The AI Gold Rush

Speaker 2

It's Saturday, the 5th of July, 2025. We're doing our weekly tech roundup with my friend, Keith Teer, the publisher of That Was The Week Tech Roundup. newsletter uh this week's uh newsletter is all about what he calls the ai gold rush keith and i are both away in july so we won't be doing a show uh for about another four weeks our next show will be at the end of july so i thought we would might also use this opportunity keith to talk about the first six months of uh 2025 at the beginning of your Interesting editorial this week. You describe what you call the AI gold rush as not just another wave, it's a phase shift. What does a phase shift mean?

Speaker 1

Could mean a lot of things, Andrew, couldn't it? It's one of those vague terms. Well, you're using it. That's why I'm asking. So what I'm capturing in that phrase is the substance of this week's articles in the newsletter, which really describe a bunch of changes underway that are quite seismic in their impact. And the first is the bifurcation between highly valued and funded AI startups and everyone else. The second is the shift away from advertising as a business model towards, let's call it toll booths, where AI has to pay the publishers. And the third is the changing job market, which is huge. And there's probably more, I think, in the editorial. There's eight, actually.

Speaker 2

Yeah, there's eight. And we'll get to those in a second. But are you suggesting then your eighth, Keith? It's always good to get to the end before we start. Your eighth is that 2025 isn't just another year. It's the inflection point. What's the difference between inflection point and a phase shift? Or are they the same thing?

Speaker 1

Is it just... Yeah, a phase shift describes the moment.

Speaker 1

An inflection point describes the likely future. And the moment is a tipping point, if you will. A phase shift is more like a tipping point. An inflection point is a tipping point where you fill in the blanks on what it is tipping into. So...

Speaker 2

So let's introduce maybe another phrase, a simpler phrase. It's a big deal what's happening. You call it in your editorial the AI gold rush, but I think that's a little misleading because when we think of gold rushes, we think of rushing to things that probably don't exist. So in California, for example, with the original gold rush, there wasn't a lot of gold, but we're not talking about that. I mean, this is real, is it, in your mind at least?

Speaker 1

Yeah, I mean, Gold Rush is appropriate this week because of the huge deals Meta is giving to new hires. I mean, massive deals.

Speaker 2

$100 million deals to individuals. We talked about, or someone predicted, I think it may have even been Zuckerberg or Musk, that you would have billion-dollar companies run by individuals. Zuckerberg's making that possible.

Speaker 1

Yeah, and also what OpenAI did with Johnny Ive, which was $6 billion. arguably for a company that was mainly about him.

Speaker 2

Yeah, although, Johnny, I would rather invest $100 million in an AI whiz than Johnny I, if it has been. Wouldn't you?

Speaker 1

I think you're right, Andrew. But you need more than one of them. That's the problem. So it adds up.

Speaker 2

So anyway, so it's beyond the gold rush. It's more than a gold rush. I'm not sure what words we would describe. It's for real, first six months of the year. As you say, a phase shift, inflection point, whatever it is, it's a big deal. You spend a lot of time, Keith, looking at the VC market. Your startup is... reinventing venture capital. You argue in the editorial that venture capital is being forced to play a new game. What is this new game?

Speaker 1

Venture capital really is three different things. So what I'm talking about with the new game is what the later stage investors are having to do, which is the third of the three types of venture capital. I think the seed investors are pretty much doing what they've always done, which is trying to find future big outcomes early. And the guys in the middle are doing what they've always done, which is doing B rounds and C rounds to jump on the bandwagon of success. But the growth guys are raising huge funds, I mean multi-billions, and are making bets where they're prepared to underwrite checks over a billion dollars for a gain of three, four, five X.

Speaker 1

When we get to the post of the week this week, you'll see one such investor, which is Index Ventures, who are one of the middle guys, returning something like $14 billion. Maybe it isn't as much as that, but a very large amount from just three deals this week. on three exits. So, you know, that doesn't sound like venture capital. Venture capital is meant to be, you know, reasonably medium-sized or smaller checks chasing 100 times your money. But the game now at the late stage is much, much bigger checks chasing two or three times the money.

