Druck能否开启加密货币牛市的下一阶段？

1. AI Bubble or Incumbents' Game?

Nothing said on Empire is a recommendation to buy or sell any investments or products. All right, the Halloween episode is here. I would highly recommend switching over. Although Spotify has video now, so you can watch on Spotify, but switch over to YouTube. We got the Halloween episode of Empire. We got Haseeb. Haseeb had a great performance last week, so we invited him back. Santi, look at Santi. Halloween Santi. We love to see it.

Going the distance. I’ve got to earn my keep in this pod. I feel threatened. Rob came in, and I’m starting to get the signals that it’s time for me to transition and find another career path.

Dude, I’m trying to level up, guys. I’m holding on for dear life.

I should hopefully recognize your outfit.

Yeah, I am the stablecoin mastermind, Rob Hadick. I was recently promoted to managing partner at Dragonfly, and I’m excited to tell you guys about why all of Wall Street finds crypto to be a joke.

There we go. I like it.

Time to grow up, guys. Time to grow up.

It’s time to grow up. The challenge is always that you need to mention, in any episode, Rain and Polymarket. What else?

Oh, I’ve got them queued up. I’ve got them queued up. I’ve got them up—at least Rain and Polymarket in—

And a horror film.

And a horror film.

And a horror film. Okay.

Those 3 things.

That I learned from last episode.

Yeah. Yeah. In light of Stanley Druckenmiller tweeting about TradeXYZ and Elon also posting about TradeXYZ, I’ve got to bring out the hat I haven’t worn in a long time.

Which turkey?

Wow.

It’s the chicken hat. Bulls make money; pigs get slaughtered. But in this kind of market, the chickens are doing the chicken dance. It’s a good time to be investing in crypto assets. It’s a good time, so we’re wearing the chicken hat for this podcast.

Okay, I was thinking of the turkey chart, but the chicken hat is also a good interpretation.

My first reaction was the tall turkey. You know, it goes up.

Yeah, exactly.

Yeah, that 80% drawdown.

Yeah, but then it drops.

It drops. It drops.

2. Google's 12% Earnings Drop

And we won’t be dropping. Well, speaking of dropping, we’ve had Google report earnings. The stock was down 12%. I don’t think there have been more than 3 or 4 times that’s happened, ever. It’s kind of crazy to see how momentum-driven markets are.

3. Druckenmiller Bets on Trade XYZ

Hold on. I refuse to talk about Google stock on a crypto podcast on the day that Stanley Druckenmiller, potentially the greatest hedge fund manager of all time, is making a strategic investment in TradeXYZ and saying, “Hold on, please.” He says—

Christmas.

“As a market participant, I’m looking forward to using”—actually, a double “to.” That’s how you know it was AI—“I’m looking forward to using their innovative product to enhance my investment process.” So—

How is he going to use it? I mean, was he the one who had a position on Hyperliquid? There are a bunch of fund managers—I think Jones did.

Yeah, I think D1—Dan, I think is his name—has a stake in PURR. But let’s discuss. To me, this is one of the most monumental things that has happened to crypto in a very long time. I don’t want to overstate it or be hyperbolic, but I think it’s a defining moment for the industry.

I’m sure you guys remember COVID. Markets were crashing, and it was like, “What is happening?” Then, in May 2020, Paul Tudor Jones came out and called Bitcoin the fastest horse in the race. I remember that day so well, because it instantly removed the career risk for every hedge fund manager to buy Bitcoin. If you said you weren’t long Bitcoin, it was like, “Okay, are you smarter than Paul Tudor Jones?”

Bitcoin was around $9,000 the day Paul Tudor Jones said that. Over the next 12 to 18 months, it went to $60,000, $65,000, $70,000. That was the bull run. I think this is that moment for onchain capital markets and DeFi. We had DeFi 1.0, and then DeFi 2.0, I guess you could call it. I think what’s about to happen is that onchain volumes are going to go up 8, 9, 10 times. This next year is about to get really nutty. Good for Hyperliquid, good for TradeXYZ, and hats off to this team. This is a monumental moment for the whole industry. Rob, what do you think?

I was very, very surprised to see ICE making an investment in TradeXYZ and then following up by doing a HIP-3 market, on the back of what Kraken was doing. Kraken recently announced more or less the same thing: they were going to have a U.S.-regulated HIP-3 market that would require KYC, where only U.S. participants could engage. That sounds like what ICE is also announcing it’s going to do, which is interesting. I almost find its investment in TradeXYZ more surprising than the market launch, because most of TradeXYZ’s business is launching these international, unregulated swaps, or perps.

