, saying that we might be able to get more done if the U.S. was perceived as more of an honest broker and perceived as willing to be a little bit more neutral in the negotiations. I think what he seems to be saying here is that Trump was willing to do that. He was willing to pressure Netanyahu as well as Hamas.

Obviously, you saw in the last few days that he said that if Hamas did not agree to this, all hell would break loose. So, he was willing to pressure both sides into accepting this deal, and you have to give him credit for it, for where things stand right now.

You are seeing, it seems like virtually everyone's giving him credit—even The New York Times. Even Jon Meacham was on MSNBC this morning giving President Trump credit for doing the seemingly impossible. The reason I say “even Jon Meacham” is because Meacham frequently criticizes President Trump as being a fascist or having authoritarian tendencies. Even he had to acknowledge that Trump had pulled off the seemingly impossible here today.

All right, Chamath, obviously this is going to have a pretty positive impact on the region. As we look at—if this does, and again, we have to be cautious here because, as Sacks correctly pointed out, in the Middle East sometimes these things fall apart and they're delicate—let's assume that this peace deal does make its way to fruition. What is it going to look like in the region after peace between these two nations?

I don't know. Let's start with that: nobody knows.

Nobody knows. Yeah.

I think it's been a very unpredictable place for a very long time. That being said, what is the rational thing that has to happen?

Most of these countries are facing a very obvious problem, which is that they have a resource that is becoming increasingly less valuable. That resource is oil. The practical reality is that you want peace so that you can focus on the forward monetization of your reserves.

When that oil sits in the ground, the longer it sits in the ground, the less valuable it is to you because it doesn't actually add to your balance sheet. The faster you can monetize the oil, the faster you can deploy it into other things, including services for your own citizens.

The longer you wait to do that, the problem we're seeing is that other solutions come around the corner. Eventually, in the 10-, 15-, or 20-year time frame, you'll have an abundance of electrons from nuclear. In the meantime, you have an abundance of electrons from gas. You have an abundance of electrons, frankly, from solar. All of these things will ultimately diminish the net long bid for oil.

If there's instability, you can't focus 100% on monetization. If there's stability, then the entire focus needs to be on monetization of the oil asset. I think that's a very healthy thing because it starts to create market-driven incentives that will only accelerate all the positive things that are happening.

By the way, the other thing I just want to say, which is personal, is shout-out to a friend of the pod, Jared Kushner. Aman [?] tweeted this, which I think is just superb: in one week, Jared did the largest LBO ever and then also helped negotiate Middle East peace. It's unbelievable. It's really incredible what these guys pulled off, and it allows that region to be a real pillar of what happens in the next 50 to 100 years.

All right, and I think it's well said, Brad, that some stability in the region will help them with the transition from an oil-based economy to what you and I learned in our trips there—some of them together—that they're looking to transition to private equity, solar, renewables, AI, owning sports teams, venture capital, tourism, and building new cities.

All of that is hard to do, and then also getting investments if people feel the region's unstable and dangerous. It's not a good place to invest capital. So, let's keep pulling the string as to what the best-case scenario is here if we have true stability in that region.

You nailed it. The conditions for all of those things, economically, are safety and security. You don't have any economic growth in the region unless you have that. You guys covered it pretty well, but maybe just highlight a broader feature of this presidency.

We might call Trump the “moonshot presidency.” If you think about everything he's trying to accomplish, even Hillary Clinton said that if he pulls off peace between Ukraine and Russia—which is also going on simultaneously to these efforts—she herself will nominate him for a Nobel Peace Prize.

I look at the presidency, and there are moonshots going on everywhere: China, AI, reindustrialization, India-Pakistan, Ukraine and Russia, and what he's doing in the Middle East.

And if you think about moonshots, we all do these in Silicon Valley. These are high-risk, high-reward efforts. We want to back people who go for moonshots, but the reality is that most people don't have the courage to go for moonshots because there's a high probability that they won't work.

But if you think about this for a second, after the horror of the October 7 attacks, the chaos in Syria, the attacks by the Houthis, and the escalations fueled by Iran, the idea that the president, Witkoff, and, as you're right, Chamath, our friend Jared Kushner, could have overcome all of that is remarkable. That could have been all-out war across the Middle East. It could have plunged the entire Middle East into utter chaos.

But instead, he decapitated Iran. We brought Syria into the fold. We're on the verge of expanding the Abraham Accords, and now you have this historic signing. We saw it with our own eyes when Sacks and I were over there on the president's visit.

He is respected as a strong leader and liked in every one of these capitals that you go to, and it's paying off. If this happens, Witkoff, Kushner, and the president all deserve the Nobel Peace Prize. If even 80% of it works out, I think it's worthy of a Peace Prize.

We're probably going to have—at least the rumors are that there could be—a China deal coming up at the end of this month or into next month. If we could resolve the issues with Taiwan, even something like strategic ambiguity, and if there's a chance of Ukraine and Russia being settled, now you've got the trifecta. You've got the 3 biggest hotspots taken off the table.

If you can get 2 of those, it would be great for humanity. Polymarket is showing a 90% chance that all Israeli hostages will be returned by the end of the month, which would be tremendous. Also, getting aid to these poor people suffering in Gaza would be nice to see. Anybody else have any thoughts as we move on to our next story?

I'll just say that the funny part here—if there's anything funny about the story—is watching the entire global left, which has been demanding a ceasefire in Gaza for almost 2 years, suddenly go strangely silent now that President Trump has seemingly engineered that ceasefire.

I guess the Nobel Committee is announcing this Friday. I do think that a record like this would absolutely win any other president the Nobel Prize, or at least get him a nomination. President Trump has ended 7 wars in 7 months before this one. This would be the 8th if he pulls it off.

7—Obama won with nothing.

He just got elected and he got a Nobel Prize. It was just kind of like a gold star.

Yeah. Okay. Well, let's move on to our next topic. The National Guard has deployed troops into Chicago to protect ICE agents, and Portland might be next.

Obviously, a bunch of protests have broken out in Chicago after a number of these ICE raids have occurred. Over the last month, DHS has been executing something called Operation Midway Blitz in the city. Eight hundred illegal aliens have been arrested.

There was a notable raid on September 30 by Border Patrol and FBI agents in Chicago's South Shore neighborhood, where 37 illegal aliens were arrested in an apartment complex. Wow. This included flash-bang grenades and a Black Hawk helicopter, similar to what we saw in Los Angeles back in June.

Trump has 500 National Guard troops standing by from Illinois and Texas in case any protests get out of hand. Trump says they are needed to protect ICE agents who are being targeted.

Illinois Governor JB Pritzker and Chicago Mayor Brandon Johnson are obviously opposed to ICE actions. They are calling Trump an authoritarian and fighting for states' rights in this regard. They filed a lawsuit to try to stop it, and a hearing is scheduled for Thursday.

Obviously, the president had some legal losses in a number of these National Guard activities, and we can get into the details of those. Mayor Johnson also signed a couple of executive orders aimed at slowing down ICE agents.

They've established some ICE-free zones across the city that prohibit ICE agents from using any city-owned property. Another one bans city employees from aiding ICE unless required by a criminal warrant.

The White House called this a “disgusting betrayal” and said Johnson was prioritizing illegals over U.S. citizens. Your thoughts, Sacks?

Well, I think the place starts with the Democratic attitude toward crime. We've seen many examples of this recently. You had the murder of that young Ukrainian woman, Iryna Zarutska, on a subway by Decarlos Brown Jr. He was arrested 14 times, and still, his judges and prosecutors refused to punish him or keep him off the streets.

There was testimony from Steve Federico before a congressional subcommittee, in which he recounted how his daughter Logan was murdered by Alexander Dickey, a man with over 25 felony arrests in San Francisco. Here you've got leftists still trying to get diversion for Troy McAlister, who killed Hanae Miyuki and Elizabeth Platt on New Year's Eve 2020. That was the case that activated all of us to get Chesa Boudin recalled successfully.

