伊朗战争、油价冲击、退出路径、AI营收爆发与公关噩梦

All right, everybody. Welcome back to the number one podcast in the world. Friedberg's out saving the world, creating new potatoes, or, I don't know, quinoa, maybe some Brussels sprouts. I'm not sure what he's working on at this point. In his place, his personal favorite bestie—he always says that when I'm not here, I want Brad Gerstner in the seat. Welcome back.

We haven't seen you on the pod since your shout-out at the State of the Union. Take us behind the scenes for a brief moment here, Brad, of what it's like to get a shout-out from POTUS at the State of the Union. Did you know it was coming? Did you choreograph this thing? Did you choreograph that, or was that more spontaneous?

Had no idea it was coming. In fact, I found out after the fact that it wasn't in the speech, and the president added it to the speech. So I don't even think it was a few days before it was going to happen.

We got an invite to the State of the Union, and listen, it's an institution. This has happened every year for 250 years in the country. I've never been. I did know he was going to talk about Trump Accounts, so I figured if I'm ever going to go, that's the time to go.

I have to say, I'm just a sucker for democratic institutions and democratic traditions. It was an extraordinary night. Set aside the headlines about what Democrats did or Republicans did, just the fact that whether it's a Democratic president or Republican president, this happens every year. You have to go report on the State of the Union. So it was a special night.

We did dinner ahead of time. We're in the chamber. The chamber, as you all know, is very small. Just to your right was the first family, and Jared, and Ivanka. We were there to observe, like everybody else. And wow, it was quite a moment.

I want to just say, you did a great job, because when you sent your heart out to all of America, I took it. I took it. I took it. I was like—

You held it out, and you cupped it at the right angle.

Right. Right. You had to get up a little bit extra and out and out 5 degrees. You're being a Nazi. But it would have been no bueno. Those would be some super-racist Trump Accounts. Keep your protractor and your ruler out when you send your heart out.

We're signing up over 100,000 kids a day to these Trump Accounts. We have millions of kids who've already claimed their account. We have nearly 30 million kids in America who are eligible for at least $250 if they just go claim their account. These things are going to be live on July 4th.

What it really showed, I think, to the country was that it accelerated after the State of the Union, because the president really believes this is a way to get everybody in Main Street America into the game of capitalism and get them all directly owning the great companies in America. So it meant a lot to me in that regard that it highlights the importance of the program. I was deeply grateful to the president for not only making sure this happens, but the shout-out is pretty cool.

Good for you, bro. I have an interesting idea for you. I'm sure it's come up already, but with this whole discussion of UBI, somebody said to me, “Oh, you know, I really like these Trump Accounts your friends did, the Invest America, because it's like the start of UBI.” And I was like, “Well, that's not exactly the intention, but I get it.”

With wealth disparity going on in the country that has a lot of people concerned, what if there was a giving pledge around equities? People could opt into it; they don't have to. But if somebody like, I don't know, Larry and Sergey or Zuckerberg said, “I want to pledge 5% of my shares to go into kids' accounts over the next 20 years,” what an amazing, beautiful thing that could be. It would be incredibly material to get whatever it is—a tenth of a share, a hundredth of a share, a thousandth of a share of whatever company. Has that come up yet as an idea? I'm sure it's obvious, right?

It's come up. Stay tuned. But yes, we're going to have some banger announcements as we head toward July 4th.

All right, let's talk about the war in Iran. Obviously, there are much more important issues than financial ones—life, death, the freedom of the people of Iran—but we're uniquely qualified, I think, to talk about the economic fallout, second-order effects, first-order effects, and there has been massive volatility over the last 5 trading days.

Just talking about Brent crude oil, and we'll key the discussion off of that type of oil: It spiked to $84 on Friday. That was day 7 of the war; $119 on Monday, day 10; dropped back down to $84; jumped back up to $100 after 3 commercial ships were hit in the Strait of Hormuz on Wednesday. Those ships, by the way, were not oil tankers. They were carrying cargo. They were flagged as Thai, Japanese, and Marshall Islands. Brent crude is currently at $99 when we're taping this. It'll be at something different by the time you listen to the pod, I'm sure. But it's quite a spike.

And here's a second chart that shows you the spikes over time. I was old enough to remember the oil shock of 1978. We had to get in line at the gas station based on your license plate number, and you had to wait an hour or 2 to get gas. Gulf War, obviously, it hit $100 in 2026 dollars. In 2008, we hit kind of a peak moment, $216 in today's dollars. That was the peak oil discussion. Demand from China went off the charts. When Russia invaded Ukraine, we hit $115, which would be $133 in today's dollars.

So this is not new, but it is significant. And breaking news today, Iran's new supreme leader, Mojtaba, says he's keeping the Strait of Hormuz closed as a tool to pressure the enemy. The Wall Street Journal on Thursday quoted a senior fellow at the Middle East Institute saying that reopening the Strait of Hormuz will require ground troops. Polymarket gives a 27% chance that U.S. forces enter Iran by the end of March, and 57% by the end of the year. So sharps over at Polymarket believe we will have boots on the ground.

Let me stop there. Brad, your thoughts on what happens when oil hits this kind of number and we have this uncertainty of, hey, this could last 2 more weeks, or it could last 6 months, it could last a year. Nobody seems to know, and how it resolves—we just had a really interesting talk with Graham Allison—how it resolves is also a major unknown. Your thoughts?

Right. So first, obviously, there are huge direct costs as oil prices go up, right? Oil is a component of a lot of consumer and enterprise products. It also hurts consumer confidence and enterprise confidence.

Goldman Sachs is out today with some analysis where they updated the economic knock-on effects. They raised their PCE inflation forecast from 2.1 to 2.9 for the year. So that's a huge jump in terms of their expected PCE inflation. Core PCE, which excludes oil, they forecast up from 2.2 to 2.4. So they're saying even if you excluded the direct price of oil, the knock-on effects are going to cause a little more inflation.

They lowered their GDP forecast by 30 basis points for the year. They also expect higher unemployment as a result of this for the year. All of that is weighing on sentiment in the market. Remember, just a few months ago the S&P peaked at 24 times; now we're at 21 times.

But I think the market may be getting it a little bit wrong, right? The Trump doctrine—I tweeted about this last week—I think the Trump doctrine is far more pragmatic than the neocon doctrine. I think Trump has a very limited set of goals. He wants to destroy and degrade threats to America's national security interests. He doesn't want to spread democracy. So my suspicion is these impacts are shorter duration, but right now the market's having a little bit of post-traumatic stress, flashbacks to Afghanistan and Iraq, and wondering if we might be wandering into a quagmire.

All right. And just in terms of the doctrine, he has said he wants to see the people rise up there. So it might be splitting hairs, but I think he might not actually be for regime change. He says he wants the regime to change.

All things being equal, I don't think he minds if the people bring it to themselves. The question is whether the U.S. is going to put boots on the ground and try to spread Madisonian democracy like the Cheney doctrine was. I think this is very different.

Chamath, your take on the economic impact and any other things you'd like to add about the war in Iran.

I think the most important thing that I saw this week was that President Trump was asked about the war, and he said the war would be over very soon. What did the market do? The market literally took oil from $120 a barrel to $90 a barrel almost in a nanosecond. I think that sort of tells you what everybody thinks.

To the extent that the market really didn't believe it, oil would not have budged. If anything, it would have faded those comments, and you probably would have seen oil stay at around $120 or even go slightly higher. So the fact that there was this reflexive move, I think, is a belief by a lot of the sharps that there is no path to a sustained conflict.