What’s surprising to me is the idea that they’re going to do 24/7 perps in the U.S. Did any of you guys hear the Odd Lots podcast where they interviewed Terry Duffy and he talked about perps?

No. No.

Okay. Terry Duffy from CME has been very aggressive in being anti-perps and anti-24/7 markets. He filed a lawsuit against the CFTC, trying to enjoin it from allowing perp markets to be treated as futures instead of swaps. It seems clear that this is going to be a big fight in the U.S., particularly with CME on the other side, not wanting these markets to be allowed to trade in the U.S. The CFTC doesn’t seem to have made a clear decision one way or another.

If you look at the ICE-OKX partnership, they’re offering perps on a bunch of stocks. More than 40 companies have basically delisted their—sorry, these are tokenized stocks. More than 40 companies have forced their tokenized stocks to be delisted, because under SEC rules, a company can elect not to allow its stock to be tokenized and traded. So it’s not obvious to me how this is going to play out. I don’t think the answer is, “It’s a party, and everything is going to happen TradeXYZ-style in the U.S.” It seems to me there’s going to be a ton of infighting now that ICE is in the game. The grown-ups are playing, and it’s going to be like King Kong versus Godzilla duking it out over the city.

I don’t have much insight into the regulatory stuff, but I’m curious: how do you think Stanley Druckenmiller and his team are using TradeXYZ? We’ve seen the famous picture where they’re pricing the Cerebras IPO, and on the floor there’s the HIP-3 market around the pre-IPO perps.

I think that’s exactly it. I think it’s the pre-IPO markets. There are 2 things: pre-IPO markets and weekend markets, with pre-IPO markets being more important than weekend markets here. The entire global market is based around the success of a few companies. I’d argue right now it’s about what’s going to happen to Anthropic, what’s going to happen to OpenAI, and what happens to the SpaceX IPO. Basically, all markets are downstream of that. Hyperliquid and TradeXYZ have successfully captured that market.

Yeah.

For simplistic reasons, if you have private exposure to Anthropic or OpenAI, could you hedge that with decent size on Hyperliquid?

Really hedge that much size?

Not with his size, I would say.

I don’t think it’s about hedging. I don’t think they’re looking at it from that perspective. I think there’s just no historical market pricing on these IPOs. It’s all a guess. It’s literally like Bloomberg tweeting, “Oh, there’s actually more demand”—“It’s 10 times oversubscribed” or “5 times oversubscribed.” That’s how IPO pre-IPO markets have historically worked. There have been really, really, really illiquid markets around these things, and now, for the first time, there’s a liquid market someone can look at.

So, Santi, I think the dream would be, yes, maybe a hedging vehicle, but right now it’s just the first time we’ve ever had a semi—

I’m looking at this right now: Anthropic on TradeXYZ has $17 million of open interest, less than $4 million of that is hedgeable.

Nobody’s hedging anything on this.

But still, it’s the first time you have a price.

Yeah.

Where things are getting interesting is if you have a good reference point. People are saying Anthropic might be at $1.5, but if Hyperliquid all of a sudden shows a drop of 20%…

That’s a signal in and of itself that he and his team are going to take in, right? In the same way that you saw oil really dip down over the weekend during the war in Iran, it allows him to make sense of things in real time, ahead of others.

I think Polymarket is probably the most instructive example here, because ICE also invested in Polymarket. What they realized was that the data it’s producing is extremely valuable. Right now, it’s very difficult to hedge a meaningful geopolitical risk on Polymarket; the liquidity just isn’t really there. It’s going to be a while before that’s possible.

They have an institutional business where they can try to do these large off-order-book hedges, but most people are looking at the prices and odds to inform other decisions they might otherwise make. That’s where the data is really valuable. A lot of things are downstream of what the market thinks the Anthropic price is going to be. Getting that data quickly, packaging it up, and selling it to hedge funds is an obvious product you can sell right now, even before U.S. hedge funds of any meaningful size can participate in these markets.

Second, you’re going to get liquidity fragmentation, because TradeXYZ is mostly overseas. It’s primarily non-Americans trading on these venues. If you’re talking about a U.S.-regulated product and you’re a CFTC-regulated futures exchange, you need to KYC everybody. You must have market surveillance. You must know who everybody is, which means you cannot use the existing liquidity on TradeXYZ for the U.S. product. You’re going to have 2 order books, the same way Polymarket does. I think the answer is—and I’m reading between the lines; I don’t know—but my guess is they see that they can immediately start to sell the data.