You have so many examples of leftist prosecutors, judges, and activists always trying to get repeat felons off, trying to get them diversion, and never prosecuting them. They're in favor of zero bail, and they've turned the jails into a revolving door. That's why you see enormous amounts of violent crime in places like Chicago and Washington, D.C.

In D.C., the president had the ability to clean up the streets and the homeless encampments and make the streets safer. No one denied the feds had a right to go in, and they did. There have been immediate positive results. Homicides, carjackings, and so forth have fallen dramatically.

Everyone who actually lives in D.C. is very happy about this situation, and the restaurants are full again. People feel safe to go out. So the question is, where else can the president help these cities and go in?

In Memphis, they have been invited by the Republican governor to go in and help, and so they will. With Portland and Chicago, the problem is that, frankly, you've got mayors and governors who are opposed to doing anything about the crime problem.

Moreover, they refuse to support ICE and DHS in their mission to enforce immigration laws. Finally, you've got violent leftist protesters like Antifa who've taken to the streets, and they are trying to thwart deportations of violent aliens by ICE and DHS.

In response to that, the president has sent somewhere between, I think, 300 and 500 National Guardsmen to support ICE in Chicago. It's not a broad-based mission to clean up the streets because we still have to figure out what the legal authority to do that is. They're just there to support ICE.

Yet you see all this hysterical bluster and rhetoric by JB Pritzker and other Democrats, basically claiming this is authoritarian. It's not. It's a very limited operation to support ICE and DHS in their lawful mission to enforce immigration laws.

There's no question that the president has this authority. Both Eisenhower and JFK sent federal troops into the South to enforce federal desegregation laws. For example, when James Meredith was barred from attending the University of Mississippi, JFK sent in 30,000 federal troops to bust open the doors of Ole Miss.

By the way, they were not National Guardsmen. These were the Army. Presidents have sent in troops. They have sent the National Guard to enforce the law in American cities. He absolutely has the right and power to do this.

Like I said, this is somewhere between 300 and 500 guardsmen sent to back up ICE, which is being assaulted by violent protesters and rioters from Antifa. What you're seeing from the mainstream media and liberals is an attempt to mischaracterize the situation.

All right, just to put some numbers on that for the D.C. National Guard being called out: Obviously, Trump, as the president—whatever president it is—clearly has the right to do that for a certain period of time.

If you look at this Washington Post story, I'll send it to you for a post, 61% of people said in D.C., according to the Washington Post survey, that the military presence has made them feel less safe. Seventy-nine percent of people in D.C. in the same survey said they don't want the National Guard in the city.

So there is a lot of conflicting data here. But overwhelmingly in D.C.—

That's an opinion.

No, it's an emotional—

A survey. I'll give you a different survey.

If we're going to do surveys—

Yeah.

Yeah, sure.

There was a poll last week from TIPP Insights that showed that Trump had a net positive rating among voters in cities, 47 to 44, and that his numbers have improved recently.

Look, I think this is tremendously popular with the citizens of these crime-infested cities. We saw this with D.C.: The locals, the residents, including among the Black population, were very happy with the fact that Trump sent in the National Guard and crime has been massively reduced.

Sorry, can I say something? You know why it's popular and it's working? Because of what Wes Moore did. Wes Moore is the perfect example of what I think a Democratic governor should be doing, which is, obviously, he has to blather on with his anti-Trump rhetoric, but what does he do behind the scenes? He sends in state troops, and he makes sure that there's a surge of policing in all of these crime-ridden areas because he knows that the policy works.

He just wants to get the credit for it, and he doesn't necessarily want somebody else to get the credit. This is where I think what's happening in Illinois is a bit of a head-scratcher, because if you see how badly run and crime-infested Chicago and the state of Illinois are, why wouldn't J.B. Pritzker do what Wes Moore does? I think the honest thing that they should be doing is recognizing that the citizens on the ground in these places want to live in a safe and peaceful place. With more policing and more troops, it's just statistically true that crime comes down.

At the end of the day, probably the citizens of these places care less whether they're National Guard troops or state troops; they just care that they're there. This is why I think asking emotional reactions to me doesn't make as much sense, because I feel it's a very partisan way of opining on the person who gives you the resource. I suspect if you ask the exact same question inside of Baltimore about the state troops that were sent in by Wes Moore, a Democrat, you'd probably get a different response. But the outcome is the same: crime goes down. That's what everybody wants.

Well, listen, nobody wants abusive tactics, but let me just give you some numbers that cause me to believe this feels more like a bipartisan issue, and we have the pendulum swinging back more than something that should cause us concern. Court-ordered removals and deportations averaged 200,000 per year from 1997 to 2000. That's under Clinton: 200,000 a year. They averaged 300,000 a year, peaking at 400,000 under Obama.

Deportations plummeted under Biden. At the same time, we know that illegal crossings into this country exploded higher. Now they estimate that Trump is back to the trend line of 300,000 to 400,000 deportations and removals of illegal immigrants. These are court-ordered removals. I might have expected the Trump number to be way higher, given that we just went through this massive step-up in illegal crossings, but I think it's really important to point out that this has been a bipartisan, consistent thing that has happened in this country for 30 years.

I'm not talking about the tactics or exactly how it's going on, but deportations have been carried out by both parties, and it hasn't been an issue before. So, it is curious that it is more of an issue now. The other thing I would just underscore is that if you look at what Daniel Lurie is doing in the city of San Francisco, he is live-blogging on Twitter and Instagram the arrests of 100 fugitives per night. I think he's had 3 or 4 of these nights now in San Francisco, taking these people off the street following a mayor who wouldn't even arrest people, with no prosecutions, because he understands as a Democrat that what citizens expect first is safety.

They want the ability to walk their kids to school, to walk to dinner, to play in a park, and not have to worry about being safe. I think I build on what Chamath said. The smart Democrats are realizing this is a bipartisan issue. This is what people want. It's not inconsistent with what Presidents Obama and Clinton did in their administrations in terms of deportations. I don't think this is the hill that they want to die on. JCal, I'd love to hear your thoughts on what's going on with these deportations.

Well, I've been very vocal that I don't agree with the violent nature of the deportations and what ICE is doing. But obviously, since Sacks and I worked together on the ouster of Chesa Boudin, I care—coming from a law enforcement family—about law and order. I think it's super important, and I obviously agree, but I want to level the conversation up a little bit and maybe talk about the president's approval ratings and how this all connects.

I've got a couple of charts I'll share with you guys here and get your reactions. It's not pretty for Trump right now. His polling is at an all-time low. It's worse than the last couple of seasons of The Apprentice, in fact. When you look at this chart here of Donald Trump's approval rating, you'll see there are 4 dips here. This is the net approval rating.

What you'll see is Liberation Day, when he did the tariffs. Obviously, he went a little too far on the tariffs, and that was pretty shocking. It went down to -10%. That's taking his approval minus the disapproval, and that's how you get the net approval. Then the LA protests—remember the violent protests and the ICE agents chasing people down in fields? Americans didn't like that either.

Not releasing the Epstein files in July caused another dip, down to the lowest point of his presidency, about -10%. Now, with these Chicago ICE raids, I believe the violent nature of them is the cause of this. Going on a little bit even from here, if you look at his wheelhouse, this is where Trump won the election. I think we all agree on the Trump 2.0 platform, which won a lot of moderates like myself.

I know on this show people like to say I'm a Democrat. I've actually voted Republican almost exactly as much as I've voted Democrat, and I am a strong moderate. If you look at immigration, his net approval rating has plummeted from +10% to about -5%. On the economy, he's down 15%; he was up 5% at the start. On trade, he's rebounded a little bit. On inflation, he's down 27%.

This is his wheelhouse. If you look at Bari Weiss's organization, CBS News—congratulations to Bari on taking it over and on her sale of The Free Press to CBS—only 17% of Americans believe Trump is making things better. This came out this week, and 58% don't want the National Guard in our cities. People oppose what Trump is doing right now, and he is at an all-time low.