There's going to be a lot of chest-bumping from the Iranians, obviously, because they need to save face, and they will want to set up whoever comes next to have the most successful chance of governing. So my perspective is that that was a trial balloon. I think it validated what everybody thought, which is that this is going to be a short-run thing.

I agree with that. The downstream impact is, I think, exactly what Brad said, which is it could show up in some short-term price spikes. But then on March 11th, you saw what Chris Wright did, which is the president activated a whole bunch of member countries in the IEA.

I think Chris released 172 million barrels. I think there's a coordinated release of about 400 million barrels of petroleum. That's going to dampen the effect of any price spike. On top of that, I think the estimate is that there's probably another billion or so barrels that one could release from strategic stockpiles.

I think that these two things together paint a picture that probably the worst is behind us. And I think now it's about finding the off-ramp. Sacks, your thoughts?

Well, I agree that we should try to find the off-ramp. I agree with what Brad and JCal said about that. Look, we've degraded Iranian capabilities massively. Their army, navy, and air force have all been destroyed. This is a good time to declare victory and get out, and that is clearly what the markets would like to see.

You are seeing, however, a faction of people, largely but not exclusively in the Republican Party, who want to escalate the war. They're calling for things like ground troops or regime change, or they simply want the pounding of Iran to keep going on and on. I saw an op-ed in The Wall Street Journal to that effect, saying that we shouldn't try to find an off-ramp; we should just keep going with this.

I want to lay out some of the risks of what an escalatory approach could entail. First of all, we're all seeing that the Strait of Hormuz is closed right now. We don't want that to persist longer than it has to, but there are actually worse outcomes than that.

If the Iranians get hit, if their oil and gas infrastructure gets hit, they've already said they're going to engage in tit-for-tat retaliation against the Gulf states. We recently saw the Iranians blow up this giant oil depot in Oman. You saw some of those images. They could continue to target the oil and gas infrastructure across the Gulf states.

If that happens, it won't really matter if the strait gets reopened, because you won't be able to restart oil and gas production in the Middle East. So that would be, I think, a much worse outcome that could result from escalation.

Furthermore, there's an even worse scenario, which is that the region is very dependent on desalination plants. I think something like 70% of Riyadh gets its water from desalination. I think it's something like 100 million people on the Arabian Peninsula who get their water from desalination. It's basically a desert, right? Those desalination plants are soft targets.

You already saw that one desalination plant in Iran got hit, and then Iran, again tit-for-tat, hit a desalination plant. I think it was in Kuwait. I could be off about that. But in any event, if you see that type of destruction continue, you could literally render the Gulf almost uninhabitable. You're just not going to have enough water for 100 million people, and human beings cannot survive very long without water.

That would be a truly catastrophic scenario. We're talking about destroying the Gulf states economically and also from a humanitarian perspective. So I think we have to take things like this into account when you hear people preaching or advocating for escalation.

You also have to consider the impacts on Israel. It's hard to know exactly how much damage Israel is taking right now because there's a social media blackout. But what you're starting to hear trickle out is that Israel's getting hit harder than it's ever been hit before in its history, and we're only 2 weeks into this. If this war continues for weeks or months, then Israel could just be destroyed, or very large parts of it.

I think Israel is a harder target than the Gulf states. Its infrastructure is more hardened, and it's further away. The Gulf states are vulnerable to drones and short-range missiles, whereas Israel is mainly vulnerable to long-range missiles. Nonetheless, at some point its air defenses could become exhausted, if that hasn't happened already, and Israel could get seriously destroyed.

Then you have to worry about Israel escalating the war by contemplating the use of a nuclear weapon, which would truly be catastrophic. So there are a lot of scenarios here, a lot of really frightening scenarios about where escalation could lead.

Even though the United States is a much more powerful country than Iran, Iran essentially has a dead man's switch over the economic fate of the Gulf states and potentially even beyond that—the habitability of some of these countries. So I do tend to think that this is a good time to declare victory.

I think Brad, you're right that the president has never said that democracy promotion is one of his objectives. Yes, JCal, obviously everyone would welcome it if the people rose up and chose a new regime, but that's not something we've said we have to accomplish. This would be a really good time to take stock of where we are and try to seek an off-ramp.

If escalation doesn't lead anywhere good, then you have to think about how you de-escalate. De-escalation, I think, involves reaching some sort of cease-fire agreement or some sort of negotiated settlement with Iran. We can get into more of what that looks like, but I think the big picture is that if escalation could lead in all these horrifying directions, then I don't think that's the right approach. You have to look at de-escalation.

JCal, where are you on this?

Complicated. I have my personal feelings on regime change, and since we don't have the information that the Mossad, the CIA, and Trump have, I do think Trump would only do this if he had a very high probability of success and an off-ramp. However, it's not looking good with the off-ramp right now, and it could be quite chaotic.

If the neocons get their way and the people on Polymarket are correct—the sharps who say there's a 57% chance we'll have boots on the ground—I think this is kind of the end of Trump's second term. If you were to put together the series of mistakes that he's made and that the administration has made, they're really at the heart of why people voted for him.

Take starting a war like this, specifically with Iran. That's what we were told was the reason to vote for President Trump. He was not going to take us down this path. He was not going to risk World War III. He was not going to risk a nuclear possibility, as Sacks correctly points out.

Now you have all the MAGA supporters, from Tucker to MTG to Rogan, Matt Walsh, and Megyn Kelly, all up in arms, saying this is the end of MAGA and a massive betrayal. There's the 1B betrayal, where Trump wouldn't release the Epstein files. We'll put that aside because I don't think that's as important as starting World War III.

Then there's the insane, unnecessary cruelty of ICE agents that we've talked about many times on this podcast, which he has corrected by getting rid of Kristi Noem. You start putting these things together, and if this continues for another 6 months, it's basically going to result in the Democrats doing a clean sweep in the midterms.

Here's the chart that I think the Republicans really need to look at to see how misguided this all is. This is the chart that should be absolutely terrifying. Nobody wanted this war, or very few people wanted it, besides the neocons, probably the people of Iran, and the Israelis.

The chance of the Democrats sweeping is now up to 45%. This just happened. The Democrats are going to sweep. They're then going to win in 2028, and the entire agenda of MAGA and Trump's 2.0 will be gone.

Then you look at just absolutely ignoring the working man. Inflation going up above 3%, as you pointed out, is a likelihood. Brad, unemployment is ticking up—still very low, but it's ticked up 10%, worth keeping an eye on. These foreign-affairs issues are the least important to the American people. They're very low on the list of priorities.

People are looking at Trump and what they believe is the enriching of his family and all these business deals, and then they—

Kitchen-sinking it.

It's literally what I was thinking.

You're bringing everything.

I'm bringing everything. 1, 2, 3, 4. Number one, starting the war that everybody said he should not do—and that was why we should vote for him. Number two, the Epstein files. Number three, the ICE cruelty. Number four, not working for the American working man who doesn't own equities.

Those are four. 1, 2, 3, 4. It's not a kitchen sink. This is not my personal feelings on this. This is my assessment of the situation. If he doesn't find an off-ramp quickly, they're going to lose both houses in the midterms. That's, I think, the thing Trump needs to really consider, and I think he will consider it. I think he's going to find an off-ramp.

Right. That is the topic. The topic was whether we're going to find an off-ramp or not find an off-ramp in Iran, and I think Sacks made the argument that there's danger that the neocons and others are arguing that we expand and put boots on the ground. You're saying if he doesn't, it'll be a disaster. Chamath and I both say he will, right? And so—

No, wait. Wait. Will what?

