《大而美法案》、Elon与Trump、美元大幅走弱、哈佛的资金困境、Figma IPO

All right, welcome to the show everybody. Number one podcast in the world and uh man going to be really exciting come September, September 7th Days at 9th, the All-In Summit in its fourth year this week. Every year you pre-announce a bunch of people who don't show up. I got Elon to come three out of three years. I have gotten a couple of trillion dollars in market cap for your freeberg. This year is going to be better than ever. Allin.com/summit. And lots of exciting news around the tequila boys. Did you see MSN covered me sending a case of the tequila to Bezos? We open source it to the fans and they've just gone crazy with it. We've uh been crushed with orders. Apologies if we're taking a little bit of time with customer support. Uh, but you can go to tequila.allin.com. Deliveries begin in late summer, right around the time I think of the uh, summit.

Friedberg and I were in Las Vegas this past weekend. It was a blast, as always. We had 20 of our besties or so with us.

I wish I could have been there, but I was on the Snake River doing a whitewater rafting trip. You camp on the side, eat mediocre food, don't shower for 3 or 4 days, and then you have this view. It was absolutely stunning.

What were you talking about? I was getting lobbied to cancel your family trip and come to Vegas. I was like, “I don't remember inviting you, Friedberg. Do you remember?” Is that where Snake River Farms comes from—the high-quality meat?

Yes, I'm sure it is. Idaho border.

Yeah, Idaho border. So that's where Snake River Farms is. Not as good as Lone Mountain Wagyu, but these are 2 elite steak producers. Get the culotte, aka the picanha.

The Big Beautiful Bill has passed the Senate. Let's get into this docket. There was a bit of drama in the House after an all-night session and over 24 hours of deliberation. The Senate passed the Big Beautiful Bill on Tuesday. Friend of the pod J.D. Vance cast the tiebreaking vote after a 50–50 deadlock.

Three Republicans voted no, and Trump won over Senator Ron Johnson. On this very podcast, he said he would be a no unless more spending was cut. There were significant changes to the bill, and we'll get into that in a minute. It's got to pass the House again before Trump can sign it into law.

Lots of drama. Trump set the deadline for July 4, which is Friday when you're listening to this, and it will, in all likelihood, get there. According to Polymarket, there's a 96% chance the bill will pass by Thursday and a 97% chance it will pass by Friday, July 4—Independence Day, the day we publish. So it looks like it's pretty much a lock.

The biggest change in the Senate was that the 10-year AI regulation moratorium for states was removed. Ted Cruz tried to negotiate cutting it to 5 years. A suggestion I made was that 10 years was a little bit too long for folks, but that didn't work. According to the National Conference of State Legislatures, over 1,000 bills related to AI have been filed by state lawmakers in 2025, so expect a lot of state-by-state AI regulation.

I think this is a good place to stop. We don't have Sacks here to talk about AI regulation, but he talked about it the last time. Friedberg, your thoughts on states? We talked about states' rights here in relation to abortion, guns, gun regulations, cannabis, and many different debates over who should get to decide for the country. Where do you stand on this one? Should states have a voice in how AI is deployed within their borders, or should the federal government take this, and if so, for how many years? That seemed to be a sticking point.

Look, I'm a big believer in the construct of our federated republic in the United States, where states can operate with as much discretion as they choose to, with the laws that they pass and how they intend to govern. There are, however, things that affect more than the state. We have interstate commerce, international commerce, the open internet, and a lot of other systems at play here that extend beyond the boundaries of a state.

The way that AI tools, AI systems, AI services, and AI-related jobs are being deployed and activated, I do think it's critical that we treat AI regulation at a national level, at a federal level. To date, there have been, I believe, 70 state laws or statutes passed thus far in the United States, with, I think, over 1,000 having been proposed to date, if the numbers are right.

Having a patchwork of regulations on, for example, model development, or telling software companies what software they can deploy, would make it practically impossible for internet service providers like Google or OpenAI to service customers across state boundaries in a way that's actually going to meet the needs of the customer. This is a huge detriment to consumers and a huge detriment to the job market if we end up creating a patchwork of regulations on AI this early.

One of the things that I would point out is that much of the early legislation, like what we saw in California, was naive. If you read the legislation that was not signed by Gavin Newsom in California, there was a definition of model parameters as a boundary condition for whether or not something would go through a regulatory review process.

That was such a naive assertion this early in the phase of AI technology development. Think about the implication: everything is LLMs, but there are other models, like vision-action models, that are going to be used in robotics—models that might be used in predicting and driving physical systems, like self-driving cars, for example.

The way that regulation gets written will often be done in a naive context, in a limited fashion, by regulators that don't fully understand the breadth of the technology that's coming our way. I do think that it needs to be leveled up and allowed to operate at a federal level.

I think it is absolutely critical—not just for technology companies, not just for the industry, but for the opportunity for job creation, for improving the economy, for driving GDP and productivity growth—that we get something passed that forces AI regulation to be done at a federal level. Even though I generally agree that states should be able to self-govern, this just has too much of an interstate and global interaction.

That was the biggest disappointment for me in the negotiation that's gone on with this bill. Of all the things that I thought were going to be passed, and that I thought were maybe not ideal, getting this AI federal preemption in that bill was probably the best thing—the top of the list for me—in terms of what was critically important for the U.S. economy.

Yeah, Chamath, your thoughts here? Obviously, abortion is an interesting one to look at because that went to the states in furtherance of Friedberg's position here. You have many different rules and regulations between Texas, Oklahoma, and whatever it is. Everybody's got a different view. It's chaotic, and we don't have a singular point of view on it.

Therefore, women in one state have very different rules than in another state. So where do you fall on this one? Federal legislation, or is it just going to take more time to figure this out? I mean, that's another angle here.

Yeah, this is nothing like abortion. Abortion is an individual issue between a woman, her doctor, her partner, and her God. A state being able to have rules and then allowing that woman to choose which state she can go to is quite reasonable, because I think that it acknowledges that there's no one-size-fits-all. So that is clearly a state issue.

But the reality is that AI is of national security importance. It is the tip of the spear of our ability to have technological supremacy. That is how we have economic supremacy. Not governing it at the federal level, I think, is a mistake.

It makes as much sense for states to legislate AI as it would make sense for states to have competitive armies. It doesn't make sense. There are certain things that are just so fundamentally primary to American supremacy, safety, and security that they need to be regulated at the federal level, and I think AI needs to be at the top of that list.

So I was disappointed that we weren't able to see the forest from the trees here. Now, I'm sensitive to part of the reasons why many of the states have very specific issues around child safety that they mentioned in Tennessee. There was a lot about copyright ownership related to musicians and their content. I get all of that.

Those are legitimate things that need to get ironed out, but I think it could have been done in a federal framework. Not having it, I think the risk is that if you have 50 different sets of regulations—and it won't even just be 50, as Friedberg said—it's going to be all kinds of pet organizations that pop up in all of these states.

What happens? I think what it does is it slows down startups and smaller companies, which won't have the economic heft to fight these regulations, work with them, or figure them out. It will advantage a handful of incumbents.

And the problem with that is that our incumbents aren't yet mature, and this industry is still developing. So I think this is one of the unfortunate parts of what is otherwise a pretty reasonable bill.

Well, you can look at gambling, I guess, and cannabis legislation. These are things that Americans largely want to do, and they are fighting their way state by state. At some point, we may have a federal regulation about online gambling, right? Right now, it's a patchwork. Same thing with cannabis and psychedelics—all a patchwork of different states.

Why do you lump together things that are about individual rights with AI, which I don't think is about that? People might look at the rights of rights holders, but these AI things are going to hit people on an individual level. So I think when you bring up this topic, there are people who feel this is an individual sovereignty issue, right? But there are other issues of individual sovereignty, like the ability to take drugs, and there are still federal mandates that come through the FDA. You're not allowed to short-circuit or circumvent that framework at a state-by-state level.