So you have to ask yourself why. I think this is because he's drifted from the 2.0 policies: closing the border, which is very popular; lower taxes; no taxes on tips, which is extremely popular; and being pro-business and pro-innovation. Thank you, Sacks. You're correct. This is all the best of Trump. DOGE reducing spending—the best of Trump—and relentlessly trying to stop wars, apparently on the cusp of succeeding with the one in Gaza right now.

This is what Americans, including myself and moderates, really like about Trump. What don't we like about the Trump 1.0 platform? The violence of January 6, ICE agents in masks beating powerless, hardworking immigrants, sending in the National Guard, if you remember from the first administration, the kids in cages, the Muslim immigration ban, and pardoning all the January 6 rioters who beat police savagely while trying to overturn election results. This is what people don't like. They don't like the chaos.

What Americans do love is the Constitution. They love the rule of law. This is where he can do really well. 85% of Americans have a favorable opinion of the Constitution, and 94% say it's really important to protect liberty and freedom. When you see Pam Bondi and Stephen Miller, Trump's ratings go down. When you see Howard Lutnick, Scott Bessent, Sacks, Elon, and DOGE, his approval ratings go up.

I would like to lobby my friends who are in the administration or around it, and the people I know in the administration, to think through this, because what's at stake is the midterms. The Democrats are going to shellac the Republicans in the midterms if this continues. Nobody wants this type of violence. Nobody wants to see people beaten. Nobody wants to see mothers thrown on the ground as their children are crying just because they came here 20 years ago or 10 years ago, when Republicans were the biggest proponents of letting people in from Mexico to work here because we needed cheap labor.

We as Americans have an obligation to these people who we brought into this country to treat them with compassion, and we're not doing that. That's my little monologue here. You heard from Brad that the deportations have basically mean-reverted. Do you think that there were the same categories of people being deported before as now, and that there's just less press coverage—that it was okay then and it's not okay now? Or just explain this idea of mean reversion and whether we're getting the same and right people out of the country that we've been getting out for the last 20 years.

Yeah. What you're seeing is many lawsuits now from people who are Americans or properly documented, suing ICE, because the way they're pursuing this—with their masks and just randomly going after people and racially profiling people—is resulting in the wrong people being picked up. That's really what I find the most offensive about it.

I don't mind, obviously, the border being closed. I've said that many times here. I don't mind criminals being deported. I think that's a great idea. But I think we should look at the economics of this as well. We're spending $30 billion on ICE now. We've tripled the budget. We're spending $100,000 per person deported. That's a large number.

If we simply fined that 79-year-old car wash owner $10,000 every time he hires an illegal alien, this would stop. There are much better ways to execute this. I have to ask myself: What is the intent of this violence? What is the intent of sending the National Guard in? Is it actually a pure intent of wanting safety and security? I don't believe that's all that's at work here.

And then a follow-up question: What do you think is the difference between Wes Moore, the governor of Maryland, sending in state troops and Donald Trump, the president of the United States, sending in federal troops into a state to help diminish crime?

Yeah. It would be the Constitution and the United States of America—not a federal government ruling over all 50 states.

The way this was set up by the Founding Fathers was that the states had rights, and among those rights is the security of the people in that state. It's a very rare situation.

So your issue—or not your issue, but your diagnosis—would be that what Wes Moore is doing is right. JB Pritzker should probably do that too. I would even take it a little further.

Yeah, I think actually the best thing to do if Trump is intent on saying, “Hey, crime’s out of control in San Francisco,” which we all experienced and you guys continue to experience if you’re in the city—

I don’t live in the city.

But if people did go to the city, I think what I would do if I was the governor is say, “You know what? We could use your help, and we’re going to call up the National Guard ourselves. We could use some help in these 17 locations. We’d love to collaborate with you.”

So it’s not an outcome thing for you. It’s a procedural process that says it needs to go through a different pathway.

It’s the violence and the cruelty of it that I object to, and it’s the inefficiency in terms of the cost of it. I don’t believe that mothers should be separated from their children or that fathers should be separated from their children. I don’t believe it should be done in a brutal way, and it doesn’t need to be. It could very easily be done with other tactics.

That is what I take great offense to: both of those things, the inefficiency of it and also the derailing of the Trump 2.0 agenda, which I was in favor of. I will call balls and strikes when Trump tips back into those authoritarian tendencies of the Trump 1.0 agenda that I’ve talked about. I’m not in favor of that. I don’t like it.

I would much prefer to see us be compassionate to immigrants. I would much rather see us take a softer hand and a path for these folks who’ve been here for 20 or 30 years. I’d like to see a path for them to become citizens, because I think this country was built by immigrants, for immigrants, and I believe we should continue that.

But it should be legal immigration. If people were here illegally because Republicans and Democrats who are super in favor of NAFTA wanted them here, we should take ownership of that and be kind and compassionate to them. If they’re not criminals and they pay their taxes, which almost universally they do, we should give them a path to citizenship.

All right. Well, the first thing I want to do is point out that you mentioned that DOGE was one of the most wonderful and popular things that President Trump did. I agree that it was a great thing that he did, basically trying to streamline the government and make it more efficient. But it was not popular. It had only a 35% approval rating, while 57% disapproved.

Why is that? Because the media pounded on DOGE and Elon, turned him into a villain, tried to mischaracterize what he was trying to do, and blamed him for every possible thing that was supposedly going wrong in the federal government. Obviously, the polling is not always a great indicator, even when it is accurate. I remember when Ann Selzer told us that Trump was going to lose Iowa by 10 points. Obviously, these pollsters often have no idea what they’re talking about.

In terms of the merits of this, you’re completely ignoring the fact that there are lawless mobs who are the ones creating the violence—not our law enforcement officers, not ICE, not DHS. In the city of downtown Portland, for example, you have literal Antifa terrorists who are extremely organized. They’re aggressive and they’re well-funded. This is not just some sort of ragtag bunch.

The White House uncovered a bunch of left-wing NGOs who are funding Antifa along with other leftist protest groups. These NGOs have laundered over $100 million of our money to fund terrorist violence in our streets. Antifa even has a safe house near the epicenter of the unrest in downtown Portland that local cops and local media know about, but nothing’s being done.

It’s this group that has basically terrorized downtown Portland and led to violent assaults, not our law enforcement officers. But when this violence breaks out because of Antifa, the media blames the Trump administration for it. I’m not surprised that the polling reflects that, because that’s what the media does.

Look, the bottom line here is: What is the policy that the Trump administration should pursue? Previous administrations have pursued deportations, and so has the Trump administration. I don’t think you can have 10 to 15 million illegal immigrants allowed in during the Biden years and then just say there’s going to be an amnesty. What are we supposed to do?

This situation was not created by Trump. It was created by Biden. They opened up the border. There were holes in the border walls. Millions of people streamed through. They were bused all over the country. Trump won on fixing that problem, and deportations were part of the mandate. Now he’s doing it.

It can’t just be a one-way ratchet. When the Republicans are in power, they close the border. Then, when the Democrats are in power, they open it up, allow 10 to 15 million illegals in, and then we can do nothing about it when we take power. No, that’s not the way this works.

There has to be a correction to what Biden did, which was outrageous. I remember plenty of liberals, while that was going on, saying that the videos we saw on Fox News were cherry-picked. They said that there was no real situation at the border. We were gaslit constantly about it, whereas every single person who actually visited the border said the same thing: that it was outrageous that Biden had opened up the border and that people were streaming through. They were running through, and Trump won on a mandate of fixing this problem.

Now that he is fixing it, people claim that he’s the source of the violence. No. Antifa and left-wing rioters are the source of the violence. You don’t see this violence in red cities, and you don’t even see it in the blue cities that have offered to cooperate with Trump, like Memphis or D.C.