He will find an off-ramp in the nearer term because the Trump doctrine is not the neocon doctrine. As much as people want to talk about Iran, Iran, Iran, I think, as I explained last week, this is about China, China, China.

You have to remember that at the end of this month, he has a pivotal 3 days with Xi Jinping in China. This is going to be an absolutely historic convening of the 2 superpowers that run the world. One is us—we are the established power—and one is China, which wants to be reascendant.

I would bet dollars to donuts that there's going to be an enormous incentive for Xi to negotiate a grand bargain in those 3 days and do something historic for himself. I think the president will use that if he thinks it creates leverage.

I think that's a great insight. How does the Strait of Hormuz open? If this war is dragging on and Israel, which seems to be the driving force in this, keeps it up with Iran, how do we ever get the strait open again?

I think the off-ramp is that the United States declares victory, does what Sacks says, and says, “Listen, we degraded and destroyed. That's what we came here to do. We did not come here for some experiment in democracy. We wish the best to the Iranian people to do the things they need to do.”

If Iran does not back down, if after that declaration Iran continues to destroy cargo containers moving through the narrow straits, I think you're going to see Iran's neighbors, Israel, and others get very involved as it pertains to Iran because it's in their interest.

Listen, the United States produces 20 million barrels of oil a day, and we consume 20 million barrels a day. This is a modest problem for the United States. This is a massive problem for China. This is a massive problem for Asia. This is a massive problem for all of our friends in the Gulf who are trying to dodge Iranian missiles right now.

So, there are a lot of people in the world who will take up arms to deal with the Iranians if the United States isn't there because we can take care of ourselves.

Your position, Brad, just to confirm it, is that we are going to leave the war in the next 30 days, and then if the straits are not open, China, India, and all the Gulf countries that are impacted by it will protect it. They will fight Iran.

I think they'll put a lot of pressure on Iran not to continue firing missiles at their ships, right? At the end of the day, this is not just an American problem.

Let's be clear: we're always involved in this part of the world. The only question is whether we're going to have an active armada that's engaged in active military activities against Iran.

What I'm suggesting, again—and listen, anytime you try to clean up a mess like this, there is risk. This is not a risk-free initiative by the United States, nor was Venezuela. But let me steel-man the alternative.

Doing nothing and allowing Iran to procure the ingredients for a nuclear missile when they are set on the destruction of the United States and U.S. interests, doing nothing in Venezuela while the Monroe Doctrine is totally wrecked and we let our adversaries take up positions in South America—those also have risks, right? Those carry a lot of risks.

And so, we're weighing these 2 risks. Again, for me, I don't like the fact that we're engaged in military activities here, but I will tell you I am very much on the side that if we're going to go protect American national security interests, you go in, you degrade their capability, and you get out. And I think that's what I hear out of the president.

Chamath, you had a follow-up.

All roads lead to China. I think you're going to see Xi offer up a grand bargain, and I think it's up to the president to decide whether he wants to take it and see what he wants to add to it to get something done. But I just don't see them meeting and coming out with nothing. I see them going in and coming out with something that's historic. And I think that all of this—Venezuela and Iran together—is all about China.

Let me just say one thing as to that, Chamath, because I think the point is absolutely spot-on. Probably the single greatest takeaway for us from an investment perspective at the start of this war was that the Chinese didn't take up arms on behalf of Iran, aren't defending Iran, and didn't cancel the summit with the president.

Because they need him. They need the oil. 20% of their entire domestic consumption is oil from Venezuela and Iran. 20%, but it's not 20% because it's literally 100% of anything that's feedstock, anything that's transport—cars, buses, planes. They are in an enormous world of hurt.

Now, they have a strategic petroleum reserve as well, and it's quite robust, but it's not robust enough to sustain 5 or 6 months of this. It's not that robust. So, at the end of the day, who is going to be hurting the most? It is China.

And so, if you play this game theory out, the reason he kept it is because now he needs the summit even more. Could you imagine if the president canceled? That would be a disaster for the Chinese. So, the fact that it's still on the books, if I was Xi, I'd be like, “How do I negotiate and help find the off-ramp? How do I end up fixing this faster?”

All right. Remember, you have 25% unemployment of young men inside of China—25% today. What do you think it goes to in 5 months with no oil?

That's the unemployment rate you should be focused on, Jason.

Oh, the China issue is a separate one.

Separate. No, no. That was separate from my point. My point is, I was bringing up a different point.

The kitchen sink didn't include the Chinese. I get that. I'm just adding to your kitchen sink. I didn't have a kitchen sink. I have 4 very salient points. All right, Sacks, I'll give you the final word here.

Well, look, that was a bit of a broadside, Jay Cal, where you kind of did kitchen-sink it. But here's the part I'll agree with you about: it doesn't take a political genius to understand that long wars are unpopular. It will hurt the Republicans in the midterms of 2028 if this does turn into a long war.

Fortunately, I think the president understands that. His political instincts are impeccable, and he's always favored short, decisive, swift military actions, whether it was Midnight Hammer or the Maduro raid. I think that is his inclination and preference.

And I think we are pretty much at, or close to, a point where the president's going to have to decide on next steps. I think he's indicated that we have completed our objectives, and I think it's just important that we don't let this neocon wing of the party try to expand the objectives or aims of the war because, frankly, they've always been wrong about everything.

I mean, these are people who never wanted to get out of Iraq and Afghanistan, and would still be there after 20 years if they had their choice. So, I think it's just important not to listen to those people.

And look, it's not just 1 op-ed in The Wall Street Journal. The Wall Street Journal is kind of the tip of the spear representing that whole neocon establishment. And I think it's just important that this is the time to, frankly, ignore those voices and let the president do what I think his political instincts are telling him to do, which is to wrap this thing up.

I'm in strong agreement, and it is my hope, too, that he wraps it up quickly and that we don't have any more loss of life.

All right, we'll keep discussing this ongoing breaking news story in the coming weeks, but back to our zone of excellence: AI and tech. OpenAI and Anthropic are scaling revenue and costs faster than we've ever seen in the history of, well, business—the world. Revenue at these 2 companies is growing at unprecedented levels.

Here are the reports, and I believe you're an investor in both these companies, Brad. Anthropic hit a $14 billion run rate last month, in February. That means they have grown revenue from $1 billion to $14 billion in 14 months. Yeah, 12x year over year. They're valued at a meager $380 billion last month. This feels like a bargain given the growth.

OpenAI ended 2025 at a $20 billion annualized run rate. And they've grown revenue from $2 billion to $20 billion in 24 months. They're valued at $840 billion last month. And, man, it looks like Sam Altman has Dario in the rearview mirror. He could get lapped any moment.

Lots of debate. Where'd you find this? What the hell is this?

That one, I made that. This is Dario closing in.

What is that? A velociraptor? What is it? A T. rex? Famous scene from Jurassic Park. Oh my God.

But I don't think anybody expected Dario to be coming around the bend this fast, but he's right behind, apparently, and they're obviously winning the business-to-business side of the business. The J-curve on these companies is insane. $250 billion, $500 billion—who knows what gets invested before these companies reach profitability, Brad, but you're invested in these 2 companies.

Unless you sold when Sam Altman told you he would buy his shares back on the famous BG2 episode. I don't think you sold it.

I bought a lot more since then, Jason. I bought a lot more since then.

Well, I mean, listen, love your children equally. They're both incredible companies. Anthropic unquestionably has a lot of financial momentum, and OpenAI is seeing a lot of momentum themselves, right?

