Mar 20, 2026 ยท 2026 #9. Read the transcript grouped by speaker, inspect word-level timecodes, and optionally turn subtitles on for direct video playback
Edit labels for this show, save them in this browser, or download a JSON override for the production folder.
Transcript Playback
Public Markets Price Outcomes
Human Transcript
Timed transcript
Blocks are grouped by speaker for readability. Expand a block to inspect word-level timing.
Speaker 2
Hello, everybody. It is Friday, March the 20th, 2026. Another week in Silicon Valley. More madness, more investment, more obsession with AI. This week, Jeff Bezos announced His intention to raise a hundred billion dollar fund to transform companies with AI. These aren't naturally AI companies. Bezos himself looks increasingly, at least in some of the pictures that appear in media, like a Hollywood star. Meanwhile, Our obsession with gambling is picking up. This week, Major League Baseball named Polymarket, exclusive prediction market partner. And Polymarket's also in the news. It's opening a bar in Washington, D.C. And of course, it's been in the news recently because anonymous bettors on Polymarket cashed in on the Iran strike hours before it happened. So AI and unreason. Our obsession with betting seems to be defining the world. But for Keith Teer, the publisher of That was the week newsletter. The market's perhaps a little more rational than some of us might think. His editorial this week is on public markets and price outcomes. It's a bit of a boring title, Keith, but it's an interesting thesis. You're suggesting that The markets are actually pretty rational, especially when it comes to these private AI companies. Of course, the expected IPOs of Anthropic and OpenAI later this year are expected to be the supposed Netscape moment for our AI economy.
Yeah, look, the point of that title, you're right, it is boring, but public markets are prediction markets. And they're driven by sentiment. And the sentiment is driven by a belief in what's going to happen next. So what's going to happen next sets the price of any asset. It's interesting, over the last week, NVIDIA had its big GTC conference and predicted that it will do a trillion dollars in new revenue between now and the end of next year. Markets initially put the stock down because they'd already said they'd do 500 billion this year. Then the CEO corrected himself and said, no, this is a trillion of new revenue between now and the end of next year. And the market's putting back up again. So markets are really weather vanes on the future.
It's a good prism through which to look at all these different things going on because indiscipline doesn't get rewarded by markets. And markets are where you get your paycheck. So almost every player, including governments, keep an eye on the market for good reason.
Yeah, and I'm still waiting for Donald Trump's Liz Truss moment when the markets are responding there. It seems to be we're on the verge of a Liz Truss moment when it comes to Trump in the Middle East. But when it comes to the markets, Keith, and this idea of the public markets not... not making its decision yet on ai i wonder if they have i mean i was looking at google stock and if you look at their numbers over the last year there's nothing that dramatic they've been climbing steadily from 150 now to about 300 so hasn't haven't the markets already determined the value of ai if they're if they're bullish or relatively bullish on google why wouldn't they be equally bullish on open ai and anthropic
Well, the markets always have a current opinion. So you're right about that. And they also have a view of the future. And if you look at the three IPOs we mentioned, should they happen at the scale that's being rumored, those three companies would raise more money going public than the entire IPO market for the last 10 years. So the market is about to deploy cash at a scale that's unheard of into AI. So even though the markets have already spoken, they haven't finished speaking.
The way you present the markets, though, in your editorial, you seem to be suggesting that the public markets and the people who would buy shares of Antropic and... OpenAI and perhaps Google are entirely rational, but aren't they slightly crazy? I mean, what about all the irrational exuberance of previous booms and busts from tulip mania to dot com?
Yeah, well, look, the editorial was triggered by an IPO that happened yesterday for a fund that's somewhat similar to SignalRank called VCX. It's called Fundrise, and the ticker is VCX. Its shares started trading at $34 and ended the day at 100 and something. Today it's at 104. So it's a 300% gain, which is also, because it's a fund,
300% valuation above the net asset value of what it holds. So yes, retail investors are paying three times what the shares are worth. And by the way, guess what the shares are? Anthropic, OpenAI, SpaceX, Databricks. So VCX is actually... And Duril as well, which is... And Duril. So it's a proxy... for what the market thinks these companies are really worth. Think of that as future asset value. Now, if you've got this window of current asset value and future asset value, and the market is prepared to pay something of the future asset value today, then these assets are going to trade at a premium like VCX is.