When you see cooperation from these cities, they are actually cleaning them up. But in Chicago and Portland, you have violent protesters, and that is the source of the problem. The media can blame it all on Trump as much as they want, but fundamentally, that is what’s going on.

Okay, let’s move on to our next topic. Anybody else? I’m surprised, as someone who normally supports law enforcement, that you’re not backing these guys up.

Oh, I support law enforcement. I do not think they should do it the way they’re doing it. It’s the violence and the cruelty that I object to.

Okay. I don’t mind people being held accountable for crossing the border. I don’t mind the deportations.

How do you think it should be done, Jason?

Yeah. A very simple method would be to go to the people who are employing them, because the people who are here illegally are here to build a better life for their families and to live the American dream. That’s why they’re here. They’re here to get a job and have a better life for their children if they are working illegally.

These are the most industrious people I’ve ever met in my life: immigrants, including the two of you. The most industrious people are immigrants. We know that.

A very simple way to do this, if you wanted to do it without cruelty and without spending $100,000 per person, is to go to where they’re being employed. If you saw the terrible video of a 79-year-old man being tackled by 2 ICE agents, and he’s suing the government for $50 million, that’s not the way to do it.

You go to that car wash and say, “I’d like to see everybody’s papers.” Everybody runs. They show you the papers. Whichever person is not actually employed there legally and doesn’t have their papers, you give a $10,000 fine. Then you come back the next day and give a $20,000 fine. If you did that, this would be incredibly effective, and you wouldn’t have the brutality you have now.

I’ve talked to many of my relatives who are cops and many of my friends. They look at what ICE is doing and see it as unnecessarily brutal. All you have to do is go to where they’re working, give the fines, and then round up people who are, in fact, criminals.

I have no problem with any gang members or felons being taken down, and sometimes you have to do that with force, obviously, if they’re criminals. But for everybody else, we can do this compassionately, and we can do it much more efficiently.

We could do it by just paying people to self-deport, and we could increase that rate. That actually has been working. The story hasn’t been told yet. I think we’ll see in the statistics that of the 300,000 or 400,000 people we wind up deporting, probably a third of them are going to be people who self-deported. We’ll share those statistics here as we go, but it’s the violence—

I’ll just note that D.C. is the city where Trump was first able to get involved in trying to stop the crime, because he was able to bring in the National Guard and everyone understands that the feds run D.C., so he has a free hand.

Have we seen any of the problems that we’re talking about in D.C.? No, we haven’t. It’s calm. People can go out at night. They’re going to restaurants again. The citizens are happy.

Why? Because the things you’re talking about are the result of violent Antifa protests in Chicago and Portland. The local politicians, like JB Pritzker, have an incentive to blame everything on Trump, and they’ve been turning a blind eye to these local gangs that are creating all the problems.

Yeah, we’re in strong agreement that Antifa—and I’ll add to it, the Proud Boys and Oath Keepers, who brought tons of guns to the Capitol and ran January 6.

I agree with you. All of those radical organizations should be handled. We're in strong agreement.

The fact that you're bringing up the non sequitur of January 6 tells me that you've lost this debate.

I'm not looking to win a debate with you, Sacks. I'm not in the debate club, but you can win whatever debate you want. I'm just giving my opinion. But, to give a little nuance here with the DOGE thing, I agree that it was unpopular with the libs, but what I was talking about was the moderates who won Trump the election this year, which you had a big part in, David.

You turned a lot of the moderates—the fiscally conservative but socially liberal people in our community, the tech community. You're directly responsible for that. I might argue this podcast, and you and Chamath and your support of Trump and the Trump 2.0 agenda, were a big part of him getting elected. I've heard from people in and around the administration that they believe that was a key part of their success this time.

So, when I say that people were in support of DOGE, I'm talking specifically about moderates and the people who backed the Trump 2.0 agenda, just to clean that up. I do agree with you.

Yeah, but you can roll your eyes. Well, my point is—

I'm not trying to win a debate with you. I'm just telling you my opinion.

Polls can give us interesting information, but they're also a snapshot in time. What really matters is the results of the policies.

And I think that if you look at D.C.—hold on—you look at D.C. just a couple of months ago, before Trump went in, the left was basically shrieking hysterically that it was authoritarian, that it was fascist, and so forth and so on. They had all these arguments that it was going to be brutal for the population. Trump goes in there, and it's worked amazingly well.

I'll let you have the word.

I don't think we should be abandoning the poor populations of these inner cities to the predatory crime that they've been subjected to for decades, which their Democratic politicians have done absolutely nothing about. Quite frankly, these cities are something like 90/10 Democrat. If we can just make them 75/25, then I don't think Republicans will lose elections anymore. So, I'm glad Trump is trying to do this. We've seen these examples over and over again.

This is why Wes Moore is doing what he's doing.

Exactly.

I think the tactic works, and I think, Jason, we are moving away from whether the tactic works or not to who is trying to frame themselves to take the credit. If you look at the Democratic path to the presidency, we've heard this before, but Wes is sort of at the front of the line. I think he's seen that this is the right place to be on this point.

So, I think it's like he doesn't want the success if it doesn't accrue to him. That's why he sent in state troopers in Maryland into all the hot spots. I think what will be interesting is, in 3 or 4 months from now, when we see the incremental crime statistics and whether we see Wes wrap that up in a big bow and say, “Look what I did.”

I think if you do see that, then I think what you will have seen is a successful policy that was co-opted at the state level because they want to take the credit for it, and—

You'll find an adequate number of journalists who will want to paint that as, “Oh, wow, look at this invention that has happened.” But the artifact will be that Trump initiated and instigated it.

Now, if people copy it, that's a smart thing to do if it works.

Absolutely. You cannot deny that law and order is at the top of everybody's list in terms of just living a productive life.

All right, next up on the agenda. Oh, Bestie Brad, you're going to love this one. AMD and OpenAI just closed a massive GPU deal. It could be worth $60 billion or more over the next 5 years. Obviously, these two companies are run by friends of the pod: Lisa from AMD and Sam from OpenAI, both of whom have been on the pod.

AMD stock rocketed up 35% since the announcement. OpenAI committed to purchasing 6 gigawatts' worth of AMD's next-gen GPUs. A little bit of an interesting wrinkle here: AMD granted OpenAI warrants for up to 160 million shares, or 10% of the company. With that massive increase in their stock price, Brad, I think this means that OpenAI is going to get a lot of GPUs essentially for free.

A big dialogue has started, Brad, around these round-tripping, related-party transactions. Is that how some people would describe them, specifically the SEC? Do you have concerns about the AI trade and the amount of circulation of shares, GPUs, and cash that's going on right now? I know you're in the thick of this.

Yeah. And maybe before we get into round-tripping and circular revenues, which I do want to hit, I think it's super important that we talk about that. Let's just break down the deal a little bit, because there's a lot going on, and I think it's important we put context around it.

So, first, this is a bet-the-farm bet by Lisa Su, right? She's giving away 10% of the company if the compute gets deployed. She said a couple of weeks ago, “We're in year 2 of 10 of a compute buildout across the country.” I have a couple of charts here. Why would she make a bet-the-farm bet, Jason? That's an important question.

Remember, Nvidia just did a huge deal with OpenAI. They didn't give away any of their company. In fact, they got the right to buy part of OpenAI. So, look at this chart. This is pretty wild. Just in 2022, 2½ years ago, Nvidia and AMD had basically the same revenue: $25 billion, right?

This year, Nvidia will do 10 times that, at roughly $210 billion to $230 billion, and AMD will do $33 billion, not much more than they were doing when the companies were tied in 2022. So Nvidia's captured nearly 100% of the incremental AI data center revenues over the course of the last 2½ years.

And I think there's this notion that somehow Nvidia just popped out this special AI chip. But I think if you listen to Jensen and study the company, it's because they have this ecosystem of software, networking, and extreme co-design. The unit of compute is no longer the chip, right? It's the entire data center, which is composed of 5 to 10 different chips. Performance per watt—power being the constrained resource here—is everything. Nvidia's just crushed the competition.