But the single most important question this year was: Would AI revenue show up? And just 60 days ago, 90 days ago, there was tremendous skepticism. No way all of these infrastructure investments were going to pay off. There's no incremental revenue coming out of AI, including many of our friends.

But in January and February, we really had kind of a nuclear moment, right? The splitting-of-the-atom moment. I mean, we had a $6 billion month out of Anthropic in February, right? It was widely reported, okay? Let that set in for a second, right? $6 billion in a month, and it was only a 28-day month, okay?

That's more revenue than the annual revenue of Databricks and Snowflake, 2 of the greatest software companies of all time, after 12 years. They could do in the first 4 or 5 months of this year the total revenue of SpaceX this year.

What is driving that? Just explain to the audience what's driving it. Is it token use? Is it cloud subscriptions?

We crossed a threshold with Opus 4.6, and we saw it again with ChatGPT 5.4, where the models and the agents on top of them—whether it's Claude Code, Codex, or ChatGPT—are no longer competing with IT budgets. They're now augmenting labor. They're competing with labor budgets.

You could not possibly have a $6 billion month by displacing IT budgets. Millions of other companies across America are saying, “Oh my God, let's spin up these agents and have them do things for us,” and they're willing to pay for it because the product of that effort is worth the money to them.

The revenue and usage momentum, I will tell you, continues in the month of March, and it only accelerates from here. As Kevin Weil has said, “The models and the agents are the dumbest today they will ever be.” We're in the early innings of compute and algorithmic capability. To me, that's the observation of this moment.

Should they go public? I said yes. They should go public for several reasons. There's tons of institutional demand. They need cheap access to money to continue to build out the compute they need to support.

There is more compute constraint in these businesses today than they've had at any time in the last 3 years, so they need access to capital. Finally, I think you have to have the retail investor in the game. These are 2 of the most important companies in the history of capitalism and in the history of America. It's destabilizing not to have them public.

Jensen said last week that he expected the $40 billion he recently invested in these 2 companies would be his last money in because they were both going to go public. He thought they would both go public this year. I think they're preparing and heading down that path.

Listen, I want to get some of these shares into the accounts of all these kids that we're opening up because they're really, really important companies to the future of the American economy.

Chamath, you had some insight into the quality and durability of this revenue.

There's not a single good example that we can find of sustained positive margin expansion and impact of AI inside of a true corporate enterprise that is not right now a small test. There isn't one.

So where does $6 billion come from? Everybody has to show up to their board and have an AI checkbox. There are thousands and thousands of companies, and when you have tens of thousands of companies as customers paying $200-plus a month, it's not that hard to show up with that kind of revenue.

The real question is the following: If you take the Databricks and Snowflake example, the companies that use that software generate enormous revenues and enormous margins. These products are in critical production workflows that underlie those revenues and profits. That is just not true with AI today.

We have all kinds of claims, but we are still experimenting. Why are we experimenting? Because we know it's important, but we don't yet really know what to do.

You can't just slot this into a critical workflow in healthcare and all of a sudden show up where, if you make a misdiagnosis or a mischaracterization of a procedure, you can get fined and go to jail. The companies that are in healthcare don't do that.

If you're in financial services and you make a mistake about somebody's portfolio or make a misallocation and point to a model, you will get sued and you will be in trouble. None of these things have transitioned from “It's interesting, it's experimental” to “It's the core critical operational workflow.”

The interesting thing is that there will be a transition in revenue quality when that happens. A great example of this is Amazon. Why does Amazon issue an edict that says you cannot use this stuff inside of AWS unless a human now reviews and approves it?

Because what happened? They had 3 or 4 SEV-1 faults from a bunch of code that was written by agents that brought down AWS. I've told you, I love AWS for one reason: because it's hyper-reliable. I hate AWS for the same reason: that hyper-reliability comes at enormous cost. I pay it, but I pay it to never have a SEV-1.

The reason they have 12 nines of accuracy is because it's humans and deterministic code that never fails. It doesn't mean that 2 companies can't get to $20 billion, $30 billion, or $40 billion in revenue. What it means is we have to be honest: This is an industry that's early. We are all figuring it out.

There are a lot of test budgets that are going at it. It will slowly and methodically emerge into production, but let's not oversell what this moment is.

Okay, Brad, I want to give you a well-constructed question here to respond to, and then, Sacks, we'll go to you if you have some input. Of the $20 billion, how much of it do you think is experimental? What percentage is experimental versus production?

Strip out the consumer spending, because that's half of it.

Okay, so let's put aside the consumer. Sure, great idea. We'll put the consumer subscriptions aside. They're obviously getting value, or they wouldn't be subscribing.

By the way, for consumers, it can be extremely faulty, and there's no SLA that you're getting. These are phenomenal products.

Yes, for $20 a month, they're well worth it, and consumers have decided they're worth it.

And also, for the individual engineer—of which I suspect there are a few million—who gets to pay $200 a month and have their company subsidize it, the company knows that these costs are being incurred. But there is no tick-and-tie at the end of it where you then review the code in a different way because you're worried about hallucinations, as Amazon just demonstrated.

There's a story in the FT about Amazon having some blast radius from AI-generated code, and they're putting controls in place. We'll put that in the show notes. Brad, let's get to the specific question I asked. Of the tens of billions of dollars in revenue between the 2 companies that's not consumer, what percentage do you think is production quality versus experimental, to Chamath's point?

I've coined the phrase “experimental run-rate revenue” versus “annual recurring revenue.” I think Chamath's point is a really important one. As an investor, I have to discern what's repeating, what's recurring, and what's not.

What I would suggest is, of course, there's a lot of experimentation because these things haven't been around that long. But I suspect that Palantir, the U.S. government, the U.S. military, Nvidia, and a lot of other major enterprises would argue they've gone full production.

In fact, it's existential to the wartime effort going on in Iran right now. That doesn't sound to me like experimental as much as it sounds like production capability.

I will tell you what will prove this one way or the other: In the month of March, do revenues continue and go up?

Well, that's not true. The experimentation can go on forever.

That's not true. That's not true. Okay, well, if the experimentation goes on forever, that sounds like recurring to me. We have scratched the surface of the number of companies that even know how to adopt AI, so these numbers will go to the stratosphere.

I'm not debating that. Look, I'm on the same side of the bet as you are. I want these numbers to keep going to the moon. I'm just being much more circumspect and honest with myself to say: I see it on the ground. I sit on top of these models. I am paying these models millions of dollars a year.

What I'm telling you is my revenues don't go up faster than their revenues. I'm consuming more tokens every single day. Do I get more economic output? I am not.

I would say that my team is at the leading edge, and so I suspect a Fortune 1000 company is steps behind my team. If I am spending triple every 3 months and not seeing my revenues triple, I suspect these other companies are in a similar situation.

I'd finalize it, but ask it on Friday when you're with Michael Dell, because I've had this conversation recently with Michael Dell. Michael said that a year ago, companies weren't seeing ROI. Today, they're seeing very big ROI in their AI investments, but I think that's the question on the table.

But what companies? Which companies?

Yeah, let's talk about Goldman Sachs. Of course, they're probably seeing an ROI. Sacks, we're agreeing: The company is experimental in large part, and this is a new tool. By definition, you have to experiment before you put it in production. What's your take on this grand debate? How much of this revenue is experimental versus real?

Well, look, when you're talking about enterprise revenue, what you're really talking about is coding assistance. That's been the breakout use case. It's really the first big breakout use case on the enterprise side.

The consumer side has been more research and writing, that kind of stuff—the chatbots. But enterprise is all about coding assistance.