Yeah, that actually is at the core of the cannabis debate: would they enforce the federal laws on cannabis? They haven't, right? So that's what I'm talking about more as it relates to AI.

Sure. Because AI means everything, and I think that's what makes this issue very difficult. AI is going to affect everything, right? So probably the people who want to slow this down and have the states do it are saying, "Let the states have their say on these granular issues." I'm not saying this is my position, but I think that's what I'm reading into their position: This is going to affect different states in different ways, so let them figure it out. Take some time, and then we'll build toward a framework nationally, which I think is what happened with the internet and Section 230 as well.

Although I don't know the entire history of it, we did come up with an idea that the federal government makes it a common carrier, so anybody can post anything on a website, and then you can request to take it down. That was federal legislation. Those things could have been handled on an individual basis, and then what would have happened with the internet? You would have had to have AOL or GeoCities have a different patchwork of regulations, like you're saying.

So I'm guessing the states probably do want to have a say on self-driving, specifically, and I don't know if they're wrong. What do you think about self-driving? Should it be federal? I mean, that's an interesting lens to look at it through. I don't know the details well enough.

It seems to me that self-driving touches both the states and the federal government.

Sure. Federal highways and local roads. Yeah, so just from first principles, there are self-driving cars now in Palo Alto going up to the city from Waymo. You've got Austin and three different providers. Should the city or the state have a say in that? It kind of feels like they should.

Yeah, sure. I mean, I don't think a city has any idea how to regulate a model.

No, that's where it's obvious. How would they ever even know? Another piece that's in flux and has been changing is energy policy: solar, EV credits, and maybe even disincentives—the $7,500 credit for buying cars, as well as carbon credits and a fee of $250 for everybody who owns an EV every year.

Friedberg, maybe you could talk a little bit about this part of the debate, because we've been spending and hitting incredible goals as a country in getting more renewables and more energy online. But that comes at a cost, right? That is a big part of the debate here. It obviously impacts Elon, and we'll get to him in a moment, but that impacts Elon and Trump's relationship in a major way.

The old-school motivation and incentive for what people would call green energy or green tech was to decarbonize electricity production and develop and deploy at scale these new technologies at the time, like wind and solar, as an alternative to carbon-based technologies that put carbon dioxide in the atmosphere, like oil, coal, and gas.

So the IRA, the Inflation Reduction Act, created tremendous government incentives for building solar and wind farms to create electricity in the United States. This proposal that's been going back and forth in this bill, and some of the debate that's triggered many folks online to denounce the bill, takes away those incentives, takes away those credits, and takes away the EV tax credit starting in September, as an example.

Why can't they do these in digestible chunks? It should be no more than 300 pages per bill, no more than X number of issues, and then break. Good conversation to have another time, but we should talk to people in Congress.

But there's a very specific reason why: the logjam that happens with the filibuster. It makes it impossible to get things done in that traditional context. So what's happened in the last couple of years is you end up with these massive, mega-bills where they jam everything into one bill that they can pass with 50 votes, and then with 51 votes they make these last-minute changes. Then it all gets dark. There's no rigorous debate and no transparency into what's going on and why.

The second argument, away from the green stuff, is the more modern argument, which I've made, and I know Chamath agrees with me on this: We have to increase electricity production in this country. The current plan has been to move from 1 to 2 terawatts by 2040. Meanwhile, China's moving from 3 to 8. China is adding an entire United States' worth of electricity production capacity every 18 months right now. They are so far ahead of us.

Ultimately, electricity production is what drops the price for things, increases jobs, increases GDP, and increases your ability to make stuff. The more electricity production that comes online, the more energy-sufficient we will be from an AI point of view.

Now, the secretary of energy, Secretary Wright, gave an interview, and he said we have more than enough capacity with gas, oil, coal, and nuclear. His argument at this point is that we don't need to subsidize solar and wind, where the government pays for these programs and private-equity investors can make money on them. This is a big part of the argument he's making. I'm not making it. But he's saying this will drive traditional energy investments in traditional systems that we can scale up quickly to meet our energy-production goals.

I will say, from my point of view, I'm not a huge fan of being dependent on government-subsidized energy at all. If the government is having to play a role in funding stuff, there's something really questionable in terms of our sustainability on that energy-production source over time. Will it actually be able to make more of it? What you want is not just to make a bunch of energy in the next 6 or 12 years. You want to make sure that you've got an engine for creating new energy production on a continuous basis, so we climb nonlinearly up the energy-production curve. That's what we need to do.

I hope that one of the key outputs of this—which a lot of people are really unhappy about, the loss of demand for solar and wind because of the loss of the government programs—is that it creates a natural market force for nuclear. I hope that we actually see better proliferation of nuclear, which Secretary Wright has advocated, and President Trump signed these executive orders a couple of weeks ago to reduce the regulatory burden on nuclear and increase the ability for nuclear to proliferate much more quickly than has been the case historically.

While a lot of people are saying, "Hey, this is going to take away all this energy-production capacity that's coming online in the next few years, or is supposed to," the counterargument is that we're going to create a natural market force because the energy demand will still be there. Someone is going to show up and say, "Hey, I want to make electricity because there's so much demand for it." Maybe they'll say, "Let's try this new nuclear deregulation system and see if we can get it to work." I don't know. The jury is absolutely out. We have not seen this energy proliferation begin in this country, but I do think one of the potential benefits of removing those clean-energy tax credits is that we actually see natural market forces drive demand for a more naturally sustainable, scalable energy-production source. Your thoughts?

I announced this a few weeks ago, but I dipped my toe into real estate. I've never done real estate before, but just outside of Phoenix, we're building a 1-gigawatt data center. That's a $25 billion total investment cycle, of which we'll be into it for $2 billion or $3 billion of equity when it's probably all said and done.

What I can tell you is that the reason why you can underwrite that deal, or why I was able to underwrite the deal, is that it's located downstream of a nuclear reactor, so there's effectively infinite energy that we can tap. But we still need gas and a bunch of other things.

The problem is that if we put in an order right now, we can't get a gas turbine viable and turned on until 2030. It's not a technology issue. It's purely a supply-chain issue. So I think what our energy policy needs to make sure we contemplate is that many of the things that we are debating are no longer issues of production, but issues of supply, transmission, and distribution.

As long as we can make sure that we allow energy to be made wherever and whenever possible and then stored however possible, we'll be on the front foot. The big risk is that if we don't have that ability, and then all of a sudden we have these incredible achievements in technology and robotics but don't have the electricity to power them.

That's the big risk. As long as we can fix that, I still worry that with nuclear, the issue won't be the federal regulations. I think it's good that the president cleaned it up. The real problem is going to be the local and state regulators and how quickly they're willing to turn these on.

The reality is that these are 10-year projects. So even if you say, "Go from today," the earliest these things can really be turned on is 2032 or 2033. That's far too late, and that's a lot of risk too, Chamath, because what if you get blocked at the last minute?

This is where solar and other projects—we're big as well. I'll tell you a story. In 2020, right before the pandemic, I don't even know if I should say the name of the company, but I almost went after a coal company to buy it. I had been obsessed with this company for a year. It was cheap. It had fallen on hard times. It was like $1.2 billion, $1.3 billion.

My team went up in arms. They were like, "Oh, my God, it's so dirty. You can't own a coal company." It was such a huge miss on my part. I should have just bought the bloody thing. Then we went into COVID, everything went upside down, and, lo and behold, that company has quintupled since then.

So I'm a fan of all forms of energy production. I think the marginal cost of energy has to go to zero, which means that any single way you can get your hands on electricity production is a winning trade over the next 20 years in the United States.