So now put yourself in the shoes of Lisa, right? She's clearly not on the wave. This tsunami has come. Jensen's riding it. He's capturing 100% of the wave, and she's not even yet on the wave. Her MI350 was just not competitive, and so they have one shot.

Either the MI450 gets adopted and they get back into the game, or they're out. She's a total warrior. I think she believes in the MI450. She went to her board and she said, “Listen, this is—we've got to take the shot here. We've got to bet the farm.”

If it works, she's going to get $150 billion of incremental revenue just from OpenAI for building out 5 gigawatts, right? And on top of that, of course, it could unlock a lot of the other market, because now it will have validated that they have a chip that works.

But it's far from a done deal, far from a conclusion, whether the MI450 is going to work. Can it compete against Vera Rubin? Can it compete against Rubin Ultra? All super-important questions. So that's the AMD deal.

Why have we switched from saying, “Oh, Colossus bought xAI's data center—100,000 H100s, 200,000 H100s,” to now framing these in gigawatts? I think it's important for people to understand why that's now how deals are being framed. Just in the last 3 months, it's changed.

Yeah, I think there are a couple of constraining features. Number 1, chips change, right? So 100,000 H100s is not apples for apples with the number of GB200s, Grace Blackwell 200s, or GB300s. And so it's hard to talk about them. You're comparing apples and oranges when you're comparing these data centers.

So a unifying metric of comparison is the power that goes in. Everything starts with the power. It's the constrained resource. So we can compare a gigawatt of power, because remember, it's power in and it's tokens out. We can compare that over time and normalize across all these different chips.

And it's not just these 2 companies. You've got Trainium by Amazon, TPU by Google, Cerebras, Groq—there are a lot of folks in this game. So, again, I think what you saw this week in this flurry of announcements is that you have the market leader that's basically captured 90% to 100% of the incremental demand for the biggest thing that the data center and chip market has ever seen.

Every other player is looking at what they have to do to have a shot to capture part of this wave. They're high-risk, high-reward bets. The other thing I would say is—and, again, shut me up, and anybody can jump in if they want to—but I want to decompose this.

We're hearing these estimates of 100 gigawatts, $4 trillion to $5 trillion of buildout over the course of the next 4 years. If you look at that same chart I just showed you, that's not what Wall Street estimates are, right? There is a lot of disbelief on Wall Street as to what's going to get built out.

So if you look at that estimate, starting in 2027, it basically flatlines from 2027 through 2029 for Nvidia—less than 10% CAGR in terms of the growth for Nvidia. They have, by 2029, doing $360 billion in revenue compared to $210 billion of revenue this year.

To put it in gigawatt perspective, that means going from 4 to 5 gigawatts per year in 2025 to 9 gigawatts in 2029. That's a big step up, but a long way from the 100 gigawatts you were hearing bandied about on CNBC this week.

That's for a different reason. Nick, put this quote up there, which is this famous quote from Nathan Rothschild, where he said, “I care not what puppet is placed upon the throne of England to rule the empire on which the sun never sets.”

The man who controls Britain's money supply controls the British Empire, and I control the British money supply. Why is that quote so interesting as applied to AI? I think what you're going to see—and you're seeing it in those revenue graphs—is that there's a traffic jam happening in growth, where it won't be the ability to actually build next-generation silicon, but the energy inputs and the ingredient inputs that will constrain growth.

The companies that control those elements of the supply chain will come into power and rise to power. What's one example of this? When you look at the architecture of Nvidia's chips, one of the things they've made a huge bet on is HBM, and this is a memory structure. What's so interesting about that is that when you look at the HBM market, it's effectively SK Hynix, which takes up the majority, and then Samsung. Now AMD's next architecture actually needs to sit on top of HBM.

They're going to have to step into that supply chain and try to ask for share. Where will that share come from? This is where the people who control that supply—SK Hynix and Samsung—will have leverage. This is what's interesting about a deal that OpenAI announced 2 or 3 weeks ago. Sam was in Korea, and you saw him shaking hands with SK Hynix and Samsung.

My initial thought was, why is OpenAI doing a deal with the memory maker? Then it occurred to me: Wow, he's buying forward capacity on HBM, because now he can allocate that share. When I saw that deal and that equity, it reminded me of this Rothschild quote. Sam has allocation, and now Sam can allocate allocation and, as a result, get a tax. I believe the warrants and the equity are effectively that.

What's a different example? Brad mentioned this before, but energy will be the gating item, beyond a shadow of a doubt. If you control electrons—any form of electrons, hydrocarbon to electron, electron to electron, photon to electron—it doesn't matter. You will then be in a position to start asking for equity, upside, and participation in these companies in a way that you could never do before. You would have just been a linear member of the supply chain, a low-margin participant.

When I look at where we are, I think the really interesting question now is to ask what the second- and third-order inputs are that are critical to allowing the big foundational model makers, Nvidia, AMD, and Broadcom to thrive. That is where I would start to look, because those who control those resources are going to dictate the pace and scale of this AI expansion.

Okay, coming around the horn here to you, our czar of AI, David Sacks. When you see this massive amount of dealmaking occurring, it's got to warm your heart a bit. American exceptionalism at work here, people swinging for the fences. What's your take on Brad and Chamath's overview of where we're at here in the fall of 2025 in the AI race?

Yeah, I don't really like to take sides on these deals, for the obvious reason that we want to be supportive of everybody and just have a healthy environment for competition. As long as there's investment going on and as long as there's a lot of competition, those are good things. That's what we want to see.

I tend to think this OpenAI–AMD deal is evidence of that. There was the OpenAI–Nvidia deal. There was the Nvidia–xAI deal. There's just a ton of investment going on, and that's what we want to see. That's all really good news.

I was struck by Brad's chart about Nvidia's revenue in the out years. The amount of capex, or investment, that's going to go into data centers and compute more than a few years from now, I think, is really hard to predict.

Impossible, right? Yeah.

Very hard, because on the demand side of compute, it's going to depend on the new applications that get created. For example, demand for tokens depends not just on AI chatbots, but on the new agents that are coming along. You have these new video-generation models, Sora, and now xAI has just released something.

We don't really know what the demand for tokens is going to be. I think it's going to be huge. I think there are a lot of applications that haven't even been invented yet. We are in the very early stages, and those applications are going to drive demand.

It's a little bit like in the early days of the internet. We had this dot-com bubble where everyone thought there had been too much fiber built out, and then all the fiber ended up being used because social networking came along, Facebook came along, and YouTube came along. When all that original fiber in the '90s was built out, they didn't even know that photos and videos were going to be a thing. The bandwidth eventually all got used up.

I think in a similar way, all of the compute will eventually get used up by new applications. I guess the thing we don't know is how much more efficient the infrastructure will get at producing these tokens. That's the offsetting factor: The chips are getting so much better.

We're moving faster than Moore's law here. Every year, Nvidia is releasing a new generation of chips, and it's doing inference at multiple times the efficiency of the previous year, on, I guess, a token-per-watt basis, or whatever you want to measure it.

That's an offsetting factor. There could be more efficiencies created in the model architectures with sparser architectures. Those will be offsetting factors as we get all these efficiencies. But if I'm arguing with myself, you've got Jevons paradox: As the cost per token comes down more and more, that's going to enable AI to be used in more and more contexts where previously it wasn't economically efficient. Now it will be, so that's going to fuel the demand.

I think this is a huge boom.

Yeah, your point is a great one. Everybody's looking for the bubble, right? We've got this pattern recognition. Everybody's like, “What? It's dark fiber. It's got to be dark fiber. There's no way that you can 10x your revenues at Nvidia without it being a bubble.”

But the reality today is that Nvidia trades at a lower price-to-earnings multiple than it did when it was $200 a share. Think about what the term “dark fiber” means. We were laying fiber in the ground that remained dark. We did not have the demand for the fiber. There is not a dark GPU in the world today.