Yeah, but you're talking to me like you're some sort of algorithm, Keith. I mean, we all know that There's so much irrationality when it comes to these sorts of things.
Well, do you think this is a... That's a good question. Do you think it's irrational that people are playing three times the real value for that BCAX? Or is it rational? And the answer already lies in, well, what will it be worth in future?
Yeah, but... I take your point on Fundrise, but there aren't that many retail cab drivers or day traders. They're not buying Fundrise. They're waiting for the OpenAI and the Anthropik IPOs. Those are going to determine. Those are going to be the Netscape moment, for better or worse, rather than Fundrise. How much money has Fundrise, by the way, raised?
Nearly a billion dollars. It's big. And by the way, it's worth double clicking a little bit. It has a long history of selling private market shares to smaller investors through a vehicle that's been around for a few years, Fundrise. It has roughly 100,000 investors on its platform prior to going public. And so... It's a middleman. It's a kind of... No, no, no, it's not. It's an active investor directly in the companies. It does invest in the companies. And it is taxi drivers. It's quite a... Or now, these days, Uber drivers, I would hope. Or Uber drivers. So actually, it's a bit of an unusual animal. It's not a big bank, like institutional. It is a bottoms-up kind of trading platform.
Well, let's... get to your take on these imminent IPOs, Anthropic, OpenAI, SpaceX. You talk in your editorial about the key word being trust, just suggesting that whether it's the formal markets or the Uber drivers, they don't currently trust OpenAI and SpaceX and Anthropic. Is that right?
Well, that was last week's discussion where we acknowledged that both both, we called it love and hate last week, but you could call it distrust and trust. Many of the things have common roots. I think the people buying VCX definitely do trust, and the people unaware of it maybe just are in the middle and then the people who are explicitly not buying it or shorting it distrust and and the current state of play is that all those opinions exist alongside of each other so ultimately we have to take you know our finger on the pulse of the future and ask what's going to happen next and i think i think my view is that what's going to happen next and i'll stick to this is that open ai currently valued just under one trillion dollars is going to end up being valued around 10. And so... Right, in the long run now.
What do you mean? Andrew. What do you want me to do? Go to bed with you? I mean, $10 trillion. Are you trying to punish me or reward you? Yeah, I think I'm trying to... Beat you up. I'll spank you in the bedroom. I mean, you can say anything you want. You can say 100 trillion, but what's the reasoning for that?
Well, look, here's the thing. Why wouldn't it be, is a good question. If NVIDIA truly does a trillion dollars in new revenue, by the way, Anthropic did 5 billion in new revenue in one month. 5 billion new revenue in one month. These things are pulling in revenue like there's no tomorrow.
Well, it kind of does last forever. Why not? I mean, if you build something substantive that's enduring because it actually is useful, it actually might be forever.
Well, you may think it's 10 trillion. I mean, I'm not sure where it's... worth having a serious discussion on this. Let's talk about your take on Anthropic. You come back to the issue of Anthropic and its fight with the government, that that story is still in the news. Yeah. My sense is you're not really 100% convinced on your own position. Is that fair?
Well, your position is that Anthropic is a corporation and it shouldn't interfere in the business of government. And if the government wants to buy its product, then it shouldn't tell it what it can and can't do with it.
Well, I'm not giving Anthropic advice. I'm pointing out that it's going to get punished for it by markets as well. Ultimately, because... Although it hasn't yet. In fact, it's been rewarded. It's rewarded because it's living in a bubble of people who agree, you know, and think in the same way. And whilst it's private, that can be sustainable. As soon as it's in public markets, that will not be sustainable. You look at Anderil, which is the opposite of Anthropic, and Anderil is already public and public markets are pricing it. Anderil landed a $20 billion contract with the US military this week. and it's trading at multiples, forward-looking multiples that you might say are crazy.