4. Google Cloud Enters Custody

Because, like you said, ICE is probably already looking at the data: What is oil being priced at over the weekend? What do pre-IPO markets look like before the IPO opens? It’s not just about pricing the open. Everything is downstream of that. How you price CoreWeave bonds is going to be downstream of what people think the Anthropic IPO…

Totally. ICE owns the NYSE, but it also has a colossal business. I think it does about $5.5 billion from trading fees and trading services, but it also does $2.5 billion selling data. Its data business is half as big as its trading business.

5. DeFi Primitives Beyond Perps

The interesting thing about this investment—or, excuse me, the other thing that happened in the last couple of days that this is really overshadowing—is that Kraken is going to be routing, I think, a lot of Hyperliquid trades. They’re building a HIP-3 market or something like that. Santi, maybe you know more than me, but compared to the Kraken-Bitnomial instance, which routes the trades through its own broker, NinjaTrader, which it got through an acquisition, what ICE is going to do here, it sounds like, is offer Hyperliquid to all of ICE’s participating brokers. I believe that’s hundreds of members. It’s not just one exchange going through Kraken; it’s hundreds of participating members.

So, yeah, pretty… Santi, how do you think about this from an investing lens? Are you just going to buy? Do you go long HYPE in the open market? How do both of you think about it?

HYPE has already rallied a lot as of today, so by the time you’re hearing this podcast, the rally is probably already baked into the price. The big question is, what does this do? My view is that this isn’t as simple a story as it appears on its face. Like I said, there’s going to be a lot of fighting and a lot of roadblocks.

This is obviously very disruptive in the classical sense. If you’re Nasdaq, you’re very upset about this. If you’re CME, you’re very upset about this, and you don’t want a weird regulatory backdoor by which the NYSE can basically have a monopoly on a product that nobody else can sell. Nobody in this industry is going to say, “Yeah, cool. How innovative of you. I guess you guys can just go do that, and we’ll hang around and watch.”

There are probably going to be injunctions against them trading this product, and courts will have to get involved to decide whether it’s allowed. Judging again from the listing of 24/7 perps and the CME suing to enjoin the CFTC from allowing these perp markets to trade, my guess is you’re going to see the same thing happen for everything on TradeXYZ.

6. Midterms and Crypto Politics

I would definitely agree with that. I think it’s incredibly bullish. Who’s going to pick a bigger fight? Jake Chervinsky has done a good job, but now you have the resources, reputation, and muscle of ICE behind you to enter the U.S. I mean, it’s the biggest catalyst for HYPE to continue to grow. We’ve had discussions around whether it’s fairly valued at $60, $70, $80, or $90. Volume feels like it’s sort of plateaued in some way, so what’s the next catalyst?

To Santi’s point, I think the market already acted. HYPE’s back up to $94, $95. The real question in my mind is: If you take a multiyear view, can HYPE grow to be—what is it now, like $100 billion, give or take? Can this be $200 billion, $300 billion, $400 billion, $500 billion? What do you need to believe in order to get there? This certainly moves you in the right direction. We’ll see how it unfolds, but it’s certainly bullish to cement their leadership entering the U.S. It has so many spillover effects around perps entering the U.S., fragmentation, liquidity—all this stuff that’s top of mind. You now have a train that’s very motivated to drive through a wall. That has spillover effects not just for HYPE, but for other protocols in crypto.

From a regulatory standpoint, I would say—

Yeah, 100%.

How do we think about this? We’ve entered a new era of crypto. This was a smaller announcement this week, but I’ll read this from the crypto team at Google, which actually posted this: “This is our largest launch in the digital asset space to date. If you need crypto custody at scale, Google MPC provides security without trading off performance.” Google launched MPC crypto-key security this week, which they’ll sell to institutions, right?

Fireblocks is the one that should really sell to Visa or Mastercard.

So, look, I think the time is running out for crypto-native companies to scale. I was actually talking to someone at Google about this probably 6 months ago, and he said, “I think we figured out something on the MPC side that is better than how any of the other custodians in crypto do it today.” As I look at this, I’m thinking that if you’re not scaling enterprise sales, go-to-market, and international expansion, time is running short for you.