There's not going to be a dark GPU in the world next year.

Good point. There's no dark GPU.

I don't think that's the question. I think the question is, what is the ROI on input and output tokens per use case? What is the ROI of a random use of a model? I think that's a very good question.

How many of those tokens are generally gross-margin negative? How many of those are unprofitable? How many of those are profitable? I think that's what people are trying to ask when they ask that question. They may just not know the technical intricacies to ask it that way, but I think that's what people are trying to debate when they're trying to guess. I don't think we totally know yet.

There are other solutions that have not yet hit the ground. I mentioned Bittensor. I think it's an open-source project that's worth looking into. I started a little fund of my own with my own capital to start exploring this, because there are going to be other solutions that come to fill all of this Jevons paradox that's happening. Chamath, your thoughts?

Yeah. Well, I think the thing to remember is that token generation is this very complicated quadratic that just consumes a ton of power. There really is no way to ignore the fact that, even with more complicated capabilities, the models are becoming much more complicated. These context windows are getting longer or bigger.

I just think that we are in this massive capital race. Unfortunately, I think what that will create is the need for all of these other things. For example, there may be architectures where you just have to use SRAM because you can't get HBM. I don't think we know what that does to a whole bunch of these use cases today, because it may just be completely unperformant.

Separately, I think Meta was the one that published this: There's a big disconnect right now in the market around the failure rates of certain classes of hardware, and not at the chip level, because at the chip level these things are really performant. But at the server-rack system level, these things are failing 9% to 20% of the time. That's very complicated. How do you deal with all of that? How do you deal with these cascading failures, all these things?

We are pushing the boundaries of physics. We're at 2- and 3-nanometer scale. Yields are good. All of these things, I think, have to get worked out.

And that's what makes it such a great opportunity and so exciting, I think. Brad, let's talk about the TAM. I've been working a bit on what the TAM of this is. In the modern developed world—what we used to call the First World—you've got about 500 million business users. They spend, on average, Brad, about $3,000 on SaaS products.

If you were to just put those numbers together, it gets pretty large pretty quick: trillions of dollars in revenue for those users. If you look at consumers and just use an analogy—maybe you'd be interested if you think this is a valid one—consumers pay for Netflix and Disney+, whatever it is, probably a couple hundred dollars a year.

Then you have 1 billion users in the developed world. I'm not counting frontier markets and the developing world. We're talking about a TAM that's going to be a couple of trillion dollars a year. I think you would agree.

So the buildout of a trillion dollars doesn't seem as farcical if that is, in fact, the prize. I think I've set the table very well.

Wait, hold on. It's not a trillion. That's just the capex; then you have to think of what the opex is. The opex is probably the same multiplied by the useful life, which is probably 20 years. So, if you think there's a trillion of capex going in, you have to expect tens of trillions of opex over the useful life of all this stuff: upgrade cycles, people, power, and water. It's a much bigger spend than we think, so we have to be careful that we're not underestimating, frankly.

Absolutely. And then there's the useful life of these, right, Brad? What is the useful life? CoreWeave thinks these things last 4 years?

Well, they underwrite to 6.

They're at 6. Other people are saying 4, so that seems a little challenged, too.

Here's a rule of thumb: each gigawatt data center is about $50 billion of investment, roughly, between the chips, the land, power, and shell. It's about $50 billion per gigawatt. And you have OpenAI and Elon talking about scaling to 10-gigawatt data centers. That's a $500 billion investment per data center if it gets that big.

I think that's right—per data center, if it gets that big. To be clear, they're still working on the 1-gigawatt data centers, but from there they're going to go to 2 or 3 gigawatts next year and then probably to 5 over the next couple of years. So you could get to a 10-gigawatt data center in, I don't know, probably 5 years, something like that. Brad's kind of convinced me that we should all be taking the over on this, as he usually does.

Welcome home, Brad. We appreciate you. When I think about Nvidia's market, I think the big question to me is—I would bet the over on demand because I think that it's very hard to underwrite demand for applications that don't exist yet, but I'm confident they will be there. The thing I don't know about their business is whether the GPU market bifurcates between training and inference. I think everyone understands that Nvidia has by far the best training chips, but there are a lot of competitors, including Chamath's company, Groq. There's Cerebras, and obviously AMD wants to compete. Huawei is kind of a wild card there, and we can talk about that.

I should also mention all the ASICs that the big AI companies are using, using Broadcom IP. You've got the Google TPUs, and Brad, you mentioned Trainium, which is Amazon's chip. I think OpenAI wants to develop its own chip. So I think the question is, given that inference will be 99% of the market and training is only 1%, could Nvidia be in a situation where they've got training locked up, but when it comes to inference, there are just a lot of other alternatives because people have developed cheaper inference chips? I'm not saying that's going to happen. To me, that would be the big question.

Well, listen.

I think, first, I go back to the point you made. We're at this critical juncture in a global AI race, and we should all be celebrating the fact that there is this level of investment going on in the United States. This is the definition of creative destruction. It won't all work, but the risk capital and the entrepreneurial boldness of Sam, Nvidia, and Lisa—this is what's going to keep us ahead in the global AI race.

None of this would be happening if Washington had not unlocked power. Jensen said this on my podcast—or our podcast—last week. We've unlocked power. We've unlocked capital. People are now able to underwrite to these deals. First, I just think that this is incredible. We have this level of competition in terms of Nvidia's ability to stay ahead of everybody else.

Listen, I think at the end of the day, it's power in and tokens out. It's perf per watt. I've talked to many CFOs of hyperscalers, and they said you could price the competitor chips at zero. Jensen has said this publicly many times: price the competitor chips at zero, and it's still a more effective and economic decision to choose Nvidia. I think that's what it comes down to. They have to be at perf per watt because power is the constrained resource, going back to what Chamath said.

Okay, I want you to just give me a number, Brad. The total TAM of this in 5 years—what's the spend going to be in the modern world? Just think about that for a second.

I think, Jason, you made a great point about total TAM. We see about $100 billion already in incremental generative AI revenue in 2026. That has to grow to well over $1 trillion by 2030. If you look at it on a top-down basis, with $125 trillion in global GDP, if you get a 10% improvement in productivity, that's $12 trillion a year of productivity improvements. If you get 5%, that's $5 or $6 trillion. That's what's leading Masa, Sam, and others to place these heroic bets.

Okay, great. We just have to lightning-round this. Here's the round-tripping discussion that people are obsessed with: Nvidia at the center of this, a $4.5 trillion company, the largest company in the world. As you can see here, between xAI, OpenAI making language models, AMD, Intel, CoreWeave standing up data centers, and so many other companies—Oracle is obviously in the mix with its cloud, Google with its cloud and Gemini, and they have their own chips as well—what do we take away from what Chamath and I saw up close and personal with OpenAI and its partnerships?

I guess I'll start with you here, Chamath. Is this analogous to the round-tripping we saw in the dot-com era? Are these conflicted-party transactions concerning to you at any level? Is CNBC and everybody hand-wringing about this over their skis, or is it a legitimate concern? I'm not sure that they're framing the concern properly.

There are a lot of industries where these things are standard practice and well accepted. One example that was explained to me this week was how the auto OEMs dealt with car dealers and their car dealer networks, where you would give what's called a floor plan loan. If you're trying to sell a Ford F-150, you give your dealership a loan. They then buy the car; it pulls forward the revenue for Ford. Then the dealer goes and sells the car to you, and they take the arb and pay the financing cost.

So this kind of stuff has been well established for a long time. What I would rather ask is, where is the PCAOB expert that says this is not good? I just think that these are too complicated for me to understand. They get into the vagaries of accounting law, but I suspect that these companies are well advised, and I think there are other industries that have been doing it for many years. If there were a real issue, it would have been exposed much sooner, quite honestly.