Although, I mean, to compare Andrile and Anthropic might not be really fair on Anthropic. I mean, Andrile is a military tech company, so it has to trade with the government, whereas Anthropic conceivably could become a multi-trillion dollar company without doing business with the US government.
Yeah, I mean, that's why I said depends what opinion. I'm a user of Anthropic. Claude Code, Claude Cowork, and Claude itself are in my daily work every single day for large parts of the day. And I'm a big fan of what they built, a huge fan of what they built. By the way, I don't really disagree with Amadai about AI, weaponized AI. either. So most of my points are really about how you navigate these differences, not whether you have them. And, you know, I don't think he's doing his company any favours by his choices. That's really all I'm saying.
Although, as I said, I mean, there was a piece this week on CNN about how Anthropik may benefit from his fight with Trump. I mean, I don't think I don't think Amidai is that crafty in suggesting that, thinking that if he stood up to Trump, it would benefit his company.
Well, look, it's popular to have his point of view. There's no doubt about it. And, you know, I've been there in the past where I've taken points of view in my business. I remember when the DOJ sued Microsoft for being Monopoly, Microsoft was my partner and I took the opposite view that Microsoft was about to be disrupted by Google and the internet and the DAJ didn't need to do anything. And it was an unpopular view, so I did the opposite. But It is common that CEOs have points of view, and he's got a point of view, it's fine.
You're in Palo Alto, so you know lots of VCs and entrepreneurs. There was a piece this week in the New York Times about Silicon Valley mustering behind-the-scenes support for Anthropic. Do you think that there's some degree of unspoken support for what Amidai is doing with Anthropic in Silicon Valley itself?
Well, that is, I don't know if it says it explicitly, but actually that's OpenAI lobbying on Anthropic's side behind the scenes. It does not want Anthropic to be deemed a risk. And if you ask why, it really comes to a much bigger discussion about the relationship between AI tech companies and nation states. And it does seem clear that even though AI is not a weapon in the way that Anderil is, AI is inevitably going to play a role in weaponry. And there's really no way of avoiding the fact that in a world of conflict with nation states fighting each other, they are going to use AI to conduct business.
I mean, that goes without saying. But I wonder, and you and I have talked about this before, whether the old equation of governments somehow being more powerful than corporations and corporations being in the business of supporting government, whether that's going to be turned on its head, that increasingly these... I mean, if you're right, and as I said, I'm a little bit skeptical, but who knows? I mean, in 10 years, anything can happen. If OpenAI is really worth $10 trillion in 10 years' time, I mean, the current largest company, most valuable company in the world is, what, I don't know, is it Apple? It's about four or five?
By the way, we do have a commenter from YouTube who says, given that 20% of the world's capacity of LNG was destroyed and data centers being bombed in the Gulf, is this the popping of the AI bubble? And the answer, profit generation lab is no. It is not. I don't even think it's a bubble. But it definitely isn't going to pop. In fact, the opposite is, it's going to double at least year over year for the next five years. By the way, that's how you get from 1 trillion to 10, Andrew, is one becomes two, two becomes four, four becomes eight.
Yeah, well, as I said, I'm not going to enter in. To me, it's a fantasy. I mean, you can go and watch a Disney movie if you want, but when it comes to valuing companies in the future, especially going, I mean, OpenAI isn't even worth a trillion. It's not even a public company yet. But if it is indeed, if we do have an AI company, whether it might be Google, it's already at four or five billion trillion, or open AI, or even Antropic, worth $10 trillion in five, 10 years. How will that change the relationship between the corporation and the government?
Well, look, US GDP is about, I think it's 120 trillion worldwide, of which the US is if I'm not mistaken, the largest... What is it, about 40, 50 trillion? You know, I'm blanking on... What's a trillion between friends? Let me just not throw out a number out of my head. U.S. GDP right now is 31 trillion.