Okay, hold on. That’s a little too strong. I don’t know that much about this product. I read the article, and I don’t know why this guy at Google is tweeting it, because this is like a leak in CoinDesk. These are anonymous sources that CoinDesk is reporting from. This isn’t a Google announcement of this product. I’m very confused why he’s tweeting a CoinDesk article. If you read the article, it basically looks like a Google Cloud product.

No, no, this isn’t a leak. This is Rich Whitman, who runs crypto at Google Cloud. He tweeted, “Every transaction in crypto begins and ends with custody. We’re excited to extend our work to the most critical part of the stack, which is trust.” I do think this is a real product.

Sure.

What’s interesting here is that Google is one of those organizations that comes and goes. From the public-facing side, it would appear that they’ve ebbed and flowed, but I actually don’t know. It sounds like they have a pretty big team that’s continued to do work throughout the years.

The fact that they still have the unit there is interesting. Obviously, Google has many different divisions and experiments, but I wonder what they see in terms of the total revenue opportunity. That’s probably what you’re saying, Yano: they probably figured out, “Okay, there’s a pretty decent market to go after here that we can probably win with most of our customers,” rather than giving that business to BitGo, Fireblocks, or some other provider.

I don’t know if this is a full custody product. From the description, it seems to me that this is a Google Cloud product. If you understand Google Cloud, there are a lot of building blocks they sell that you can use to build a business with their technology on the back end. But as far as I can tell, it’s not an actual custodian.

It’s not like Fireblocks, where they support all these assets and transactions, provide contract signing, and have an interface and a mobile app where you can approve things. They’re not shipping any of that. They’re just shipping the back end: purely the cryptography layer inside enclaves, doing MPC using some algorithm they came up with. That’s it.

A Fireblocks competitor could use this technology on the back end and then build the application around it, turning it into a custodial product. But by itself, this is like Amazon KMS, its key management service for storing private keys. That’s not enough to build a turnkey product.

That’s super important to call out. They’re not delivering the nice app and everything you get with Fireblocks. They’re not doing any of that.

Correct.

Why wouldn’t they go directly into custodying with Google, quote-unquote?

They’re selling the service to someone who would do that.

They just don’t want that liability? Why wouldn’t they offer it?

That’s just not what they do.

Not their business.

That’s not their business. They don’t do that for anything, right? It’s Google Cloud.

It’s like AWS. AWS doesn’t sell apps; they sell the building blocks for you to build apps.

Yeah, I know.

7. Inside Dragonfly's Partner Shakeup

This is a complete side note, unless there’s anything else on Google. Haseeb, your predecessor here, Rob, was demoted from Empire but promoted by Dragonfly. Congrats to Rob and Tom.

What was his old role, and what’s the new one? Tell me what happens inside a venture fund, just out of pure curiosity. What’s the difference?

The first thing is that titles at VC funds are totally made up. They mean whatever the people there think they mean. Once upon a time, analysts at VC funds were called analysts or associates. They’re not called that anymore.

Then a16z came around and gave everyone the partner role.

General partner, which used to just be called partner, now implies that a person has investment authority. They can write checks, they’re part of the investment committee, and they vote on investment decisions. They also have economics in the fund in the form of carry—significant carry.

Managing partner is often a term that implies the person is managing the firm. People report to them, they often can’t be fired, and they’re built into the management company itself. They’re often key men on a fund. A key man on a fund generally means that if you leave, the LPs can decide to shut down the fund because you’re integral to the entire operation. If you’re gone, they might not want to continue with the fund.

Often, managing partners are designated as key men within a fund, which usually isn’t a good thing from the fund’s perspective. You don’t really want key men. You want the fund to keep going if you get sick or ride off into the sunset, so everyone can keep doing their jobs.

The promotion from general partner to managing partner—which both Tom and Rob at Dragonfly received—basically means that the fund is the 4 of us. The 4 of us make decisions together. De facto, nobody can fire anybody else, and decision-making is now joint among the 4 of us. There are no reporting lines anymore. VC funds aren’t like companies; they’re partnerships, like a law firm or Goldman Sachs. That’s the new world order at Dragonfly.

How does it work when you want to make an investment? Is there a veto? Feel free not to answer, but I’m curious how you think about it. With 4 people, sometimes you might not reach consensus because it’s an even number.

We have 2 mechanics for this. When a deal goes to a vote within the investment committee, you only need a majority. If 1 person is a no, you can still get the investment done. That’s important because good deals are often contentious.