You say PCAOB—Public Company Accounting Oversight Board. I just think there's all the quarterly regulatory filings that are required, all the Sarbanes-Oxley compliance that's required, and the number of auditors that are looking at these things. I think it does come down to the fact that there are many industries that use these forward-flow agreements, if you want to call it factoring or forward-flow agreements, if you want to call it that.

I just think that it's newer in our industry, but it's important to acknowledge that it's very well established in many other legacy industries. There's a lot of it all of a sudden. So I guess, Brad, the uniqueness to our industry is one piece—or I'll give it to you, Sacks. The uniqueness to our industry is one piece, and the number of these is just making people's heads spin a little bit and makes their mind wander. But conflicted-party transactions have to be very well documented in SEC filings, et cetera. These are largely public companies. But go ahead, Sacks.

Well, look, these companies all have lots of very expensive lawyers and accountants. So I'm sure that from a GAAP standpoint, a securities standpoint, or a compliance standpoint, what they're doing is fine. Now, when people talk about round-tripping, if you want to talk about the spirit of it, the underlying concept behind it, I think what it comes down to is that Nvidia is effectively extending credit to its buyers. Why do they need credit? Because this buildout is so capital-intensive.

Elon, every model is quadratic. It gets worse.

They're all running around the world looking for capital. There was a photo of Sam the other day. I forgot what country he was in. It might have been the UAE or Saudi Arabia or something. In any event, they're all looking for as much capital as they can get to fund this buildout. In certain cases, obviously, Nvidia is extending effectively credit in exchange for equity or whatever it may be.

I think the question you have to ask is simply: are these companies that are on the receiving end of this credit effectively creditworthy? In other words, will this investment yield economically viable results? Because if, let's say, Nvidia extends credit to OpenAI and then OpenAI buys chips from Nvidia, Nvidia recognizes that as revenue. The question is, is there economic substance to that, or is it basically a shell transaction?

My view on it is that there is substance to the transaction because there is downstream demand for AI and the applications that are eventually going to get built, and for the APIs that are being put out there—for basically the tokens that are being generated. I mean, Brad will have the exact numbers, but I think OpenAI's revenue ramp rate is something like $5 billion to $25 billion to $100 billion. Brad, what can you tell us about the years on that?

Well, I don't want to get over my skis about what's been disclosed, but I think they'll exit this year at over a $20 billion run rate.

I think there are a lot of people who think they could hit $50 billion next year and $100 billion after that. So this again gets back to your point, David. There are no dark GPUs in the world today. The tokens are being consumed. Every consumer in the world wants answers. They don't want to use 10 blue links.

Every enterprise wants to deploy this to make their enterprise better. National sovereigns want to do this in order to advance their military and other things.

I want to come back to this very point. We should be on the lookout for sham transactions. When you have moments of excitement like this, they of course will lead to somebody on the edge doing stuff that is a sham. What's a sham? There's no economic substance. There's no end buyer for your product.

So I call up Sacks. I say, “Sacks, I've got a bunch of inventory. Nobody wants to buy it. I'm going to wire you $1 billion. Buy $1 billion of my product, post it to revenue, and we're good to go.” It doesn't cost him anything, and I get to book $1 billion in revenue. Clearly a sham, clearly illegal. That's not at all what's occurring in the case with NVIDIA.

Just put it in perspective: From 2025 through 2027, NVIDIA will generate $450 billion, half a trillion dollars, in cash flow. What they're investing is peanuts. They are making equity investments in companies that are peanuts. It's not credit to these companies.

In the case of OpenAI, they don't even have to use NVIDIA chips. They just told the world they were going to buy all this AMD, right? They're rumored to be building their own chips. So the point of the matter here is, of course, it lubricates the system by making these investments, but they're tiny equity checks in these companies.

I think about it this way. CapitalG is one of the biggest investors in Duolingo. Duolingo is one of the biggest advertisers on Google. Nobody's crying foul that Google is somehow round-tripping revenue back into Google advertising, right? So long as there is end demand for the product and it's not a sham, I think these things are fine.

All right, I think we'll leave it there. There's going to be a lot more news on this.

But people love the picture of the plug plugging itself into its butt, or there's another one of a power strip plugged into a power strip.

Can you find the plug plugging itself into the butt?

The elephant. Yeah.

There it is. That's what we're talking about. Yes. I somehow don't think the on-off switch is going to make a difference here.

By the way, whenever I see this, I think, “Is it dangerous to do that?” And then I remember, no, there's actually no electricity or current.

But I always think that it is, anyway.

I think the concern we'll have is if there's a downdraft. Right now, they're saying a 10% to 30% correction in the next year is what people are buzzing about. I guess it's always some percentage of a correction.

If there's a correction, some stocks go down. The tide goes out. Then we figure out who has a sustainable business and who's wearing shorts and who's not. That's what happened exactly in the dot-com bubble. The tide went out. Amazon had enough revenue to continue. Other people did. Some people maybe were caught with no shorts in the ocean.

Whenever you have a boom like this, there's obviously going to be winners and losers, and there are going to be a lot of losers. But the question is whether the winners will pay off the losers. That's what venture capital requires.

But I just think it's amazing how much of a boom is going on, and I think Brad's made the argument that it hasn't even peaked yet. Forty percent of U.S. GDP growth this year is AI, and AI companies have accounted for 80% of the gain in U.S. stocks this year.

I know a lot of people are wringing their hands and pointing to this as evidence that somehow we're in a bubble, but to me, it's a wonderful thing that the U.S. is on the forefront of building out this technological revolution and all this infrastructure. I think it's ultimately going to pay huge dividends, and we just need to not screw it up.

Yeah.

And there's just so many people who, to be honest, just want to screw this up with excessive regulation at the state level and excessive regulation at the federal level. They want the bureaucracies in Washington to manage and approve everything.

If we go down that path, we could sabotage this. If we just allow it to happen, if we allow the innovators to do what they do best, I think this is going to drive 4% or 5% GDP growth for the next few years. It'll be a really wonderful thing.

I mean, it's a great place to wrap. If only we had somebody who had decades of experience and had a really logical bent about free markets who could advise our president, we would be in great shape. Oh, wait. That's Sacks. Good job.

All right, everybody. Gold rally is off the charts. Shout-out to our friend Vinny Lingham, who told us about this for the last year or two. Jeez, it just broke $4,000 for the first time ever this week, up over 50% for the year.

Gold's outpacing Bitcoin, which is up 30% itself, and all the other major indices. The Nasdaq is up 19% in the same time period. The S&P 500 is up 14%, and the Dow is up 9%. Silver is surging even more than gold. There might be some utilization reasons for that. Silver gets used in a lot of different components, whereas gold very rarely does.

At the same time, the U.S. dollar is down 8% year to date. Here's a chart from our friends over at Visual Capitalist about gold versus U.S. Treasuries, something we talked about here. U.S. Treasuries—Chamath, you want to take this, and then we'll go to Brad. What are we seeing here with gold? Is this a safety trade? Is this an anti-USD trade? What's behind this? Too much money supply? What's happening?

No, this is just people confusing correlation and causation and making stuff up to look smart. Why is gold up? Gold is up because there are many more net-new buyers. Who is the most important net-new buyer? It's Tether.

Tether has been issuing a stablecoin called Tether Gold, where they'll actually custody the gold on your behalf, and the amount and volume of it are rising. At the same time, central banks are rebalancing. Yet at the same time, you have a lot of macro funds that have essentially decided that central banks aren't to be trusted and they don't know what to do.

So they're not necessarily long bonds, they're not necessarily long currencies, and so they're long gold. I think it's a mixture of things. There's no panacea explanation, but if you had to put it into a couple of buckets, it's net-new speculation, net-new stablecoin-related issuance, and a loss of confidence in central bank policy around the world.

And this is, in some ways, a good backstop, Brad, for governments maybe to stop spending and maybe have a little austerity. You called austerity in the tech industry in 2021. Thanks for that. I bought a bunch of stocks, went all in on Facebook and a bunch of other ones when you were talking about this.