Speaker 2

So it's all about scale, and that's another point you make. But it's always the point in Silicon Valley. You say scale isn't just an advantage. It's a necessity for survival and dominance, which is as true for these AI startups as it is for the VCs investing in them. But is that any different? I mean, scale is always the word that people throw around. They have been doing in Silicon Valley ever since I've been here the last 30 years.

Speaker 1

Well, it's different numerically. I mean, there's never been a time when people were investing double-digit billions into companies that are...

Speaker 2

But let me rephrase the question then. I mean, maybe it's a rather naive question, but is that ultimately that much difference between investing 10, 50,

Speaker

100 million and a billion or 2 billion or 5 billion?

Speaker 2

It's just money, Keith, as you would always remind me.

Speaker 1

Well, to insiders, it's massively different. If you think that something like 80% of all the venture capital raised this year has gone to five firms, five. Let's just remind ourselves of who they are. Andreessen, Sequoia, KOTU, Lightspeed. Can't remember the other one.

Speaker 2

It's a winner-take. I mean, as everything in Silicon Valley, it's a winner-take-all market.

Speaker 1

Yeah, but it's structurally changed to be much more barbellish, if you understand that reference. The barbell is when on the seed stage, there's a lot of activity, and on the late stage, there's a lot of activity, and there's not very much in the middle.

Speaker 2

Which is true of the entire digital economy. It's been true ever since I've been watching it, the disappearance of the middle.

Speaker 1

Yeah, and so structurally, it's changing shape, but if you are a startup, You know, unless you can catch one of those late stage investors, you probably don't have a chance because the table stakes to participate are so large that you need to underwrite market entry, not just growth.

Speaker 2

As more of a pessimist than you, Keith, always looking for disasters, you're always looking for the opposite of disasters. I don't know what the word is. It seems to me that if everything is going through five companies and one or two of them fail, then we have a huge problem.

Speaker 1

Well, who's we in that sense?

Speaker 2

We being the rest of the world, America, everybody else in the economy that they're controlling, essentially, or managing.

Speaker 1

Well, so my assumption is that the winning companies, and I think there's more than five, it's probably somewhere between 20 and 50 companies that are at scale in this new era. The winning companies, if we can believe the likely outcome is the growth of GDP, not just in America, but globally, due to productivity gains and the cost of labor declining, then the winner will be everybody. Everybody wins.

Speaker 2

Yeah, I mean, this is where you and I radically disagree. I don't want to get to this, yeah, I want to come to this later in the conversation, but I have to put myself clear here. I fundamentally disagree with you, but that's another issue. But leaving aside that,

Speaker 1

my question is more about... Where you wouldn't disagree is that that would be the goal. You and I would both share the goal that everyone benefits.

Speaker 2

Yeah, but the VCs don't care about that.

Speaker 1

They just want to make their money. Well, they make their money no matter who benefits, as long as they're in good companies. But it's what the output generated by this new technology...

Speaker 2

Okay, let's leave that for the moment. Let's just focus on this issue that you discuss in your editorial of scale being an advantage. It's not just an advantage, it's a necessity for survival and dominance. Let me rephrase the question to you. Does that make the broader economy more or less vulnerable, more or less liable to some huge dip, crisis, or perhaps uptick?

Speaker 1

I think the economy is super complex and there's no single point of failure. The AI economy then, leaving aside the broader economy. The AI economy, I don't really see a scenario that could lead to failure anymore. I mean, I think one of the big trends in the first six months of this year is the removal of doubt that AI is both real and impactful on economics in particular, but probably on everything. And, you know, so I don't really have that. Well, at least I can't think of a, you know, a set of events.

Speaker 2

I mean, again, I don't know.

Speaker 1

What would you imagine could happen?

Speaker 2

I mean, what I could imagine happened, for example, is that OpenAI fails somehow because it's still not a profitable company or that Altman gets revealed as some sort of fraudster, which already has been.

Speaker 1

I mean, they're at 14 billion annual revenue now. 14 billion.