This is what I was trying to wrap my head around when people were saying, "Oh, we shouldn't buy all these solar panels out of China because China is making them so cheaply, and our trade imbalance, et cetera." I was like, "Well, wait a second. What's more important—the trade imbalance and giving China some money for solar panels, or getting more solar panels to Arizona, to Utah?"

Look, let's take the subsidies off the table. Ultimately, you don't want any of these markets subsidized. The reality is, it is faster and more efficient to put solar panels up and start generating electricity. It's just 17 months from start to finish. Seventeen months.

That's probably, in my opinion, the single most valuable underwriting feature of solar: you put a dollar in, and you can start generating revenue in 17 months from that dollar going in. That's very different from putting a dollar in and not seeing it back for a decade-plus. That's scary.

You need to generate a rate of return that justifies a 10- or 15-year investment cycle. That's hard to do because there's a lot of volatility in the world, and things can go down as well as up. I don't know. My perspective is, get rid of all the subsidies at this point. Create a clean slate, but don't do anything to hinder the production of electricity because we need literally as much of it as we can get our hands on.

Then we need to find a way of storing it in a safe way, and then we're off to the races. The battery technology, Friedberg, just keeps making great incremental progress in terms of cost going down—what, 15%? Jason, I have a question for you. How many days a year do you think California is in an energy deficit? How many days a year is California in an energy deficit? It doesn't have enough energy.

It would be during the summer months, when electricity prices go crazy and they have to find all kinds of ways of generating massive amounts of baseload. How many? Like a third of the time—100?

Five. That's it.

Wow. So we've really made progress. In California, our grid is, on average, between 45% and 50% utilized. This is for PG&E. I don't know what it's like for SoCal Edison.

Yeah, down south.

The point I'm trying to make is that, in normal market conditions for residential customers, we have tremendous amounts of power, and for most needs, we have tremendous supply. So things can be pretty good. The problem is all of the stuff downstream from the making of it. If we don't clean that up, we're going to create these artificial constraints that will come back to bite us when we really need the power to do something very exciting.

All right, let's move on to Elon and Trump's relationship. Everybody wants to hear our take on what's going on between Elon and Trump in relation to this bill. We took a pass on it last time, but we'll take a swing at it now. Obviously, there's been a bit of back-and-forth between 2 of our friends, both friends of the show, Elon and Trump.

Are you friends with Trump, Jason?

Well, I'm speaking on behalf of the show. I have never met Trump in person, but I did interview him on behalf of me and Sacks.

But on behalf of the show, listen, you guys are all Team Trump and Sacks.

Yeah. Well, you're Team Trump, too.

Elon has come out hard against this bill. He tweeted, "It is obvious with the insane spending of this bill, which increases the debt ceiling by a record $5 trillion, that we live in a one-party country, the Porky Pig Party. Time for a new political party that actually cares about the people."

This escalated a bit, but this isn't as bad as the first time around. When asked if he would deport Elon, Trump said, "I don't know. We'll have to take a look," and said he might stick DOGE on Elon since he gets a lot of subsidies.

Friedberg, we skipped talking about it because it was a bit chaotic last time, but this time I think maybe we'll chime in a bit. Things aren't as hot right now, and the bill is going to get through. So let's take a swing at it. What are your thoughts, Friedberg, on the kerfuffle, the donnybrook, the brouhaha between Elon and President Trump?

There are a lot of people making accurate declarations that federal spending needs to be reduced and the deficit needs to be shrunk. We are in a debt death spiral, and they are absolutely correct. So I don't think Elon is off the reservation when he makes those comments. He's talked about this continuously, and he dedicated months of his life to operating DOGE and trying to bring to light some of the extraordinary operating inefficiencies in the federal government that should be addressed.

I think we'll see what happens. The jury is still out. For those actions to become permanent, I believe they need to be mandated in an appropriations bill. The White House has publicly declared that they are going to move forward with an appropriations bill to try to cement some of the DOGE actions. So I'm hopeful, but I think Elon is right with respect to the spending.

I will provide the alternative view from the White House that I have heard publicly stated, which is that this is not the bill to do that because it mostly focuses on these mandatory spending programs. They are addressing cost savings to some degree in these mandatory spending programs while keeping the tax rates where they are or reducing tax rates in some cases, with the expectation that this will stimulate GDP growth.

The White House view is also a view that you could look at and say, economically, I could see a path here that does make sense. You're reducing spending only on mandatory programs. You're going to come back with an appropriations bill to address discretionary spending. Then the final action that the White House might take, which we've heard about separately, is impoundment.

At the end of the fiscal year, they may come back and say, "All the money that we saved by not spending it, we can actually recover through impoundment." The CBO has not accounted for tariff revenue. Just looking at the recent deal done with Vietnam, Vietnam has about $130 billion a year of imports to the United States. They announced a deal 2 days ago with a 20% tariff on Vietnam, which, if Vietnam's import volume does not increase, is about $26 billion a year of incremental revenue for the federal government.

There's an argument that Bessent and Lutnick and others are making: "You guys have failed to recognize that we've got a few other things up our sleeve that we're going to do. We're going to do impoundment. We're going to do this appropriations bill. We've got more revenue coming in. Net-net, we will get the deficit down to where we need to be."

In fact, Scott Bessent was on TV saying over and over, "We're going to get the deficit below 3% of GDP," which is the key target here. But the immediate reaction that Elon is having to this bill is the same one I've had, the same one Senator Johnson had, and the same one many others have had: "What the f— are we doing?" We are in a fiscal emergency in this country, and we're not addressing it.

I think both points of view can be valid, and both sides can have a good argument for why the bill should be passed and why the bill shouldn't be passed. The White House has declared that this is the only way they're going to get some of the programs done that they believe they need to get done for national security, like the ICE border stuff. These are things that they believe they need to get done.

But a lot of folks are looking at the numbers and saying this just isn't enough, and you're now increasing the deficit by cutting taxes. One of the things the CBO does not do, though, is maintain a strong model, and there's a ton of economic debate on this point: how tax cuts stimulate GDP growth, job creation, and income growth for people with jobs.

The one argument is that when you cut taxes, more dollars flow into the economy, more jobs are created, more businesses are created, GDP grows, and income for other people grows. The alternative argument is, "It's a tax cut for the rich. What are you doing? They're going to put that money in their pocket."

It's only going to benefit themselves. So that's, I think, a key crux in the argument that you'll never get to a resolution on. One side will use that argument, and the other side will use the other argument.

So, with respect to Elon and Trump, I will say one thing very importantly: I don't think MAGA can exist successfully without the tech alignment. I don't think tech can exist without MAGA because of the government alignment and the importance of the government allowing these new technologies, like AI, to come to market and proliferate. I don't think these two can exist in isolation and in conflict with one another.

Elon is the de facto king of tech. He is the person saying, “Look, if I'm in conflict with Trump, tech is in conflict with Trump,” or at least that is the perception on the MAGA side. I think that is very risky for both sides, to allow a conflict to endure. I do think that both sides have heads that are going to be cooler and will prevail here.

I do think that these two are going to recognize the importance of being codependent, if you will, in being able to progress their respective agendas. Chamath, your thoughts on the kerfuffle? It seems to be winding down and maybe reaching some sort of endgame. It feels like we're in some sort of an endgame here. And if there is one, what is it?

I think it's just important to recognize that on the one side you have the most powerful person in the world, and on the other side you have the most important, powerful entrepreneur and richest man in the world. When you have people that are that accomplished, it's not as if you're going to get along 100% of the time.