Okay. So we're talking, so if what you're saying is right, And it's beyond speculation, as I've suggested, but if open AI or one of these AI companies is worth 10 trillion, it means if an open AI is worth 10 trillion, then Anthropic's probably worth eight and Google's probably worth 15. They basically make up the entire economy.
So the stock market is like one year's GDP for the world. And by the way, that's global stock markets. The US is about half of that. And so what you're saying is that OpenAI could be itself about 15% of the US stock market's value today. Of course, the stock market's value will grow. Now, in order to be worth $10 trillion, its revenues would have to be something in the order of $10 trillion divided by 20. So that's about half a trillion. $500 billion.
You're not answering my question, or you're not maybe understanding my question. How does that alter... the power dynamic between the government and corporations. I mean, at the moment, the government's trying to push Anthropic around and Dario Amidai is pushing back with some degree of success. It would seem, I mean, he's supported by some people in Silicon Valley. It certainly benefited Claude users. One wonders in the long term what the impact will be. But if these trends continue and these AI companies grow in their valuation, what does that mean in terms of their relations with the government?
Well, so let's ask the question about NVIDIA, which is the world's biggest company. How does it impact the government? The answer is not very much. The government could stop it selling its chips to China. because ultimately the law trumps the company. And so what you hope is that lawmakers are immune to being bribed. We all know that lawmakers do get, in quotes, bribed through lobbying groups and so on. So far, at least, most of the AI companies have shown limited interest in lobbying. They do do lobbying, but it's not a lot of money, relatively speaking, that they spend on it. So I think that the answer to your question is really a question more about government than it is about the AI companies. The AI companies only get power if you have corrupt governments where the AI companies want to leverage that corruption to their own self-interest. I don't think of it in that framing. I think AI is a human good. And I think governments want it, not all governments, but most governments, especially the US government, want it to succeed.
We shall see, certainly when it comes, if there is some war or crisis over Taiwan and NVIDIA and its role in all this. It will be interesting. Another piece that you linked to, which maybe speaks more to my division this week, is your friend Om Malik. He writes, and I think it's your first piece in your interesting essays of the week, is he writes on neo-symbolic capitalism, and he quotes Patrick Collison, the CEO of Stripe. I think our difference this week, Keith, and I'm in the OM camp here, is you view capitalism still in a rigorously rational way, in mechanical, arithmetic terms, and the market simply responds to the reality of the world, whereas OM and myself see things in a much more I mean, Ohm uses the word symbolic. I might even say irrational. The behavior of these peoples, whether it's Musk or Bezos or certainly Dario Amidai, isn't quite as rational as you suggest. Why did you include Ohm's essay on neo-symbolic capitalism? I'm assuming you don't agree with him.
There's elements I agree with. I think that the discussion about the near future and the relationship between capital, labor and government is probably one of the most important discussions that could be had. And he's participating in it. And so I don't think... the criteria for me including an essay as agreeing with it, it's whether it's speaking relevantly to the topic.
And let's just be clear then about, and I'm quoting from Google Gemini on symbolic capitalism and I'm quoting, symbolic capitalism refers to an economic structure where value is derived from producing, managing and consuming symbols such as data, brands, narratives and culture rather than physical goods. Driven by a professional managerial elite, this system emphasizes reputation, prestige and symbolic capital over traditional manufacturing. And a lot of it's borrowed from the French cultural theorist Pierre Bourdieu. So what I think Ohm is arguing is that and whether it's Patrick Collison at Stripe or other entrepreneurs, that they're all driven by adding to their reputations and that the traditional metrics don't really make any sense.
Well, it's interesting because in that framing, Dario Amadai is a representative of what Om is discussing. He's used social media as a platform to build reputation by having a point of view against government. And that's really a very different set of events than his products and the revenue that they're generating. Whereas OpenAI is focused mainly on growth and revenue, and hasn't been on a pulpit trying to get reputation building. So, you know, I think... Sam Altman...