We also have something called the gunslinger budget. Once per fund, as long as there’s 1 GP who’s for the deal, that GP can push through a deal even if nobody else said yes to it. The clause is that you have to put your own money into the deal as well.

Sequoia did that with YouTube, I believe. Roelof, I think, was the gunslinger for them, where most of the partnership didn’t want to do the deal.

I think so.

He came in and said, “I want to do this. Chad’s a great builder.” I think Chad, right? Or Hurley? I think he used that. Is it tied to a particular number or dollar figure as a percentage of the fund?

No, there’s no limitation on the amount it could be. You have your own risk parameters, though, so no deal will account for more than a certain amount.

Yeah. Okay, that’s interesting.

8. YC's Crypto Startup Wishlist

There was this list that came out. Y Combinator had an interesting tweet. The title of this episode should be something like “The next phase of the bull market,” or something pithy like that, because I’m looking at the timeline. We didn’t mention that Elon quote-tweeted Stanley Druckenmiller and said, “Interesting,” right? I’m getting so many flashbacks to the last bull market, when Elon was buying Bitcoin for Tesla’s treasury. The market is definitely starting to heat up a little bit.

As part of that, I think crypto venture was so hated for the last couple of years. I’ll give you guys a big tip of the cap, because Dragonfly is one of the very, very, very few funds that basically doubled down on crypto when everyone pivoted away to AI.

I’m trying to do my hair.

Oh, nice.

Oh, nice. I got—yeah, I can do a little chicken leg here.

Y Combinator came out and said, “We’re excited to fund more crypto startups.” Their request for crypto startups was AI capital raising. They said AI investment is one of the largest capital-raising events in human history. So this is basically about how you can leverage the blockchain to make AI capital raising easier.

It always comes back to ICOs. Stablecoins: new types of stablecoins, new currency stablecoins, and neobanks built on stablecoins. Number three is trading, which is tokenization, exchanges, and trading interfaces. And then four is agents. They said agents are going to be prolific traders: build agentic infrastructure, build agents using crypto APIs. So I’m very curious: when you see a list like this, is it, “Oh, great ideas. Wow, nailed it,” or is it, “Welcome to the club. You’re maybe 2 years behind”? What’s your thought when you read a post like this?

I think these kinds of posts matter more for the spirit of them than they should be taken literally. The reality is that most of these are terrible ideas for startups. I love Neil, and I love that YC is doubling down on crypto. I think it’s beautiful to see, but the reality is that VCs have terrible ideas for startups.

It’s only very rarely true when there’s total white space in an industry, like there is, I think, for many aspects of AI. There’s just a bunch of stuff—like construction—where there’s so much white space that somebody can come build some things in construction, and it’ll quickly become a big company. But crypto is a very mature market. It’s been around for a very long time, and there are so many startups you probably don’t know about that have tried many of these things, got absolutely killed, and learned that nobody wants this.

One of these ideas was non-U.S. dollar stablecoins. 99.8% of all stablecoins are U.S. dollars.

If you just read this, you might think, “Oh, wow, nobody has tried it. What a bunch of idiots, sitting around making non-U.S. dollar stablecoins?”

It’s actually making the dollar stronger. People in other countries don’t want their currency; they want the dollar.

Yes, yes. There are a lot of these ideas that are very convincing when you first read them. I’ll tell you: when I first got into crypto investing—when I first got into crypto, period—the first startup idea I had, because originally I was going to be a founder, was building Bloomberg for crypto.

I didn’t know at the time that “Bloomberg for crypto” was a meme—a vacuous, stupid idea that had been tried a thousand times and never worked. What would even be in a Bloomberg for crypto? It’s one of those ideas that everybody comes up with first for a startup. As a VC, I’ve gotten that pitch hundreds of times. Another one of these ideas is tokenizing random assets.

Yeah, yeah. Don’t tokenize random assets.

“By tokenizing an illiquid asset, we’ll make it liquid.”

Yes. I promise you, that’s not a good startup idea.

Fellas, as much as we’d always like to see bullish signs for the glorious times, there are bearish signals here. Every time Y Combinator decides to reenter crypto and put out these proposals, it’s typically not early. The last time they did that was around the peak of 2021. They’re certainly not leaders when it comes to their renewed interest in crypto.