I think a lot of the breathlessness around all of this stuff is overblown. I think the reality is that when push comes to shove, I think they agree on more things than they probably disagree on. And I think when everybody realizes that the alternative is essentially some insane form of socialism and redistribution, I think the alliance will hold and they'll find some common ground. Jason, what do you think?

Well, I think you're right. These 2 individuals are used to speaking their minds, and they're both really good at social media. There's no better media cycle than covering the 2 of them battling it out and trading barbs.

But I think what Elon did this cycle was really interesting. He started a preference cascade for Trump, and that preference cascade—I think it played a large role, if not the role, in getting Trump elected. People can debate that. I don't know that you can perfectly know what that $250 million and all that effort he put in did in terms of Trump's chances. Trump probably would have won anyway versus Kamala, but maybe not.

Putting that aside, the platform Elon has refined during this political cycle is one that resonates, and I think it's a really good idea for him to, if he's going to be involved in politics, pull that string and crisply define it. I was thinking about it over the last couple of days, and I think it hits into 4 basic groups.

Balanced budget and government efficiency—that's kind of one, fiscal responsibility; sustainable energy, which obviously he's been passionate about, and he's the leader in solar, batteries, and EVs; manufacturing in the United States, which he also is the leader in; and pronatalism, and just the population crisis. He really cares about those 4 issues.

So what I think he should do is take that America PAC and, instead of making it just pro-MAGA, clearly define what it is that he believes: a small set of issues. Then he should go and back the people who are running for Congress and the Senate and just say, “Hey, here is what I would like you to be in favor of.”

Do what Grover Norquist did with his no-new-taxes pledge. Do some sort of pledge like that. I don't know if it needs to be a new party, but just really crisply define what matters to him. If he's going to be involved in politics, then get people to agree on it.

If he wants to give them donations, as is his right, as is his PAC's right, to raise money, he could represent, I think, a very world-positive view: sustainable energy, incredible execution, and a really efficient government. He should stand for technical excellence—technical excellence, excellence at large—and America. Just go for it.

It doesn't have to be personal against Trump. One of the big problems with Trump is he has a bunch of sycophants around him, and the more you appease him and just blindly follow him, I think the perception is that the closer you get to him. I don't know that that's true. I think he likes debating stuff.

So I think he should embrace the people who debate it with him, and then those people should just crisply say, “Here's what I stand for, and I'm putting my money behind it. I hope we're in alignment, but I'm going to stay in my lane and prioritize what matters to me.” Elon's priorities are exceptionally sharp and well-defined, and he should pursue them.

Yeah. The jury's out on where this ends up and the broader picture. You guys will get annoyed because I've talked about Ray Dalio so much, but I think he's done an amazing job in the last 5 years basically explaining everything we've seen, from global conflict to the internal conflict, the rise of socialism in the U.S., and the relationship to the debt and deficit cycle.

He just posted on Twitter yesterday that he went to D.C. to discuss the budget deficit with senior people on both sides of the aisle. He said, “It's clear to me that we are unlikely to change the debt trajectory we're on and avoid the painful consequences.” And he talked a lot about this concept of absolutist politics.

This is the same reaction I've had every time I've gone to D.C. and we've met with members of Congress. If you guys remember, from the beginning I said DOGE isn't going to be it. As much as Elon can identify and resolve to a better way of operating the federal government's agencies, you cannot change the spending without a change in statute from Congress. You have to get Congress to act.

Every time I met with members of Congress, their incentive is to keep the money flowing to their districts. That's what they focus on. Every year, their districts want more in different contexts, in different forms, through different programs. And that's what their job is. Their job is to go to D.C. to represent their state's interests or their local district's interests and say, “We need to make sure that we're taken care of as the money flows.” As a result, everything keeps getting bigger and bigger and spiraling away.

At the end of the day, the way we economically save America, if we even have a shot without money printing—which a lot of people like Balaji and Dalio and others are now indicating is what's going to happen, meaning we're going to end up inflating away or printing away all of the debt that we've taken on—is to get to a 3-3-3, which Bessent has highlighted is also his goal: 3% federal deficit to GDP, 3% GDP growth, and 3% inflation.

Just to give you a sense of where those numbers sit today, the current estimate is a 6% deficit-to-GDP and 2.4% inflation. So we've actually got a little bit of room to run on inflation, and GDP growth this year is 1.4%.

Will the tax cuts in this bill increase GDP growth? Will AI increase GDP growth? The jury is out. Will the tariff revenue reduce the deficit more than the CBO is estimating? And will increased GDP growth reduce the deficit? The jury is still out.

What will happen with inflation? If the Fed cuts rates, you're going to see inflation climb a little bit more. Maybe the tariffs' effect on inflation will still come through. Some folks have said that hasn't hit yet, but it will come later this year. Will we exceed 3%? Inflation TBD.

So the jury is still very much out on whether these actions that are being taken, which Bessent has declared are going to get us to that 3-3-3 number, will do so. A lot of folks who are looking at this with a cold stare, with no political influence, with no political intention, not representing a party, and not representing an administration, like Ray Dalio, are saying, “What? There is no way we're going to get there.”

Balaji, I think, pointed out in his tweet that our debt is not just the federal debt, but there's other debt that we're not even accounting for: consumer debt. I've said this in the past. There's actually corporate debt—business debt—and consumer debt. There's also state and local debt.

The one key number that we never talk about is the unaccounted-for liabilities in public pension funds, which are on the order of trillions of dollars more. So when you add all of this up, someone's going to pay the bills on all that debt, or we're going to have to inflate away that debt by printing money.

At some point, the train has left the station. A lot of people are saying, “We're done. There's no way out of this.” So the Elon-Trump battle is an interesting side kerfuffle, but it's not really the big story here. The big story is the train has left the station.

If we're looking at this as the first year that tech really got actively involved in politics, obviously Peter Thiel has been at this for a little bit longer in his support of JD Vance, the vice president. I'm looking at this as maybe this is year 1 that Gen X is truly engaged in making a difference in Washington and setting an agenda.

That agenda has 25 years ahead of it if everybody remains this engaged, whether it's people like Sacks or Thiel or Elon or countless other people. What is JD Vance's position on this when he runs for president in 4 years? What will these other individuals—Dean Phillips, who we've had on the program—do if he decides to run?

I think we're maybe 5% into the impact this could be having. And already, Jason and Friedberg, I always give you a lot of credit for this.

Two years ago, you said you started bringing this up incessantly on the program.

I started doing this during COVID.

Yeah, it was 3 years ago. You're admitting you've been annoying for 3 or 4 years.

I think, actually, it's the sand that might make the pearl in the oyster. We need to address this. It's become a top issue of our time.

That's actually success. The fact that this issue is now the issue of our time—our budget, our fiscal responsibility, austerity—that's actually a really good thing. Maybe it doesn't get manifested in this bill, but maybe it will get manifested in JD Vance's presidential run, or Dean Phillips's. The both of them will be discussing it in 3.5 years, or 3 years, when they're on the presidential trail. So that could actually be the early sign of success for this. I'm an optimist.

Good luck. Cap that. It's over.

We sit here and park on this topic dozens of times. We keep talking about, "We need to do this. We need to do that." But at the end of the day, when the bills get passed, when Congress takes action, you get to see where their heads are really at.

I do think that this bill has indicated that the administration has a set of incentives: They want to get the actions done that they promised they would on the campaign trail, and Congress has a lot of incentives to keep the money flowing. I think we've seen that in this negotiation to get this bill done.

I'm not faulting anyone for it. Congress is looking out for the interests of the people they represent, and the White House is looking out for the people who voted for them. They said, "This is what we're going to do. We have a mandate," and now they're getting it done. That's just the way it is.

But this is the key part of the whole storyline: At some point, once the spending levels get too high and the country, the individuals, and the businesses become too dependent on that spending—which is effectively what happens at the end of every empire—you can't back out. You can't stop spending, and everyone just votes themselves the dollars.