Right. And Ohm, being a media guy, puts this down a lot to the new form of media. Interesting, Andreessen Horowitz had a big thing this week called What is New Media? I can't remember if I put it in the newsletter or not, but it's a good podcast. And Ohm is saying that basically reputations are built by direct messaging to an audience through social media, not by doing a press release. And that this amplifies this symbolic capitalism focus and that most leaders of industry are more and more focused on their reputation and imagery. And he's not wrong, but he also is wrong because none of that works unless you're making money. And think of Musk. Musk is an interesting character study in this context because Musk has, you know, drawn the ire of not just you, but many, many people due to his statements and actions around Doge in particular. Yet he's making so much money through Tesla and SpaceX that it really doesn't matter. And so he's the counter symbolic capitalist in a way.
Well, and just to be clear, and I wouldn't disagree with you there. Ohm says that Musk is the master of symbolic capitalism. He executed a perfect plan for creating a perpetual machine for symbolic capitalism. And the whole valley took notes. I mean, the two aren't necessarily in contradiction. You can have profitable companies and at the same time be a symbolic capitalist like Musk.
Well, I wonder whether they're increasingly becoming the same thing. My interview of the week, which was entitled Is Elon Human, was with Charles Steele, an investor, who's written a book on the curious mind of Elon Musk as an investor. He's much more sympathetic to Musk than I am. But I wonder whether symbolic capitalism has become real capitalism and that Maybe that's why Dario Amadei has been so, quote-unquote, brave in taking on the government, because the great industrialists or post-industrialists of the 21st century will need to take on government. And in a sense, they're stepping into the vacuum where government used to exist. And that comes back to my question of what happens if... an AI company is worth 10 or 15 trillion dollars in five or 10 years. They're going to be able to rival government. And given the lack of trust and faith and credibility in government, one wonders whether the Amidais of the future will become the Prometheuses. It's perhaps no coincidence that Jeff Bezos' new fund is linked to his project Prometheus.
Well, there's a lot in what you just said, Andrew, so let's focus on trends. I think a lot in global terms, and I think it's inevitable that the largest global companies in the future will have larger valuations than the GDP of any country, possibly even bigger than the US's GDP. And why? Because the globe is a lot bigger than any nation. And if you are dominant on the globe for something that people pay for, especially expensive things like AI, you're going to become big. That to me seems inevitable and therefore capitalism ultimately is a numbers game. All these other things are still true. There's layers of economics and then politics and civil society and the way it gets played out in the media. All of those live together, but at the base is economics. and economics is a numbers game. And so the winners win the numbers game and get to own much of the layer above, unless politics is independent and truly democratic. And therein lies, you know, the challenge facing all of us. How do you get a polis that can be trusted with the future? And, you know, I have a point of view about that. It's very different to the current situation where we have wars raging left and right in terms of what I think the future needs to look like. But that is a contested space where opinion will clash with other opinion and result in some kind of an outcome. And in the absence of that, you get uncertainty. And uncertainty kills economics.
Uncertainty is the very definition of economics. I don't know what you want from it and I don't know whether you're expecting these billionaires to be entirely rational. Another of your stories this week is on a TechCrunch story about The billionaires who made a commitment to the Buffett-Gates initiative on the giving pledge of giving their money away. Peter Thiel is leading this. Now they want out of the promise that was made back in 2010. Meanwhile, this week, Thiel, who in an economic sense is clearly a genius and perhaps equally successful, prescient as Elon Musk. He's been in Rome this week talking about the Antichrist, and there was a piece in the New York Times suggesting that some people in the church want to perhaps persecute Thiel himself as the Antichrist. So do we need to get out of this idea that somehow entrepreneurs are rational. I mean, they are in market terms, but maybe the great geniuses are also mad, like Thiel and his obsession with the Antichrist, or Musk and the fact that, for better or worse, he's on the spectrum. Is this one of the issues that maybe the public markets have to acknowledge this? Maybe One reason to be a little dubious of Dario Amidai is he's a bit too rational and reasonable. He needs a bit of the Peter Thiel or the Elon Musk in him.