My sense is that the demand for code is very scalable. Software engineering has always been an area of the economy where companies have never been able to hire enough people. Even Silicon Valley, which is the most attractive place for software engineers to work, has never been able to recruit and attract enough of them.

The rate-limiting factor on the progress of every startup I've ever invested in is not having enough engineers to code up the product roadmap. Then you look at the rest of the economy—the Fortune 500 and so forth—and they have hardly been able to recruit software engineers at all because they've all gone to Silicon Valley.

So, I think you're dealing with a part of the economy where there's always been a massive supply shortage. I don't know what the natural limit on that is, but my sense is that there's tremendous latent demand for the ability to generate code in large quantities and create new products. As the cost of code goes down and coding assistance gets better, you can code up new types of products.

Then, of course, it's going to lead to agents, which is another way of basically using the code that gets generated. My sense is that this could be very scalable. I don't know where it taps out exactly.

Where I think Chamath is right is that I think there is a change-management aspect to this in Fortune 500 companies, for example, and they haven't really wrapped their heads around how exactly they're going to use it. There was a McKinsey study that showed that a lot of these pilot projects in Fortune 500 companies were experimental, and a lot of them were proving not to be successful.

So, I do think that as you go beyond coding into company transformation and things like that, it becomes a little bit more speculative. That's not to say it won't happen. I think it will happen. I'm bullish, but I do think that we're still waiting to see what the breakout use cases beyond coding will be. Probably agents will be the next big one, but I think Brad's right that that's big enough to see this scale for a while.

Because the thing about code is you're paying for code on a metered basis right now. You're paying per token, which is an amazing deal for companies, right? Before, they had to go through this recruiting process to find engineers, source them, vet them, keep them happy, and give them all the perks—the KIND bars and everything else. To be able to buy code on a metered basis as the cost per token keeps going down is an amazing deal.

It was just metered on people. Now, to your point, it's metered in a different way, but it's still metered. Let me just say, you used this term “labor displacement,” Brad. That's the one part where I might disagree with you, because there was such a shortage—

Yeah. Of software engineers, I think when people hear the term “labor displacement,” they might start to think that $6 billion of incremental revenue means $6 billion of layoffs, and I don't think it does. The way to thread that needle is the fact that we were artificially constrained in the number of software engineers, how they could be used, how rapidly they could be acquired, and all that kind of stuff.

To be able to now turn that on like electricity is such a huge game changer and unlock for the whole economy. Yes.

And that's what I think is really exciting about it.

It's augmenting human labor, right? We're not at a place yet where it's displacing it, and this is the definition of productivity gains. I'm going to make just 2 quick points here.

The place to look for this actually moving from experimental into production is not at big companies. Big companies are actively resisting this, and they will resist it because it means lowering head count. It means the person who implements it might actually implement themselves out of a job. That is the natural resistance you'll see in big companies. That's not where to look for adoption of new technology.

Startups are the place to look at this, and that's where I am on the ground. What I'm seeing there is that startups are using this in production for their legal work, marketing, SDRs, accounting, and reviewing legal documents. This is all work for which they would normally hire consultants, outsource, or make hires.

What I'm seeing on the ground is that it's production-ready in startups that are using it in those categories. HR as well—accounting, marketing, all of that. All that blocking and tackling, all those chores, are currently being done with these LLMs. They're doing it in production and they're doing it at scale.

Just a quick second point here: here's the J-curve, and this is the question I think we'll get to in our next segment. When does this become a profitable business? You asked this of Sam in that famous clip on the BG2 Pod. RIP BG2.

Here you go: the LLM industry J-curve. I just asked Claude to make this for me. If you have $500 billion, I think you would agree it's probably going to be around that number. Brad invested in all of this.

More. A lot more.

Okay, so $5 billion is an underestimate here. When do we actually see these large language model companies hit profitability in a calendar year? It took Tesla, Uber, and Amazon a decade-plus in each of those cases to win back their investment. If you put it at $10 billion—

This is a really good chart. Here's the precise math on this.

I am building a 1-gigawatt data center in Arizona. When I greenlit that project, I thought it was going to be a $4 billion or $5 billion investment. I was like, “Okay, whatever.” Then it went to $10 billion. Then it went to $15 billion. Then it went to $20 billion. Now it's upwards of $50 billion for the powered shell, all the land, all the permits, all the infrastructure, all the people—all of it.

Sarah Friar said, I think it was about a year ago, maybe less than a year ago, that for them, every gigawatt is about $10 billion of annual revenue. So if you think about that J-curve, Jason, the way to think about it is: energy equals intelligence. For every gigawatt that they're trying to spend, they have a 5-year payback. That's roughly what it means just to get to break even. Then years 6, 7, and 8 will be where the profit is.

Now, how do you shrink the J-curve? You have better silicon. We're going to see something from Jensen in a week or 2 that uses a bunch of the stuff that we partnered with him at Groq on. There'll be other people. There'll be open source.

All those things can shrink the depth and the surface area of that J-curve so that you can get out of it faster. But right now, that thing is roughly accurate: it's about $50 billion per gigawatt, and it's about a 5- to 6-year payback just to get into the money. Then it's about $10 billion a year. The technology industry has to do something to make this better.

Could I, though, take a step back and give you just a different framing of all of this?

Please. I think the big thing that we're debating is actually something we've seen in every other technology trend when it started to get some meaningful traction.

In the first generation of the internet, when you started to see e-commerce and all these other business models. Then, in the second big wave of the internet, around the move to mobile and the move to social. And now we're seeing this big wave around AI.

I think what happens is, in step 1, entrepreneurs are A/B testing what it takes to raise money. I think what has happened is that at least some parts of the AI ecosystem have decided that this crazy, scary doomerism is the best way to raise money. Every now and then they come out and say, “All the jobs will be destroyed.” Prophetic, you know, Dario says that. “This thing is sentient.” And investors are like, “Okay, here's $10 billion. Here's $50 billion. Here's $100 billion.”

But then the second step happens. They get the money, they start to do the training, they start selling, and then the investors are like, “Hey, where's the revenue?” So then they start selling everywhere.

Then, if you see the Department of War example, all of a sudden you flip-flop. You become an unserious, dilettante-like partner to the American government. They're like, “We're going to boot you out.” That's billions in revenue gone. And what happens? The same investors that gave billions of dollars are like, “Hey, hold on a second. That's absolutely not allowed. You need to conform and get back on track.”

So what does Dario do? He flip-flops, and he's like, “Oh, I'm really sorry. I didn't mean to. Let's make good.” All of that, to me, is an industry that's still in its very early phases and still figuring out what its place in society is.

So, what is the problem? The problem is the following 2 clips. I'll just have Nick play these, and I'd love your guys' reaction.

Alex Karp

The one thing, though, that I think even now is underestimated by all actors in the industry, including in Silicon Valley, is how disruptive these technologies are.

If you are going to disrupt the economic, and therefore political, power significantly of one party's base—highly educated, often female voters who vote mostly Democrat—and military and working-class people who do not feel supported, and you believe that that's going to work out politically, you're in an insane asylum. You cannot have it both ways.

This technology disrupts humanities-trained, largely Democratic voters and makes their economic power less, while increasing the economic power of vocationally trained, working-class, often male voters. These disruptions are going to disrupt every aspect of our society.

To make this work, we have to come to an agreement about what it is we're going to do with the technology, how we're going to explain to people who are likely to have less good and less interesting jobs, from their perspective, and how is it that we are going—

And by the way, on the military thing, these technologies are dangerous societally. The only justification you could possibly have would be that if we don't do it, our adversaries will do it, and we will be subject to their rule of law.