It's a scary moment. I've said it before: I do think that GDP growth is the one path that's left to resolve this. I don't think we're going to just cut spending. We need to be really thoughtful about making sure we don't hamper GDP growth, particularly as it relates to AI, which is going to unlock a lot of new industry, a lot of new growth, and a lot of new opportunity for jobs. This is why I worry a lot about this patchwork of regulation in the states.

You only need 5 more people like Rand Paul and Thom Tillis, right? If you get a couple more of those, Chamath, this could be a completely different discussion.

If you want to pivot to the next topic, one way to look at this is how the dollar is trading. The U.S. dollar is now down 11% this year, and that's against every single major currency. The dollar was down over 10% through the first half of 2025. This is the worst start in over 50 years.

Take a look at this chart and you can see it. It's kind of shocking. Keep in mind, the dollar jumped 7% after Trump was elected. It peaked in mid-January, so if you take the pre-election dollar index number, 103, it's down 6% since then. That's not record-breaking, but it's significant.

Obviously, economists are saying tariffs and global trade are a big piece of this. As we just discussed and will keep discussing, the U.S. debt load is $37 trillion. It's more expensive for Americans to travel abroad, and it's less attractive, maybe, to invest in the U.S.

Let me just follow up on the point I just made. This is where the chickens come home to roost. This is where the rubber meets the road. You start to see the inflationary effects of the spending and the spiraling debt: Things get more expensive, and your earning power doesn't increase commensurate with the higher expenses.

The U.S. today imports $4 trillion to $5 trillion a year. That's $4 trillion to $5 trillion that U.S. consumers and businesses are buying from abroad and importing into the U.S. to use those products and services. The cost of all that just went up by 11% as the dollar declined in value against the average of all these currencies.

That's outside of increases in prices that may arise because of the tariff effect, where folks may say, "Hey, let's charge more for tariffs." Whether it's the cost of tariffs or the cost of the debt, the dollars that you have to spend now just went up by 11% to buy the same thing.

If that compounds and continues, and your earnings and assets are not growing commensurate with that, that's where dollar devaluation and asset devaluation happen. That's where, again, we open up the door to a thing like socialism, where people are like, "It's now twice as expensive to buy my groceries. It's twice as expensive to pay my rent, and I'm not making any more money. I need an alternative. We have to get together and get the government to make everything free."

That's why I'm so convinced that there's a rise in socialism in this country. In this sort of inflationary environment like we're seeing so far this year, you don't see it in the dollar inflation numbers. You see it in the dollar currency numbers. You're going to say, "I need an alternative."

The paradox to that is that free stuff is not free. It means increased spending, which means you're stoking the flame even more. So, exactly. Any thoughts here on the dollar?

I think the dollar has devalued 50% in the last 35 or 40 years. I think it's somewhat useful to look at any single couple of months in time, but this has been a one-way trade for a very long time. It's probably important to understand why that is.

I think it generally has to do with the fact that the United States finances a lot of growth, and that has been the right decision. Unless you see a complete collapse in the currency, I suspect that this decay continues to happen.

So the question is, is it a bad thing? The answer is, it depends. If asset prices increase faster than the dollar devalues, you're still ahead. You may not be ahead as much, but you're still ahead.

If you look at asset prices in the United States relative to asset prices anywhere else in the world, it is the flight to quality, which is to say that it is the thing that everybody wants to own. You see that in the equity markets, you see it in real estate, and you see it in hard assets.

I don't know. I think that until we run surpluses or completely eliminate the debt, there will always be a reason to be somewhat short the dollar. But the reality is that a lot of people still want to own these assets more than they want to own any other asset. Those assets are dollar-denominated.

As long as that continues to hold in the push and pull, it'll be a slow bleed, but it's probably manageable. That's just sort of the mathematical trend of it all. Unless there's some cataclysmic collapse in asset prices, I think this is just a thing that you have to deal with. It's sort of like the carry of it all. There are all kinds of other trades where you pay a carry, and that's okay.

Is there a relationship you can explain to the audience between the stock market ripping and the dollar devaluing?

I think the reality is that, if you think about a country that all of a sudden has to pay a tariff, let's take the Vietnamese example that Friedberg said. Let's just say that they had to prepay 1 year of it, just to make the math simple so you can understand. They have to come up with $23 billion, I think, is what Friedberg said.

$26 billion on $130 billion.

How do you do that? The first thing that you're going to do is sell dollar-denominated assets that you already own, generate those U.S. dollars, and send them to the United States Treasury. You may be selling bonds, so you would think, "Okay, well, that's not really good for asset prices."

But then you quickly realize that all of that is far outweighed by the fact that all the rest of the stuff that you own, whether it's gold-denominated or local-currency-denominated, you actually want to go and buy these dollars because you want assets that are safe in turbulence and volatility.

Would you rather be long the Vietnamese dong or the Vietnamese equity markets? To a degree, yes. But if you have obligations that the Vietnamese government needs to fund, their central bank is probably deciding that they need to be long U.S. bonds and fixed income. Other people who hold assets in that country are probably going to be, on a weighted basis, adding exposure to the United States while all of this stuff is happening.

Why? Because you're seeing the balance sheet of America burgeoning and growing: $23 billion from Vietnam, tens of billions over here, tens of billions over there. It all adds up.

Again, I think that this is a very complicated, multivariate problem, but the net takeaway is that dollar devaluation is not something that's new. It is a phenomenon that has existed through market cycles for 50-plus years. It's a decay that has happened and will continue to happen.

The way that I think about it is that there's a drag, but can the drag be overcome by the increase in asset values of the hard assets that are dollar-denominated? The answer is yes, and meaningfully so. As long as that's the case, I think you're going to continue to have a bid for equities.

If all of a sudden the MAG7 decided to delist, stop being American companies, and show up in the CAC 40 in France, we would be in big trouble. But I do not think that is going to happen. As long as there is American ingenuity and American supremacy, which goes back to the other point, we cannot kill these golden geese, nor should we kill the emerging and growing golden goose that is AI.

As long as those things remain the same, there will be a constant bid for American assets. That will keep the enterprise of America going for far longer than most people would guess.

Up to a point, but pull up this chart that I mentioned.

Yeah, I think that this is why Buffett speaks about this in these extremely long arcs: you are always ultimately going to be on the wrong side of the trade if you bet against the United States. It is probably important to ask what the boundary condition would be.

I do not think the boundary condition is the dollar. I actually do not think the boundary condition is the debt. I think the boundary condition is if something were to happen to the quality of the human capital inside the United States and its inability to innovate. Then we would probably be in trouble.

But even that story would take 50 to 100 years to play out. I just think that, for most of our lifetimes, this is a safe trade because it is a winning trade.

Well, it is winning if you can put up with the currency dropping 11% in 6 months. If you pull up this chart—and I get it, Chamath, I agree—but businesses are booming in India, businesses are booming in China, and businesses are booming in other countries in the way that businesses used to boom strictly in the United States.

There are other markets that seem to be having their day. You can see this number, which I think is really striking: in the last 10 years, U.S. Treasuries held by foreign holders have declined from 34%. This is a good thing in one context, but my point is that foreign countries, foreign businesses, and foreign investors are not holding U.S. Treasuries as much.

There is nothing bad about this chart. This chart shows that foreign governments and central banks have less and less influence on the direction of American fiscal and monetary policy.

Better. That is better. So who is going to buy our debt?

It turns out that when you are the largest economy and you are growing, there are a lot of internal people who will do it. That is a key point here.

That means the debt is going to get more expensive if there is a smaller market.