Well, to give them some credit, Neil has definitely turned around the crypto program at YC. Before Neil, it was dreadfully bad, and the previous YC leadership was pretty anti-crypto. Neil has revitalized the place, and he’s a tireless advocate for crypto there, which I think is awesome.

I just think this request-for-startups stuff is a problem. Every time I see it—even a16z used to do these a lot; VCs do them at the end of the year—I’ve never seen a good startup come out of a crypto request for startups. I can’t think of a single one.

9. The Four Stages of a Bull Market

Yeah. And to your point, Santi, there are stages to every bull market. The first phase is when nobody is talking about the bull market except for this tiny crowd on Twitter. Think back to 2023. Bitcoin actually 3x’ed in 2023, and Solana ran up a lot, but nobody was talking about it.

Then there’s the next stage: some companies are still doing layoffs, but prices are up. Everyone’s thinking, “This feels good,” but they don’t want to let themselves feel good yet because they’re still scarred by the last 18 months. Phase three is, “Okay, bull market, let’s go,” and fundraising gets crazy. Stage four is Justin Bieber launching NFTs, and you should probably sell it all. So this feels like—

We’re probably in mid-stage two or something like that.

I think they’re feeling really good, but Justin Bieber isn’t launching NFTs.

I think my point is that there are pockets of crypto that will do well no matter what the macro environment is like. Talking about bull and bear markets is kind of irrelevant as an investor. You can certainly try to time the markets, although I’ve never met someone who can do it systematically.

Yeah. Especially not anymore. These days, it’s just really hard to do that.

I think the more interesting nuance is: which sectors, and why? Tokenization—I think that’s a terrible idea. As much as I don’t know this guy who’s now running the pro-crypto program at Y Combinator—you should definitely come on the pod—it still signals a lack of edge in really understanding where the ball is in crypto.

10. Fidelity

That makes it even more impressive to see someone like Stan and David Tepper, and some guy, and Paul Tudor Jones really cut through all that and just say, “No, perps are really interesting instruments. Let’s just back that,” versus this spray approach of tokenizing random things. Anyone who’s spent meaningful time in crypto knows how bad an idea that is. It’s been tried. It’s just not it. There are a lot of reasons it won’t work. Just back the fastest horse; back the killer product.

To me, it signals that either you believe we’re late and perps aren’t sufficiently interesting, or you don’t see that perps totally transform market structure and price discovery. Get behind it. It’s still super early. Forget about all that. This is why my point about Hyperliquid is that I’m not sure if it’s worth 80, 70, or 60, but I want to see it go fast. Can it become 500 or 1,000 bucks per HYPE? That’s a more interesting conversation than whether it rallied 20%.

Well, let me stand up for it a little bit and play devil’s advocate. What’s the purpose of a request for startups? Obviously, he’s not telling you what to buy or giving investment advice. This is for founders who don’t know what to do. If you’re a founder and you don’t know what to do, and you want to build something in crypto but don’t have good ideas, you read this.

I think the best defense of it is that it’s meant to be a generative prompt. It’s supposed to increase the diversity of your ideas. You don’t literally do the thing on the list, because those are obvious ideas that are low-hanging fruit and a bunch of people are trying them at the same time. But it’s meant to get you acquainted with some of crypto’s surface area and what people might do if they don’t have any edge, earned insight, or deep experience in one particular area.

I take issue with some of these ideas because I think some are absolutely bad. But the general theme of a request for startups is that they publish these across cybersecurity, AI, robotics, biotech, and other areas, and I think they can genuinely be useful for founders to get an idea of what’s out there, what people are working on, and where the zeitgeist is. If you ask me what the zeitgeist of new crypto services is right now, I don’t know that I have a great answer for you.

You know who used to do this quite well? Vitalik.

Vitalik. Oh, yeah. I mean, Uniswap—he had a couple of really interesting ideas in his blog posts back in the day. Prediction markets: he talked a lot about prediction markets.

Yeah.

Other people were talking about it too, but Hayden sort of picked up on Vitalik’s idea. Bancor might claim some of the credit, but it was really Vitalik who said, “Hey, this is interesting. Someone should build it for the Ethereum ecosystem.”

Hayden Adams, the founder of Uniswap, has a really impressive story. He didn’t know how to code, taught himself Solidity, and then built Uniswap. It’s a remarkable example of what can be possible in crypto. Vitalik had really good ideas. Then he kind of became too woke for my liking and started hating DeFi and financial infrastructure, which is one of the many reasons I think Ethereum is deeply uninteresting.