So, if you decouple this from the support of the military, you're going to have an enormous problem explaining to the American people why we're absorbing the risk of disrupting the very fabric of our society, including the most powerful parts of our society, if it's not because it's about maintaining our ability to be American in the near term and the long term.

Now watch Sam's reaction: “Fundamentally, our business, and I think the business of every other model provider, is going to look like selling tokens. But we see a future where intelligence is a utility, like electricity or water, and people buy it from us on a meter and use it for whatever they want to use it for.”

So, if you take those 3 messaging veins on a spectrum, one is: We have a sentient supergod; we're the only ones that can protect you from it, but your days are numbered. That's Dario. Alex's is: Hey, hold on a second, you can't have it both ways. You can't both say it on the one hand and then try to run the fabric of society and flip it. You need to be much more circumspect. And then Sam's, which is: We want to sell tokens as a service.

I think the point is that this industry right now—that revenue traction, if anything, has distracted people from actually getting on the same page and being much more methodical, reliable, and trustworthy in explaining all of this and managing the expansion of it. All of this fundraising gobbledygook has actually created this breathlessness that is not useful and isn't helping. There needs to be a lot more seriousness by these folks to actually run this business thoughtfully.

You can't be a dilettante, you can't flip-flop, and you can't pressure-test or A/B-test this kind of messaging in public. But I understand why you're doing it, because the stakes are so high. You're playing this enormous poker game. I think we need to do a better job of explaining all this to people, because right now my end of this is: Look at this chart. This is now the result of those 3 messages. Here is where AI is: It is slightly above the Democratic Party and an autocratic state. That's where AI is. ICE is more popular than AI.

Very popular.

So, to me, this is really the crux of this: We are not really being honest. It would be much better if we said, soberly, there's a lot of experimenting, this revenue is great, but we don't really know what's real. We're going to try to figure it out, we're going to work methodically, and there are a lot of regulated industries we're going to work within. We're not going to flout the law and the rules. Licensure will still mean something. That's a way better, thoughtful, mature message.

Brad, what do you think? Great rant. Does the industry have a PR problem? Obviously, these recent surveys, and especially comparing them to China, show that people see AI as abundance and this incredible new technology they want to embrace. Here, people are scared. They're scared they're going to lose their jobs, they're scared about wealth disparity, and they're scared that the rich will get richer while the poor get poorer. There's a lot of fear here in the United States.

What can our industry do to turn this around in terms of communication from the big companies? They don't seem to be communicating in any coordinated fashion, and they obviously are scaring the hell out of the public.

I think it's a fair rant and a fair point. At the start of the Industrial Revolution, in the late 1800s, we had similar social responses to innovations that were occurring. We had some violent clashes, demonstrations in the street, and the entire robber baron movement. Class warfare and worse have come with other revolutionary industrial changes of this magnitude.

It doesn't surprise me that we have a lot of anxiety from people who may lose their jobs. I think there are people out there forecasting into the future in ways that are scary to the average person who's listening to this, and I don't think that's particularly helpful. Could we do a better job of messaging? No doubt about it.

But if I just rewind to where we started, I actually think this is going to be a pivotal year for the industry to demonstrate how this is really beneficial for humanity. I think we're going to be able to demonstrate that it's very beneficial from a health care perspective, from a drug-discovery perspective, from an education perspective, et cetera. But we need to have a coordinated effort because, Chamath, it's deeply unpopular in the country.

I happen to be on the optimistic side of this. Seventy percent of the jobs that exist in the United States today did not exist 40 years ago. We've gone through the digital disruption that put a lot of people out of work, but the abundance and the creation of new jobs expanded the pie for largely everyone. I think that will be the case here.

If you listen to Dario, he says the concern is that the disruption occurs at a faster and bigger rate, and that we can't keep up with that replacement. I think that's another fine point. But if we just go back to where we started the conversation, which was, are these good investments?

That's not the conversation. No, of course they're good investments. Of course you're going to make money. That's not what it's about. He asked the question. You made the argument. Jason asked the question: Are these companies simply selling tokens at a loss? And we moved into the—

No, no, no. They're selling at a profit. I'm buying them and losing money. In the 1849 Gold Rush, Anthropic, OpenAI, and all of these model makers are selling the pick and shovel in the Gold Rush. I am buying it and I'm trying to pan for gold. But as with the Gold Rush, most of these companies will go out of business.

All I'm saying is, if we are really circumspect and honest, there is still way more to figure out than has been figured out. This is not a solved problem, and I think it would behoove everybody to just tell the truth about this. It would be way better to be honest: This is not figured out.

I would say the data, the cards that are being turned over on the table, move me in the exact opposite direction.

Okay, let me get Sacks involved, and then I'll give my take. Sacks, do you have any thoughts here?

I have some thoughts on the question you asked about whether the industry is doing a good job with PR. I think the answer is clearly no. I think they are scaring the bejesus out of the public, and that's why the popularity is right down there with—I don't know, what was it?—Iran. I think we're a little more popular than Iran, but look at Iran: Iran's had 47 years to screw it up. We've only had 2.

It is very much a U.S.-specific problem. If you look at sentiment data across countries, what you'll see is that other countries are much more optimistic about AI than the U.S. I think Stanford did a study on AI optimism. They simply asked the question: Do you think AI is going to be more beneficial than harmful? Something like 80% of people in China said yes. In the U.S., it was in the 30s, and it might be even lower now.

It's not just China and the U.S. Across Asian countries, people tend to be pretty optimistic. The U.S. and Western Europe tend to be pretty pessimistic about it. I think that's less about the reality of AI and more about our media environment and who influences it.

Obviously, Hollywood has created a lot of dystopian films about AI. You've also got the fact that, as we talked about, these CEOs are doing a horrible job and keep talking about putting everyone out of business. I don't think this has been accidental. I would say some of these CEOs are speaking this way because they're not very good at comms. I think others are actually doing it because they see a strategy there. They're going for a regulatory-capture agenda.

That's a good point, Sacks. It's delusions of grandeur, plus they're positioning their companies.

Yeah. It could be for financing reasons, like you've mentioned, Chamath. They want to tout this stuff for fundraising. But also, I think that some of it is to create a regulatory backlash that they can then control—create a licensing scheme or permissioning scheme. That's a big part of it, too.

Then I think you just have the fact that, in our media environment, the scare stories are the ones that get a lot more attention than the heartwarming stories. If it bleeds, it leads. You can just see how unpopular it is for all of these reasons.

New York is about to outlaw medical and legal advice from AI chatbots, which, by the way, is probably the most obviously valuable and highest-ROI thing for a consumer. It hurts the poorest people the most. But you understand, if you're a professional association that sees it as your job to protect your industry from disruption, you might actually want to spread FUD about AI in order to seek those protections through your state legislature.

Well, because if you have companies that are fanning those flames, and those companies are the ones in the industry, it's making your job even easier.

But just think about the poorest person. They can't afford a lawyer, and they want to do their own research. They research the legal stuff in order to fight an eviction. There are poor people who don't have a primary-care doctor, they're not insured, and they find a way to deal with some medical issue they're having. This is the greatest thing.

It levels the playing field for people without resources. So, this is the craziest, stupidest legislation ever.

I give them the sky.