Speaker 2

Well, we will see. But you and I have disagreed on that. So, okay, so scale... isn't just an advantage, it's a necessity for survival and dominance. But so what? Who cares? I mean, it doesn't seem that big. I mean, excusing the pun, it's not that big a deal.

Speaker 1

Well, with my startup hat on at SignalRank, I can tell you it's a huge deal. Just to give you a sense of it, my market for what I do is roughly $700 million a year that I could be investing. I'm currently investing annually about $20 million. So the gap between the scale I would need to really do a good job and where I'm at is quite large. But my heart,

Speaker 2

I mean, I want you to succeed, Keith, because maybe you'll take me out for lunch or something. But my heart doesn't go out to middle rank investors. Who cares?

Speaker 1

No, I'm not asking for you to do that. Mine doesn't either. What I'm more doing is, look, if I succeed retail investors, meaning you, can invest in very high growth, venture-backed companies. And if I fail, they can't. So if I can't put to work a much larger sum annually, I'm failing to deliver on a human goal. So scale for me correlates to a good human outcome.

Speaker 2

Although if you take your argument to the... level, and we'll come to Vinod Kostler later, that there'll no longer be any need for work in our AI age, then investing is work. I don't know what the big deal is. Anyway, let's move on to your fourth point about tokenization, getting its, and I'm quoting you here, its moment in the sun. What exactly is tokenization and what's happening? What has happened over the last six months? And how is that connected with The AI revolution.

Speaker 1

Yeah. So tokenization should be thought of in the same part of your brain that you think about securitization, which is what the stock market is. You know, a company creates securities that you can buy and sell. Well, you know, there's all kinds of regulations around that. A company could also create tokens that represent their underlying shares, but they don't because there's all kinds of regulatory things preventing it. What happened this week is Robin Hood did it for them. Robin Hood basically went out and bought shares in companies like SpaceX and others. OpenAI were the two they announced. And they created a tokenized version of those shares sitting on top of the shares they bought. And in Europe, where the regulations are less stringent, Robinhood users can now buy OpenAI and SpaceX tokens that give them a slice, so Robinhood claims, of the underlying shares.

Speaker 2

Which actually contradicts what you said earlier. So the ordinary retail investor like myself could actually get into OpenAI or SpaceX, which actually makes me much more nervous about the long-term viability of this

Speaker 1

boom. But just to be clear, you actually couldn't because U.S. regulations prevent us from being... Okay, well, my equivalent of me in Europe then. The equivalent of you in Europe could do it. They can already buy shares via apps like Revolut, American shares. So tokenization really brings the capital sitting in crypto markets to share purchases. So it's democratization in a way. It is. And it is a contradiction, by the way. is very far down in accepting crypto payment for our index shares. What you're seeing is a convergence of three things. Private markets, which normally are not available to retail investors, public markets, and crypto.

Speaker 2

Robin Hood... Yeah, it's fascinating, Keith, and it's... Sounds to me like a recipe for, again, disaster. We're back in 1999 with cab drivers buying Yahoo stock, basically.

Speaker 1

Well, it could be a disaster if bad actors persuade people to buy overpriced assets. And that already happens in the stock market. We have a competitor called DXYZ, as the ticker, who owns $70 million of SpaceX and Stripe and other shares. and trades at five times the value of the underlying asset.

Speaker 2

I mean, you may have a competitor who is not quite kosher, but it's not just about bad actors. It's people, I mean, going back to your gold rush metaphor, it's the culture of the gold rush where everyone believes that this thing will only create profit, so they buy and buy and buy until the market crashes. That's a classic, it's a classic boom, isn't it? a bubble, a boom is a boom, which always results. I mean, irrational boom, irrational exuberance. People can't imagine losing anything and then they do. And it's always the small investor who loses first and loses biggest.

Speaker 1

Well, if you think about histories, capitalism relies on this being true. What you just said has to be true for capitalism to work, because think of the East India Company back in the days of merchant capital. Everyone piles into whatever the growth asset is at any given moment, because where the growth is, is where the value is. So growth and value correlate. but also overpricing lives there. And so you only really make money based on timing and prudence, but you can make money because there's real value.