That is not necessarily true either. One thing that is true is that we are investing a lot of money, and there is a lot of dry powder. The amount of money being invested into data centers and AI, and the amount of cash that is moving into investment specifically within our borders, has to be accretive, right?

Yeah. The amount of cash in money-market funds probably exceeds the sum total of all of the equity markets around the world.

That is insane. Now people are talking about building $10 billion, $20 billion, and $30 billion worth of data centers and nuclear power plants again. There is something going on here with American exceptionalism.

Speaking of American exceptionalism, Harvard is still battling it out with Trump, and its $50 billion endowment is being questioned. It actually relates in many ways to what we are seeing in private tech companies and venture capital. It is going to take a somewhat circuitous route to get there.

Since Trump was inaugurated, Harvard borrowed $1.2 billion due to uncertainty around its federal funding. Remember, the Trump administration canceled more than $2 billion worth of research grants to Harvard. Earlier this week, the administration formally accused Harvard of tolerating antisemitism on campus.

The White House said it would file a civil-rights lawsuit through the DOJ as soon as possible unless Harvard came into compliance and made a deal with Trump. Here is what they are looking for: the Trump White House wants Harvard to cancel its DEI initiatives, establish third-party oversight of admissions, and take mandatory actions to combat antisemitism.

Harvard has declined to make a deal so far.

You might remember Representative Elise Stefanik. She asked the FCC 2 weeks ago to investigate Harvard’s financial disclosures. She got into the disclosures when Harvard set up that $1 billion-plus line of credit and loan. They did a bond deal and issued about $750 million of bonds.

That $750 million bond offering on April 9 was sent out, and 6 days later Harvard sent out a supplemental disclosure with more information about the White House civil-rights investigation, which paints a dire financial picture. Harvard is cooked, and I think this is really good for America.

Explain.

It turns out—and a lot of people have posted about this on X—that there continue to be rampant Title VI violations with respect to admissions. I think Cremieux published something you can probably find showing that Columbia was continuing to discriminate against Asian students. Harvard’s original case was about discrimination against Asian students.

There was also a lot of woke activity at other Ivy League schools, like UPenn, and there is all this rampant antisemitism. It all needs to be fixed. If a deal gets done, Harvard will have to capitulate.

I think President Trump holds all the leverage and all the cards, and there is nothing mathematically that Harvard can do. They can stall for probably another year and a half, but at some point they will not have the budget to sustain themselves, and they are going to get into a huge world of hurt.

What they will have to do in order to finance their budget in probably 18 months is actively sell their private-equity portfolio. By the way, from 2019 to this year, that allocation almost doubled, from 20% to 40%. That is an insane asset allocation—frankly, an asset misallocation—at the top of the market into the most illiquid asset class.

When people sniff this out, what they are going to do is look at Harvard’s private-equity portfolio. Harvard was recently able to sell $1 billion of private-equity investments in managers it no longer wanted to support, at a 7% discount.

There is no smart money on the street that is going to look at any private-equity portfolio from Harvard without asking for a 20%, 25%, 30%, 35%, or 40% discount, because Harvard’s back will be totally against the wall. If you do not ask for that, you are just a bad businessman.

If you put all of these things together, Harvard is going to need to reestablish federal funding. In order to do that, I think the president has been very, very clear. For whatever reason, Harvard has refused to address these issues. But there is just something fundamentally broken in the Ivy League.

The Wall Street Journal reported on Wednesday that Harvard would face a $1 billion budget shortfall every year if Trump followed through on his funding cuts and tax hikes. He was also saber-rattling that he would get rid of Harvard’s nonprofit status or perhaps change how its endowment worked.

There are a lot of tools that he could deploy here. There is also an excise tax in the Big Beautiful Bill that taxes foundation assets. I do not know what the final language was, but there was a version I saw in which the excise tax on foundations was upward of 8% a year.

I do not know if that was the final version, but that is an enormous amount. Let us say it is half that—let us say it is 4%. If you are paying a 4% tax on your endowment every year, it starts to add up. For Harvard, that would be about $2.5 billion in additional taxes every year.

And reportedly, Friedberg, Harvard is at the table and in discussions with the White House after a couple of months of defiance. What is your take on this? Is this an important priority for America and for the Trump administration? Is it a sideshow? What are your thoughts on the larger ramifications of this?

Let me just suspend the brand, history, and legacy of Harvard for a second and talk about what we will call prestigious higher-education institutions. What are the 2 primary functions of these institutions?

The first is to educate students, and the second is to conduct research or create facilities for research. Remember, these institutions do not direct research. They recruit and enable researchers who apply for grants to get funding to do their research, and then they educate kids.

I think the internet was the first leg on the stool to break higher education. The internet democratized access to information and knowledge. You can watch MIT graduate courses. All of the core educational content delivered in prestigious higher-education institutions has been largely democratized and is broadly available for free on the internet.

That is an incredible transition that has happened for humanity, society, and the world. AI is the next leg of the stool to break. I think AI may actually break education. It may break higher education, and eventually it may make its way all the way down to childhood.

We need to rethink from first principles how we educate, how an individual gets educated, and what other benefits they get from an educational system besides core domain knowledge. There is also socialization and experience with project-based work.

But I think AI fundamentally rewrites the ability of an individual to get a good-quality education. We could see kids in Africa and kids in South Asia getting the equivalent of a Harvard graduate-school degree at a cost of zero through personalized tutoring enabled by AI and ubiquitous access to knowledge and information.

So that core function of the university is broken, and these institutions are now starting to reconcile what that actually means for the long-term viability of all of these higher-education institutions in the United States.

The research function, I think, is also being rewritten around the world. In Europe, China, and Asia, there are independent research institutions that get research funding and can show up and say, “Hey, this institution is just being used to run research. It doesn’t necessarily need to be within an educational framework. It can be an independent research institution that focuses on either a topic or a domain.”

So I do think we’re going to see more and more independent research funding happening through the grants that come out from the federal government, nonprofits, endowments, and foundations that fund research. I think there’s a real reckoning underway.

It’s almost like these guys have created a monopoly. They’ve accumulated this capital, which allows them to build these great buildings, attract these great researchers, and then get the research funding to fund those researchers. They use that to raise more capital in their endowment, build the next building, and keep this thing growing. I think that’s breaking.

I have one question for you. I like where you’re going with this, but there’s one hole that I would like you to address: I don’t think any of that is nearly as valuable to most of the kids applying, or the parents forcing the kids to apply, as the brand and then the cycle that employers put back. How do you fix that loop?

Yeah, I get it. You can YouTube your way to something and AI your way to something, but at the end of the day, Goldman Sachs loves to recruit from Harvard, and that’s a really big deal because that’s a wonderful company.

I think the Thiel Fellowship has highlighted that you don’t get exceptional performance by exclusively going to people who have higher-education degrees from prestigious institutions. The Thiel Fellowship, for those who don’t know, offered significant funding to kids coming out of high school—18-year-olds.

I think the Thiel Fellows—Jason, you probably know better than I do what they’ve created, right?

A lot of startups—amazing startups.

But how do you deal with 500 million kids graduating each year globally with no brand differentiation? I think that’s an important question. How did the Thiel Fellows do it? And how do you hire, Chamath? You just did a program for 80 or 90 to find people, right?

Yeah. We did a coding challenge. I guess what I’m saying, though, is—

So what’s the alternative to that coding challenge?

No, no. The best way to determine if a person can do a job, if you don’t want to outsource it to a logo like Harvard or Stanford—which have worked in the past—is to watch them do the job. That is the number one way to do it. The way to do that is to hire them for projects and/or do internships.

That means you have to invest in what’s called professional development, and that slows companies down. So the hack is to just pick a logo. But, for example, in our venture firm, I created a training program for associates. We invested in that, and we hire 3 at a time. Typically, 2 of them make it, and sometimes just 1 who hits our notes. The other ones move on.