Sky of the week. No, the problem is very specifically that the people who rely on these models to make a health care diagnosis or get a legal opinion to help improve their lives are being told by the makers of those tools that they're about to bring death and destruction upon the economy and the world. So, then the lawyers and the doctors are like, “Well, then maybe we should slow this down,” and they tell their lobbyists, who then go to New York and tell the New York legislators, “Hey, these guys are trying to wreak havoc.” And then they're like, “Oh, yeah, well, then maybe we should shut it down.” That is the loop that's happening. Incredible.

That was the Bernie Sanders moment this week, when he said, “We ought to have a moratorium on all data centers being built in the United States because AI is dangerous.” That was his message, and that is what he's pushing.

And actually, that brings me to another point. If you look closely at Bernie Sanders' messaging, one of the talking points he used was literally verbatim from the Future of Life Institute. I think it was something about how AI is less regulated than a sandwich shop, which is just not true.

Future of Life Institute is one of these EA-funded doomer think tanks, and they've got something like $1 billion. Vitalik Buterin donated $600 million of dog coins to—anyway, you've got this really weird quirk about our media environment: We have these EA-funded think tanks with literally billions of dollars. You know, it's guys like Dustin Moskovitz. They're the DSA, the Democratic Socialists of America. It's so weird that New York is now taking the crown of most stupid state from California. I don't know how this happened, but they just seem to be making every mistake possible.

But—

Let me just finish my point. You've got these, let's call them doomer think tanks, funded by these EA billionaires. They have literally billions of dollars. You can influence a lot of public discourse with that. A lot.

They are behind a lot of the NIMBY stuff around data centers. They've been spreading a lot of the FUD around data centers increasing your electricity prices, which, again, I think is a solved problem now because the AI companies have agreed to pay for the incremental costs and stand up their own power generation. They've been spreading a lot of stories about water usage, which is just totally made up. The modern data centers recycle their water, so they don't use up water.

But again, you've got these doomer groups who are just trying to stop AI however they can. They're extremely well funded, and they're having a big impact. I think this is actually one of the reasons why you're seeing, in the U.S. again, that AI is basically the least popular thing they can poll for, except for the Democratic Party and Iran.

And by the way, FLI, the Future of Life Institute, also funds journalism fellowships and endowments at publications. So, who probably writes negatively about AI?

Yeah, exactly. Here's Sacks. Look at this chart. It goes back a little earlier than 2023, but I had the data accurately from 2023, so we're going into the 4th year. About 40% of all protested data centers in America get canceled.

In 2023, this was a nonissue. There were a few data centers that were protested, and a few of those—40% of them, I think it was literally 1 or 2—got canceled. But then starting in 2024, when you have this divergence of messaging, or this chaotic, slipshod messaging, and it was just a fever pitch to raise money, what you started to see was this fomenting of negative perspectives by individual people on the ground.

In 2024, about 40% of all protested data centers were canceled. It was still a small number; you could ignore it. But last year was when the bottom fell out. We had about 25 data centers canceled, about 5 gigawatts that got canceled. If you use Sarah Friar's number, that's $50 billion a year of revenue that's off the table because of what happened in 2025.

That has implications for everybody. Look at the amount of taxes that would actually raise for federal, local, and state governments. All gone. Vanished.

In 2026, just at the end of February, there were about 100 data centers being protested, which, if you flow that through, will mean about 40 will get canceled. That number right now is about 7 gigawatts, so another $70 billion a year in revenue.

So, just last year and this year, we've taken $120 billion of revenue per year off the table. This is a wake-up call that this messaging is wrong. These people are not doing what is right on behalf of a very nascent and critical industry for America.

There's only so much that Sacks can do, the White House can do. All these other people are kind of at the periphery, but if the people on the ground don't get their shit together, this is a national disaster.

Just to give people a sense of where this is happening, almost all of these cancellations are in Virginia and Indiana, according to some cursory research here. And there have been zero cancellations due to local opposition here in the great state of Texas, where we have over 150 gigawatts of data capacity requests.

So, if you want to do this, come to Texas, talk to Abbott, talk to Ted Cruz. Just CC them on your tweet, and they'll have you to the poker game, and they will greenlight it.

Brad, before we move on, I want to get your opinion on open source, how powerful it is, and how powerful Apple silicon is getting. I'm not sure how you factor this in at Altimeter into your thinking, but this seems to me to be a massive headwind against the 2 big bets you have.

All of these open-source models—we've started running them and are picking up about 85% of our tokens right now. Every startup I know is saying, “We are standing up our local models, and we only use the top models, the paid ones, when we have jobs we can't do.” So, I'm just curious about your thoughts on that, and you can add to that the AutoResearch project from Karpathy that came out this weekend.

For people who don't know, we now have a group of tinkerers who are setting up their OpenClaws, setting up large language models, and trying to train them with this AutoResearch tool. This seems like a parallel track that could be material. I'm just curious if you're monitoring it at all.

I mean, first I would say that I am very enthusiastic about open source. We see it in widespread use everywhere.

But here's the interesting thing: For the advanced companies, they're doing some planning with the frontier labs, and then they're doing the execution, if you will, with the open-source models. So, they're running an ensemble-model strategy.

But here's what I think is more impressive. We have incredible open-source models nearly on the frontier, and notwithstanding that, we're seeing companies like Anthropic add $5 billion or $6 billion of revenue in a single month, which is extraordinary. We've never seen anything like it in technology.

And that's in the face, Jason, of all this open-source usage. So, what does it tell me? It tells me that the TAM is dramatically bigger than any of us think that it is. When we look back on this period, that will be the big takeaway.

It's the takeaway with Uber, the takeaway with Google, the takeaway with Amazon: The TAM was way bigger. We've crossed an important threshold. Open source will be a part of it, but clearly the frontier labs can do well even in the face of it.

All right, a little housekeeping. The All-In Summit is coming and liquidity is, I think, sold out or about to sell out. We might have a couple tickets left. You can find both events at allin.com/events. And All-In Summit tickets if you want to get there quickly before they sell out. September 13th, 14th, and 15th in Los Angeles again. And for all the All-In listeners, we're launching a survey today. This is super important to take the All-In survey if you made it to this point in the podcast. If you are an early true believer in the pod, if you're one of the All-In stands, we need you to fill out this survey. allin.com/survey.

All right, let's wrap up with this final story that went viral. The millionaire tax has hit Washington State. Howard Schultz, the longtime CEO of Starbucks, has bailed and gone to Miami—Surfside. He bought a condo in Surfside. He pulled a J-Cal.

I think you mean a sexy boo?

Well, yeah, but I was never a Starbucks liberal before I left the state of California.

Listen, I don't know how many times I have to make this correction. I am a moderate. I literally voted 4 elections in a row for Republicans. People have asked me for the receipts: Pataki, Giuliani, and Bloomberg. I literally, for almost a—

Okay, okay. You voted exclusively Republican. Washington's millionaire tax passed this week. Here's what the tax is: People making more than $1 million a year will pay an extra 9.9% in tax starting in 2029.

The Budget Center estimates the tax will impact 30,000 households, bringing another $4 billion to the state's general fund. The funds are supposed to go toward public schools, higher education, and health care.

In a huge coincidence, on the same day the new tax was passed, Howard Schultz, the billionaire Starbucks founder—

Huge coincidence, did you say?

Yeah. Just unrelated stories.

Unrelated news.

He will be leaving Seattle after a 44-year run because he found out about these incredible Cuban sandwiches.

There was an opportunity to buy a $44 million condo in Surfside.

He couldn't pass it up. It just happened to be on the same day that they passed a millionaire tax. He had the Cuban sandwich at Le Sandwich, and he fell in love. Schultz has been getting crushed after saying, when he ran for president, that he would be willing to pay more taxes.