I was with the founder of what I’d consider one of the best up-and-coming DeFi protocols yesterday. They were very tied to Ethereum mainnet. They wanted to build on the L1 because their thesis was that institutions trust the L1 more. That was their view. They’ve been decently successful: it’s working, they’ve raised money, and they’ve got a lot of TVL growing.

But they said, “Look, at the end of the day, a lot of the TVL—especially in the world we’re building in—is actually on Base right now.” And they said Base—

What type of primitive?

Yeah.

Really? Is it like a money market? What is it?

It’s a lend-borrow.

Okay, a money market. How is that fascinating? What are they doing that isn’t being done today?

I’m not going to share the whole startup. It’s different. They said the Base team has done unbelievably well over the last couple of months at onboarding. There’s so much idle capital sitting in Coinbase that if you can activate it in some way, it gets really interesting.

This is just something I heard. Someone who should be building on Ethereum mainnet is going to build on Base. They said the next place they’ll deploy is Robinhood. It’s just one anecdote, but—

What are the more interesting DeFi primitives you’ve come across? We’ve talked about perps; I think the answer might be perps. Other than perps, I’ve felt the innovation hasn’t been there. Maybe it doesn’t have to be, as long as you have these core primitives that are enough to take over Wall Street: stablecoins, real-time AMMs. But is there anything truly new you’ve come across? You guys see a lot of good founder ideas, putting aside Y Combinator.

I saw a paper making the rounds on Twitter that I didn’t fully read, but I got the gist of it. The idea was to take some of these 24/7 on-chain perps, tokenize them, and split them into 2 parts: overnight exposure and daytime market-hours exposure.

What they showed, which I didn’t know, is that apparently stocks perform about 100 times better overnight than during market hours. I don’t know if you guys were aware of this.

Performance meaning the price goes up more overnight?

Yeah. If you held the S&P only overnight—

I’ve seen tweets about that. Someone did come up with this strategy: if you’d just done this, you’d have had a crazy return.

Yeah, it’s a crazy return and extremely consistent. I have no understanding of why it would be true. But they were saying that if you actually tried the strategy, you’d get killed entering and exiting your positions at the open and close every day. You’d get eaten alive by fees.

You can’t really do this in a practical way unless you tokenize the position and separate the risk of the daytime position from the overnight position, with different people holding different sides. If you take these perps and use some mechanics, you can separate the 2 sides and get exposure only to the overnight movements in these stocks.

That sounds interesting from a strategy standpoint. The thing I don’t buy is the fee point. That would mean anyone coming in and out of positions gets killed by the strategy. But there are high-frequency trading firms doing this consistently, even with fees and other friction. Unless there’s a higher fee for overnight trades, which I’m not aware of. I’ve traded overnight, and there’s less liquidity, but the fee is the same.

I think the idea is that if you’re entering a big position and closing it every single day, you’ll face enough slippage with enough size to cancel out the effect. There’s basically a no-arbitrage condition on this overnight holding trade.

Got it.

That’s what I think was implied. I don’t know; I just remembered seeing the tweet a couple of days ago.

Santi, is the person building this an engineer or a trader?

I have no idea. I shouldn’t be trying to represent it. I just remembered seeing the tweet a couple of days ago.

All right, folks, other news from the week. Tell me if you fancy any of these: Instinct raising at $30 billion. This was a Rob Hadick favorite to talk about.

Was it?

Yeah. He likes to position crypto in the limelight, saying AI is a bubble, so you should rotate your dollars into crypto, which is—

Not the worst take. Bring back Rob Hadick.

You can’t say, “This thing is going to explode, therefore this thing might not.”

Yeah, buddy. That’s right. Well, to be clear, if the AI bubble explodes, we’re coming down with it.

No, no, no. It’s all going to rotate into crypto. Thank you very much. I love your take. You guys are investing in AI too, and you’re investing in crypto. Call it fintech; it doesn’t matter. At the end of the day, if you back a team that has product-market fit, real demand behind it, and a reasonable valuation, then—

I mean, specifically talking about Instinct, I think it’s very plausible that it blows up. These companies are getting so leveraged on compute that it could literally run out of money. They’re losing so much money per user, and it’s not Meta. They don’t have the balance sheet to fund this thing at a massive loss forever.

If investor appetite suddenly seizes up, they’re going to have to sell half the business at a massive down round. They’re building up a pretty big capital stack because it’s so expensive. And apparently Noah is unwilling to charge customers.