I think you have to invest in your own professional development in your organization. The organizations that do that then succeed and have a massive competitive advantage because they have their own training program and way to evaluate talent. That’s one way.

Doing a codeathon and scoring people is another way. Having an AI interviewer is another way. You can interview 10,000 kids instead of interviewing the 50 that you chose out of Harvard because Harvard is your first filter.

So the real question, Chamath, is: What are the mechanisms by which employers are going to create new filtering systems? We just went through 3 examples, but I don’t know if the brand holds over time.

Well, here’s the thing that what you’re saying doesn’t address: There are just a lot of kids who aren’t ready to bloom when they’re 18, 19, 20, or 21. I’m not sure that—I’ll just take myself as an example—I was a marginal performer in university, but I had co-op. I did well in co-op jobs.

Those co-op jobs were because these employers only wanted to recruit from the University of Waterloo. If I had to compete with 50,000 kids, I’m pretty sure I would not have gotten it because I was a bit of a layabout. But then, when I got into the professional working world after a couple of years, everything just clicked.

I think the problem we’ll have is that, in the absence of brand, it’s going to be very difficult to differentiate oneself and filter people. I think what Jason says becomes the huge problem: The burden of professional development for all these young kids is extremely heavy.

I think the idea of all of this tooling is good and necessary, but it’s insufficient. There’s something else that we need, because I think being able to differentiate yourself in a coding challenge is not the future either. It’s for a very, very narrow class of person.

Look, project-based work and social adjustment—there are clearly other really important skills and development cycles that are needed for people. It’s not just dumping knowledge into your brain. This isn’t like The Matrix, where you can just turn on the knowledge and gain it.

I do agree. I think there are other systems by which that will happen, and those systems will output filters. I’m just not sure it’s the same system that we’ve used for the last 250 years.

It’s not.

I think that’s the really interesting part of the discussion we’ve gotten to, Friedberg: We had an incredible system, a series of layered iconography that indicated, “Hey, you get 10 people from this group, 100 from this group, and you slot them in, and your company’s going to operate.” Parents don’t have to worry, students don’t have to worry, nor do the companies.

In this new world, companies have an opportunity by creating professional development. And you know what? The people who couldn’t get into Harvard because they didn’t have the connections, weren’t legacy, or didn’t have the wherewithal to pay for SAT tutors or whatever it took to hack your way in and have a legacy family member—or whatever racist policies they had that didn’t let Asian people in because they didn’t have the right personalities—all that nonsense goes away. That actually benefits them.

Their grades were too good. Exactly. Their performance was too high. In all fairness, you did wrong. It’s like the dumbest thing I’ve ever heard.

But here’s the good news: Anybody can build a project in the world and refine their skills right now based on all the information that’s on the internet. You and I, Chamath, are part of that group of people. We made our own luck. We made our own projects. We had some level of grit, autonomy, and self-reliance to do that.

The feedback I got from our previous discussion about this was everybody talking about how poor people, as a group, can never strive and never build anything. The truth is, it’s the easiest it’s ever been to build a company, build a project, and be independent as a creator of a product or service in the world. It’s the easiest it’s ever been. It requires the least capital and the least amount of time.

We’re selling a narrative to people that they’re helpless, when in fact they are super, super empowered. The only thing you can’t do is put yourself $200,000 in debt, because you’ll never get out of it.

Mm-hmm. I actually think this is a huge, beautiful opportunity to reset the system. But it’s not going to be like going to a supermarket, picking the brands you want, and then filling positions. You’re going to have to be self-reliant, all the students out there.

And that’s what Peter Thiel got right. You’ve got to give a shout-out to Peter Thiel here. I think, Friedberg, when he did the Thiel Fellowship, the way they picked people was that they had a mission to accomplish something in the world, and they were making progress toward it.

That’s actually the criteria they used. Did you have enough inner resolve to actually pick a mission, and did you actually do anything to steer yourself toward it? When you make it a competitive sort of program like that, they just picked the people who picked the most interesting missions and had made the most progress.

But there’s definitely going to be a hole left in society if these degrees do not correlate with performance and reality.

All right. On Tuesday, Grammarly, which I’m a huge fan of, acquired Superhuman, that amazing, super-fast AI email tool from Rahul Vohra, in which I was the first investor. Superhuman had raised $114 million and was valued at $825 million during peak ZIRP, according to Reuters.

Superhuman has annual revenue of $35 million, and Grammarly seems to be building a little bit of a suite of AI workplace tools. They bought Coda, which I was not an investor in, but I am a huge fan of that product as well. It’s similar to Notion, another product I’m a huge fan of. They bought Coda back in December.

On top of that, Figma filed its S-1 with Q1 revenues of $228 million. Remember, they were going to get bought by Adobe before it was stopped. They have 13 million monthly active users, $1.5 billion in cash, no debt, and it turns out CEO Dylan Field has 75% voting power. So he’s in founder mode. Very nice. He’s also a Thiel Fellow. Really interesting cat. I’ve had him on my other pod.

Figma is going to try to raise $1.5 billion in its IPO. That would match CoreWeave, which was the biggest tech IPO of the year so far. We are on a heater. Circle went public; that was up 7x from its IPO peak.

Chime went public slightly higher than its IPO price. You had eToro, and Hinge Health also went public. Wealthfront, which I was an angel investor in, just filed to go public. Yum, yum.

And a bunch of M&A transactions. We talked about DoorDash making 2 purchases. Sam Altman bought 2 companies at OpenAI, and tons of M&A is happening at the same time. According to Polymarket, there's a 52% chance of a rate cut in September and a 46% chance of no change. So we're definitely not getting an increase, according to the sharp money.

I think Powell said he would have cut if there hadn't been the tariff curveball thrown into the system, which is what most of us thought. And Chamath, you talked about all this sidelined cash sitting in money-market accounts. Markets are at an all-time high. Uber blew past $88, so I should be retired right now.

What are your thoughts on M&A and IPOs? It feels like we've got a really frisky, hot market right now. Does it make you nervous, or does it feel like this is where we should have been all along and Biden was putting a headwind against all this?

Yeah. So here's the crux of the issue: I think this is the intersection of a lot of really interesting things happening right now. You have Meta giving individual human beings $300 million to $500 million packages like they're NBA first-team All-Stars. OpenAI's revenue numbers just leaked. They're forecasting $13 billion in 2025, spiking to $125 billion in 2029. You have Anthropic; their revenue by 2027 is forecasted to be about $35 billion.

So what does all of this tell you? To be honest, it's telling me that the state of software is a little unclear. Meaning, I actually believe the OpenAI and Anthropic numbers. I understand why Facebook is now spending as if there's an existential risk. And I think the existential risk is that these models could be so foundational to how social experiences and work are done that they start to absorb a lot of other stuff.

So the question is, how do other tools fit into a workflow when these things become so central to how people both enjoy their free time and spend their productive time? When you look at that and look at Figma, what I would say is, on the surface, in the absence of these AI businesses, I would say, “Man, what a gangbusters business.” Growing by 40-some-odd percent a year at this scale, with adjusted operating margins—I think—of 18%. I don't like “adjusted” because it's adjusted for stock-based comp. I don't know what it is when you add that back in. But the point is, it's a phenomenal business.

The question that I think the institutional investor will have is: What am I buying? And does this revenue growth sit adjacent to core-model revenue growth? Because the big question that we have yet to answer—and this is not a Figma-specific issue; it is an industry-wide issue—is how much these foundational models absorb into what they do for what you pay them. And I don't think we know the answer to that yet.