Look, they all are irrational. That's why they get where they get. I don't know of any rational person that ever became a billionaire running a disruptive company. Contrarianism is at the very core of innovation. And so they all have their irrational side. And I would include Dario. I mean, it is totally irrational to believe you can fight the government and win.
Well, I would say none of those beat the government. The government embraced and extended their message and eventually owned it. They got incorporated. And so in surprise you're successful, like rap music, you get incorporated and become part of capitalism. And, you know, otherwise these people would be leaders in the same way that, you know, these people are not Oliver Cromwell, who went with an army and stole most of Ireland for the English. They didn't steal anything. They built things. And their success probably fuels their lack of rationality. And their egos are bigger than any of our egos. except maybe mine, and that is at the very core of their being. That said, they don't run the country, and they won't run the country.
Well, but this comes back, and you and I have been circling around this. I think in some ways they already do, and they certainly will in the future. Let's end with another great symbolic capitalist, Vinod Khosla. uh his uh post of the week which is of course a platform for symbolic capitalism on x i'm quoting him he said um ai will necessitate a change in tax structures capital gains taxes ordinary income taxes and more
And then he goes on, and this is really interesting. Capitalism is by permission of democracy. But I would actually reverse that. I think democracy is by permission of capitalism. Isn't that the old Marxist framing, Keith? Isn't it true?
Before you get to that question, I mean, Vinod is a classic symbolic capitalist, isn't he? He understands the system as well as Musk or Teal or any of the others.
Correct. I think there's two things to talk about. The first is his framing of capital and labor. I think there's a book in that. I'm almost tempted to write a book called Capital and Labor in the context of what's going to happen next, because that is probably at the very core of what's going to change. And when that changes, he's right that taxation makes no sense as it is thought of today. because income will not exist. And if there's no income, there's no tax. He wants to get rid of capital gains, which is the other side, the capital side. But he wants to find a mechanism to channel the wealth from the success of AI to society, which is a worthy desire. I think it's a great desire.
And you've articulated that. We've all articulated that. I mean, it almost goes without saying. I mean, even the Andreessen's and Teal's and Musk's of the world want that.
Right, so that's one thing. The second thing is, how will it happen? What are the mechanisms through which that could happen? And I only can come up with one answer, which is government. Because any other answer, it would either be philanthropy, like the big winners in AI just agree to donate money to society. That doesn't seem very likely. It's certainly not reliable.
What became of the Marxist in you, Keith? You really do believe that capitalism is by permission of democracy? Marxists would always argue that democracy is by permission of capitalism.
Well, let me address that. I didn't address that, actually. I think democracy, historically, certainly as we understand it today, not slave-owning democracy in Athens, but... where there's a middle class.
Right. When you get a middle class, which is the precondition for which is capitalism, and as the middle class grows and goes to college and is educated, it tends to influence the outcome of elections more and more and more. Even Trump is symptomatic of the middle class influencing an election. And that tells you that democracy thrives on economic success. Now, there are examples where that isn't true, like India is always quoted. although India is about to and is in the middle of economic success now, but usually democracy requires a healthy capitalism.
Yeah, I mean, it's a bit of an old-fashioned idea. It doesn't seem to work in China, for example. Everyone expected Chinese prosperity result in democracy. The reverse is true.
Well, there we have it. Keith believes that capitalism borrowing from Vinod Khosla, the great symbolic capitalism, venture capitalist, symbolic venture capitalism, maybe that should be the title of your book, Keith. Like Kostler, Keith believes that capitalism is by permission of democracy. I believe that democracy is actually by permission of capitalism. And my sense is as our century progresses, and if Keith is indeed right, that the open AIs of the future will be worth 10, 15,
then that will be increasingly the case. Certainly, We will return to this subject, Keith. It's one that we've been circling around and we come back to time and time again. Thank you so much. Very interesting conversation this week.