Also, China seems to be flocking into gold, too. I've been reading about that. So what's your take, Brad?

Yeah, listen, there's one camp that says this is all because the world's falling apart and we're all panicked about global governments' balance sheets. I think that's part of it, but I think Chamath nailed it, right? You got a total change in the new demand picture for gold, as evidenced by—we have a Bitcoin of gold, BTC, you know, Bitcoin, but we also have it in the form of Tether, right?

You're now going to allow people to have exposure to the actual underlying gold by buying a crypto asset. So you have new demand for it. You do have people concerned about the state of the world, but it's always been a speculative asset plus a store of value, right?

I've got a couple of charts here. You can throw them up if you want. One is simply looking at the correlation with the S&P 500 over the last 10 years, and the other one's looking at it with U.S. debt. It's probably chart crime. I'll go back to what Chamath said. I can make up any story I want to make: It tracks the growth in U.S. debt. It tracks the growth in risk assets in the S&P 500.

The fact of the matter is, it's a combination of all of these things. Hats off to the people who are up 50% this year on gold.

All right. Is that it, Sacks? Anything you want to add?

There is one other big factor. Maybe Chamath mentioned this, but China's central bank has been increasing its gold reserves for 11 consecutive months. They added 40,000 ounces in September alone. As of the end of September, China's gold reserves totaled 74 million ounces, valued at $283 billion.

They've been gradually substituting away from U.S. dollars and U.S. Treasuries toward gold. Some of this is geopolitical. China is our chief competitor in the world. They probably don't want to be dependent on the dollar and U.S. Treasuries, so they're diversifying away.

Also, you can't forget that this trend really started during the Biden administration with the Ukraine war, because the Biden administration used the U.S. dollar as basically a geopolitical weapon. They cut off access to SWIFT. They froze Russian assets in the banking system.

A lot of countries looked at that and said, “Oh, geez, if I'm totally reliant on the U.S. dollar complex, then I can be on the receiving end of U.S. foreign policy that I don't like.” So it was pretty natural that the BRICS countries would look at that and start to develop alternatives.

And just by the way, when the BRICS talk about this currency that they're going to develop, they're not talking about a man-on-the-street type of currency. The best guess about what they're going to do is that it'll be some sort of gold-backed certificates where they can settle up huge international trade flows using gold certificates.

So that might be part of what’s driving this as well. That’s a very long-term project by the BRICS countries. I don’t think we’re going to see anything in the next few years.

And that’s, Sacks, their way of sending a message to the US: “Hey, don’t use USD to yank our chain.” Is that your interpretation?

Well, I don’t know if they’re trying to send the message, or if they just don’t want to take that risk, right? They understand that if they’re part of the dollar complex, they can be affected by things like cutting off SWIFT, sanctions, banking sanctions, and so forth. Their assets, in theory, could be seized. Obviously, they’re going to try and reduce the possibility of that externality.

Yeah, makes total sense. All right, let’s wrap on Polymarket. My guy Shayne, friend of the pod, announced that the New York Stock Exchange parent company, ICE—Intercontinental Exchange—had invested $2 billion at a $9 billion post-money valuation, or somewhere in that range. They will now distribute Polymarket’s data to thousands of financial institutions globally, and it looks like Polymarket will be launched in the US imminently.

There’s actually a Polymarket market on whether Polymarket will be available in the US for trades. Polymarket was only founded in 2020. They were valued at $20 million, at $350 million last year before the election, and at $1.2 billion earlier this year. Founders Fund made a nice investment, and that turned out to be a pretty great trade. There’s now a 98% chance that Polymarket will go live in the US in 2025. I think that’s found money right there. I’m going to place a bet there. Chamath, your thoughts on this new category and why it’s so important for it to exist?

I talked to Jeff Sprecher the day he did the investment, just to congratulate him. One of the things he said, which I think is true, is that everything is becoming a market. You’re going to see this convergence where, eventually, you’ll have sports that look like markets, knowledge that becomes a market, and then ultimately equities and debt instruments also behave this way.

Meaning, they’ll be tokenized. They’ll be very fungible. You can just take a long or short position on almost anything. I think you’re going to see this convergence. What does it mean? I don’t know, but it puts a lot of really interesting assets under the microscope.

If you built a high-frequency trading organization, what does it mean when, all of a sudden, some crypto pool shows up and can do HFT much faster than you and basically chops the cost to zero because you don’t have infrastructure and corporate overhead? Or what is it going to mean to betting sites like FanDuel and DraftKings, where now a bunch of these trades can happen here as well?

I think what Shayne is leading the charge on is democratizing this and financializing all this infrastructure so it can run extremely liquidly via tokens, effectively on-chain. I think that’s a huge deal, broadly speaking.

Yeah, I’m loving this for the Oscars. I was just thinking, wow, if you’re a movie studio releasing a film, and these markets become big enough, you could hedge your bet on your own film. This could have profound impacts on every industry. Brad, how do you think about it as somebody who trades?

I just love the fact that we’ve now professionalized the wisdom of the crowds. We’ve known for 150 years that crowds are smarter than experts, with all these people hanging out in ivory towers controlling the conversation. Now those emperors have no clothes, right?

The reality is, I think it was this guy Francis Galton who, in 1907, got a group of 800 farmers together. They had to guess the weight of an ox, and then he had 5 experts do the same thing. Guess what? The crowd got a lot closer than the experts to the weight of the ox.

Here we are, 150 years later, and you’re going to have very big companies that are now taking it to the next level and creating liquidity. When people put money on something, the crowd gets even smarter. Kudos to you guys. Through your partnership with them, you’ve taken them to a higher level, and it’s awesome to see. It’s a great lesson for entrepreneurs.

There was a website, Intrade, from the early 2000s into the 2010s that was doing this. They had regulatory issues, but the idea was out there, and people just didn’t go all in on it, so to speak. Here we are. Shayne has crushed it.

Another amazing, respectful, fascinating, insightful episode of your podcast. Welcome home, fifth bestie, Brad Gerstner. Great to have Brad back in the mix.

It’s great to be chopping it up with you.

We have to get Brad and Friedberg on the same pod.

Oh yeah, let’s do it.

What if people prefer a science corner or a market corner?

I’ll allocate my 2 votes, I can tell you.

Wow, Friedberg taking a stray. He’s not even here. No, you just said, “Wondering-out-loud corner.” Maybe there’s an asteroid coming, or we could talk about the markets and have great charts. I don’t know. Anything’s possible.

Congratulations on all your success, Jason. Congratulations.

How much success are we talking about here?

Congratulations on behalf of the Trump administration. Congratulations.

Thank you. Well, I hope on your Trump-driven success.

Congratulations. Absolutely. Yes, and Trump 2.0.

What else can we do for you?

Well, yeah. I want to bring that up.

What other stocks can we get for you?

Actually, I do have one thing about laws. That would be a big—

What other people would you like to loop into this conversation? What other people would you like? I call balls and strikes here. It’s a very important role to have.

You noticed. You know what I’ve noticed psychologically?

Sacks is here, the Sacks man. The people who constantly say they call balls and strikes are never the ones who call balls and strikes. I’ve got my eyes on your balls, Chamath.

Don’t listen to what they say. Just look at what they do.

Exactly.

Yeah. Welcome to Texas. Come by the ranch anytime.

Welcome to Texas. Exactly. I love the great state of Texas. How’s that ranch treating?

It’s so great. I carry a pistol on my waist. It’s incredible. This place is great. You’re welcome.

To quote Jack Nicholson in A Few Good Men: “You live under the security blanket that we provide, and then you question the manner in which we provide it.”

Absolutely. That’s called checks and balances. To our amazing panel, great job. And to our civil servant, David Sacks, get back to work for the American people. You’re doing a great job.

I love you, boys. Bye-bye.