Bezos obviously left back in November 2023, and people speculated that maybe the 7% capital-gains tax would have influenced that. Who knows? So, I guess, Chamath, what is the end game here? These local politicians must have learned the lesson that people of means can move. They have the ability to buy new homes and put their old homes on the market. They're very mobile, and they could even leave the United States and go to Singapore, Dubai, or other locations in the world. Why are they still enacting these taxes, and will they continue to enact them until we get to 60% or 70% tax rates, lose all of the creators, and this becomes Ayn Randian?

I think that state politicians on the West Coast are very ineffective and not very smart. Jason, there's a tweet that was published—I think maybe it was an infographic—that showed the net migration rates of every single state for 2025. Washington is a few months behind California in trying to enact these stupid taxes, and the reason they're stupid is that these kinds of things don't work at the state level.

We know what it's already done in California because the Hoover Institution just published something this morning, and it's a complete indictment of what the billionaire tax was trying to do. By the way, this billionaire tax is only polling at 25% of the votes it needs. Maybe it'll find a way to get on the ballot, and even then it will have an uphill climb to get voted in.

But look at the destruction that it has done in California just by announcing it. The Hoover Institution basically ran this Monte Carlo simulation. They ran 100,000 simulations, and in 71% of those runs, it comes out with a negative NPV. If you expected-value it out, it's about a $25 billion hole.

They also found that they overcounted the number of billionaires in California, so that number was wrong. They undercounted the amount of revenue that they pay, so that was wrong. And they overcounted the estimate of how much money they would make, so they're not good at math. They're not good at math.

When you add it all up, they thought they were going to make $100 billion, but they're actually going to make $40 billion. The people who left pay $3 billion to $5 billion a year in taxes. It's going to create a $25 billion hole, and the middle class is now going to have to foot this because this is net revenue that's not going to come into the budget. That's about $2,500 per middle-class household. There are about 10 million households in California, so that's what's happened just by making the threat.

Washington had a 23-hour debate and passed the law, so I suspect when you look back on this in 18 or 24 months, it'll be as bad as or worse than California. These things don't make sense. The reason they don't make sense is that you are putting good money after bad.

We all know that money that goes to state governments is wasted; we just don't know how much. When you keep asking for more, eventually the smart people say, “Enough's enough. I'm out of here.”

We might find out how much. I think Bari Weiss is on the case. I don't know if you saw her CBS report this week, but she's going hard after fraud. Until you get fraud out of the system, I don't think you have the moral high ground to raise taxes. I think that should be the message.

Well, that should be your campaign promise when you run.

Okay. Hi, I'm Jason Calacanis, and I will get rid of fraud and lower your taxes.

Look, you may have seen an even more severe tax proposed at the federal level, where Bernie Sanders and, I think, Ro Khanna came out with their version of a national wealth tax. It wasn't just 5% once, like in California; it was 5% per year. In roughly 20 years, the federal government is going to take all of your money. That's it.

Look, this is socialism. This is another way to get to the same endpoint, which is that the government owns everything. They seize it.

Well, I mean, the seizure part of it is the nuanced point we have to get across. If you earned it and paid your taxes already, it's yours.

The government can't just decide, “You know what? We didn't take enough 10 years ago. We need to go seize that.” When you sold your hammer—

Right. When you sold it, we didn't take enough. We need to take it now.

Raise your hand if you believe the things you own are better off being owned by Bernie Sanders and Ro Khanna. Raise your hand if that's what you believe. You'd have to be an idiot to believe that.

Ro Khanna, I'll tell you: the last time we saw these proposals for asset seizures was during the Gilded Age, from 1870 to 1920. Then it was Carnegie and Rockefeller. What's interesting is that I went back and looked at it, and there was actually warfare. We had hundreds of people killed in clashes during the Great Railroad Strike, the Pullman Strike, and so on. It was all over this.

Shame on the politicians who are fanning the flames of class warfare. We all need to bring the temperature down. There are fair debates about whether states have enough resources to fulfill their obligations to their citizens. There are fair debates about fraud. All of this has to be taken on.

I think it's interesting that in the state of California, the teachers' union is against the billionaires tax because they know it's going to lead to fewer dollars for education in the state of California. Matt Mahan, who's running for governor, is against the tax. The current sitting governor, a Democrat, is against the tax. They all need to step up and explain, not just that they're against the tax, but you're either on the side of business and entrepreneurs and creativity and moving the state forward and growing the economy, or you're against it. And that's what's at stake here. And fortunately, the outcome of the battle during the Gilded Age was that America didn't abandon entrepreneurialism. We didn't abandon capitalism like Europe did. And now we leaned

One point on that. You mentioned that some of the unions in California are opposed to this asset-seizure tax. That's only because they weren't cut in on it.

Right. Because there are already rumors that the California Teachers Association is working on its own version.

Oh, boy.

They're working on a billionaire asset-seizure tax—not for this election cycle, but for the next one.

I think, unfortunately, that's going to be corrected. If this one doesn't pass, they'll correct it for 2028, and it's more likely to pass because they're all going to do it. The other thing is that by 2028, this national wealth tax will just be a standard part of the Democratic platform.

It could be table stakes, yeah.

Yeah, it's table stakes. I think the Bernie Sanders position will become the position of the Democratic Party. You even see Gavin Newsom creating wiggle room for himself to embrace this position.

If you look closely at his statements formally opposing the BTA in California, what he says is that a state can't do this by itself because it has 49 other states to compete with. So, what he's saying is that California is operating in a competitive environment. One state can't do it, and he's leaving the other part elliptical: the federal government needs to do this. I think you can expect him to embrace that position by 2028.

There is a way out here from this socialist movement. It's very simple. If you think from first principles, what does an American want? What does an American family want? What do mothers and fathers in this country want? They want to educate their kids. They want to be able to own a nice home. They want to have decent health care. They want to have healthy food. It's a very small subset of issues.

AI is uniquely positioned to solve a lot of these problems, and entrepreneurs can come in and take on these highly regulated industries if we're allowed to participate in them. In education, we have to break this accreditation cartel. In housing, we have to break regulations, like the great states of Texas, Nevada, and Florida have. And when it comes to health care, AI could have a tremendous impact. Entrepreneurs could have a tremendous impact in lowering the cost of health care and letting people solve for that with their own self-led health care.

These are the problems. If you solve for people's homes, people's health, and people's education of their kids, we're going to solve these problems, and we don't need to go to socialism and seize people's assets. That's what entrepreneurs should be doing. That's what entrepreneurs should be working on.

And that's where the government can help. That's where Trump is uniquely qualified. He is the regulatory breaker. He got nuclear back on the agenda. No other president had gotten nuclear back on the agenda for America. He can get housing back on the agenda. We have to break those things, not start foreign wars, to my earlier point, and start creating houses for Americans.

If you would like, you can moderate and I'll go first, but I always go last. You can give me a round of applause, or you can throw a tomato. Go ahead, Sacks.

It's all good. Go for it.

Thank you, Bestie Brad, for joining us. Thank you, Bestie Brad.

Love you, boys. Bye-bye.

Rain Man, David Sacks. I said we open-sourced it to the fans, and they've just gone crazy with it. Love you, Bestie. Queen of Kinwah.

And is my dog taking a dump in his driveway, Sacks?

Oh, man.

If that guy should meet me at the party, we should all just get a room and have one big, huge orgy, because they're all just useless. It's like this sexual tension that they just need to release somehow.

What? Your big feet? We need to get merch.