So if all of these things just become excellent coding tools, then all of this high-level software is free and clear, right? It's in the safe zone. But I think the problem is we don't know that that's the case. And so I think in the IPO, what you're probably going to see is people approach this company the same way that they approach all non-core AI IPOs: It's a business that you love to own for a year or 2, but if there's a compression in valuation, it's because people cannot underwrite years 3, 4, and 5.

Friedberg, do you believe that the Star Trek communicator—just double-click on your pendant and ask the computer to do something—means there's 1 piece of software in the world that does everything and this whole long tail of business software just goes away? And if so, on what timeline?

No.

Okay. Yeah, I mean, this speaks to revenue quality and revenue stability. I think you used the term “How brittle is it?” Chamath, what are your thoughts on Chamath's angle here, of that unknown? Where would you fall either way? Are we going to have a suite of products, or does it get more narrow?

Yeah, I mean, more narrow. Well, it depends on the application. I think Figma's done a classic land-and-expand in terms of who they initially go after and what the suite of tools they offer does. By expanding that, they now can offer a bunch of different people within an organization a set of tools to help them all collaboratively develop products and services.

You can see that in some of the numbers. Revenue growth is on the order of 40-some-odd percent. They ran a 43% operating cash-flow margin in Q1. So in Q1 of this year, Figma generated $95 million of free cash flow. They've got net revenue retention of about 130%. So this land-and-expand is proven out, and there's real durability, it looks like, to this business for now.

But to Chamath's point, what does 3 to 4 years from now look like? Does this get absorbed into ChatGPT? You could say that about any software at any point in time. I think the thing that makes this AI era different is that that transformative shift can happen overnight, where suddenly someone else launches a service and it's completely obvious because of what AI can do. But I think Figma's done a great job staying ahead of the curve.

The free-money trade, I think, instead of having to bet up or down, Jason, on AI, is that if I could get $50 million or $100 million of Figma, I would probably go long it and short an equivalent quantum of Adobe, and I would just book the spread. That's a safer trade because even if the AI-model thing comes around the corner and we don't see it, the company that's going to take a re-trade on valuation faster than Figma will be Adobe. So you'll be hedged, and you'll probably make money that way.

We look at something in terms of not just the quality of revenue in our investment firm. We look at the durability of it. Can this exist 2, 3, 4 years from now, and is the value accruing so much to the user that, for the amount they're paying, they just never think, “I should swap this out. I should replace it,” right?

The revenue durability of your iPhone is a good example of it. Despite people not renewing their phone every year, you still can't think there's a better option than an iPhone right now. Even my Google Pixel 9 Fold, as great as it is, just feels like that revenue is still durable. I wonder when it becomes less durable.

Can I push back on this? I think the question that it brings up is not whether the individual person can whip out a card and pay for it in 4 years. It's whether that individual person actually exists.

No, no, no, no. Meaning, we don't know what the layoff cycle and the pattern of layoffs inside companies may be with AI. Meaning, if we all become more generalized skilled workers and there may be many, many, many more companies, then the odds are more likely that you provision highly skilled, vertically specific work to a set of agents.

If that's true, then the tools that created incredible durability when the organizational chart of a company supported vertical specialization won't exist when, instead, you'd have horizontal capabilities that you work across. That, I think, is the big question that AI will bring to bear.

Again, it's not going to be overnight, but that's where people will front-run those trades. And the market, specifically if they sniff this out, will want to price that 24 and 36 months forward and say, “This is what the end state looks like.”

A way to help people understand that is to imagine you outsource HR, and you don't have a 6-person HR company like you talked about 2 years ago, and then you don't need HR software, right? So your point is, each group has a set of SaaS software and tools it uses. If that group goes away because it's just abstracted into the AI machine and outsourced, there's nobody to buy it in the organization. There's nobody going to the CFO saying, “I need this HR software. I need this project-management software.”

I think it's a really interesting point, and the way, as a founder, Friedberg, to avoid this is to have a product that services many different needs for those customers or that organization. I think that's why I like these tools, like Coda and Notion: They kind of infect many different departments in the organization, right?

We weren't going to do it this week, but we said, “You know what? We wanted to talk about a couple of issues, we wanted to see each other, and it's the slowest news week of the year, but we wanted to get together and hash out some of these issues.”

Anybody got big plans for the weekend? Any recommendations for people? I need some books to read. You guys got any shows or books you're reading right now? Any albums you're listening to? Anything you're obsessed with?

I'm reading Modern Poker Theory by Michael Acevedo.

Oh, really? This goes into GTO and that kind of stuff?

Yeah, fabulous book.

What's your big takeaway thus far? Did you write a note about something or highlight something?

No, I mean, I'm just tuning up my game. Always tuning up my game.

Oh, you're doing a little tuning?

Well, I mean, I hope it's because we have something going on in November.

You want to do—

Well, we can't talk about that.

Well, I don't know. If this poker thing happens, maybe we can't play poker anymore in America. I don't know.

Oh, yeah. The big bill—we didn't mention this—but the Big Beautiful Bill says, I don't know, maybe you can't be gambling. We didn't talk about it, but that's a reason to get up in arms about this bill. If it gets rid of the poker ledger, what do we do with the ledge?

No, the ledger. And the ledger will survive. Actually, actually, this reinforces the value of the ledger because you'll just run the ledger infinitely.

You may have to have an infinite ledger. Never settle. I was thinking my proposal for the ledger was: if you’re under 100 dimes, you roll. If you’re over 100 dimes, you clear each year at the end of the year, because it’s just not worth the tax implications. You know what I’m saying? It’s not worth the tax implications.

We may have to create an offshore blocker and fund it with stablecoins. I mean, the whole thing is going to be really complicated. Our lawyers and accountants are going to have a field day with this, just so we can flip coins and play bomb pots.

Friedberg, are you watching any shows or movies? Maybe you could give us one of your great deep pulls for a science-fiction film that people should watch this weekend if they want to get some joy. Maybe Silent Running. You like Silent Running?

Logan’s Run. Silent Running. Those are good choices.

Yeah, and I think I watched the Bob Dylan flick on the flight to Italy.

Shout-out to Timothée Chalamet. What a great film. Okay, I have to be honest with you: I was not a super fan of Bob Dylan’s music before. But then I was like, “Wow, the body of work is really impressive.”

Let me give you 2—and, sorry—Joan Baez and their music together are incredible. I was a little short on Joan Baez, too, and I was like, “This was a mistake.” I’m just going to give you 3 albums to listen to. Blood on the Tracks. I want you to listen to Blood on the Tracks. Then I want you to listen to Infidels, another one, which is like his best of the ’80s. Infidels and Blood on the Tracks, those are 2, and Empire Burlesque, a third one.

Empire Burlesque and Infidels from the ’80s—this is Dylan at the really interesting height of creativity. And then there’s Blood on the Tracks, post his ’60s and ’70s folk-rock stuff and that transition. People consider Blood on the Tracks the seminal album.

I’ll also give you a deep pull of Street-Legal, with an incredible track, “Changing of the Guard.” You’ll love this, Chamath, because you also like The War on Drugs. The War on Drugs was very influenced by that era. So you’ve got Blood on the Tracks, Street-Legal, Infidels, and Empire Burlesque.

Those are JCal’s Dylan choices. I hate drugs. The War on Drugs. Did Dylan do drugs?

I mean, it was famous that he was on speed for a little bit in the ’70s, reportedly, and that’s where he had a lot of productive days. But I think he introduced the Beatles to LSD—that was the rumor.

Friedberg, what are your thoughts on my Dylan selections, or do you have one of your own?

I’ll leave it to you, JCal. You got any movies or something?

I did recently rewatch Arrival by Denis Villeneuve. How do you pronounce his last name?

Villeneuve.

Oh, he’s doing the new Bond. That’s going to be the new Bond.