I listened to his podcast—I think it was on Invest Like the Best—and he very much said that. That was my first thought. There was a really interesting tweet from someone at Microsoft that’s catching fire. He said, “All of this makes me think AI structurally favors incumbents. The capex they committed years ago becomes the cost and scale advantage across every consumer product they launch.”

I couldn’t agree more. It doesn’t mean there’s no place for startups, but look at earnings week: Google and Amazon are all raising capex. It’s increasingly hard when you see Elon tweeting to try to drive prices up. Compute is really hard to come by and getting more expensive. Everyone—from Sam to Dario to Microsoft, pretty much everyone building these companies—is saying the same thing.

We’ve talked a lot about distribution. I’m curious if you agree with that take, and how you translate it into your investing approach.

I do agree with it in that the biggest beneficiaries of AI have clearly been large companies. Nvidia, Microsoft, Amazon—the hyperscalers are making tons of money from this AI buildout, and they have the capex to support it.

Nobody’s gone bust yet, except Character.AI, Windsurf, and a few of these players who were very early and had these weird pseudo-exits. For the most part, the money has always been there. As long as you have users, people will fund you, even if you have no business model.

When you see a contraction—and we’ve seen mini tremors, but not a proper contraction yet—we know it will come. We know how it works in crypto: even on the way up, there are contractions. We’ll see one at some point, and a lot of startups that can’t fund the capex without selling equity will get frozen out.

Big companies can just weather the storm. They can stomach it because they’re funding the stuff themselves, and they have the balance sheets and the underwritability to issue debt and keep going.

All that said, there are tremendous venture-backed companies. Anthropic is obviously going public very soon. OpenAI, Cursor—you’ve got all these companies that have done incredible things on the startup side, with some of the fastest growth rates we’ve ever seen in any category.

But from a VC perspective, it’s clear that in crypto, almost all the value was captured by startups.

Yeah. Almost no value was captured by JPMorgan, BlackRock, or those guys. They got in the game. BlackRock has obviously been one of the biggest beneficiaries on Wall Street, but it still captured almost nothing compared with what the industry captured.

AI is the other way around. These startups are making a lot of money—Anthropic and OpenAI are great businesses—but they’re a pittance compared with what Nvidia, Amazon, and Google have added to their market caps because of AI.

Yeah.

Do you think it’s necessary for the share incumbents take from a disruptive technology to be sufficiently high in order to clear the regulatory friction required to get it through the door? Any major disruptive technology has had fierce regulatory pushback. At some point, you need incumbents to be motivated to advocate on your behalf, or you build a good enough product and survive, but the regulator has no choice, as in the case of Uber, where clearly the consumer loves this. It is just—

Political seppuku to—

Ban this type of technology.

Yeah. If you look at the internet, it didn’t really have the same level of pushback. The only real pushback against the internet was the telcos, because they were potentially going to be disrupted by Voice over IP and lose their ability to charge as much as they did for phone calls. Besides that, it wasn’t really obvious who was supposed to be the enemy of the internet.

The music industry, I guess you could say. Napster failed; Spotify.

Yeah, but that’s not the internet, right? That’s like—

Oh, no. I mean, it’s a small subset of commercial interests.

Yeah.

Yeah. But the internet itself, in the same way, crypto was purely disruptive in a pretty unique way, more similar to Uber or Airbnb in the extent to which it’s purely disruptive—

Right? Every single incumbent is like, “The hotels are like, ‘Fuck Airbnb.’ The taxicabs are all like, ‘Fuck Uber.’” Crypto was kind of like that until 2023, when we started getting ETFs and BlackRock and financial institutions came in. Then all of a sudden, things were different. And, of course, stablecoins went big, and the US government started caring about us internationalizing the debt. So I think that’s what turned the tides for crypto to have actual allies. But, you know, I think for AI it’s easier, but ironically, AI is more unpopular than crypto is now.

Yeah.

So I think we’re going to see a big vibe shift post-midterms.

Yeah, I agree with you. For me, it was definitely a turning point tied to Hyperliquid and ICE. The fact that these guys are going to be beneficiaries of prediction markets, stablecoins, and tokenizing these things via a DAT or an ETF is super bullish, because now they’re on your side. They’re the biggest lobbyists, and they’re instrumental. Whatever administration comes in, they have huge weight.

I can’t help but think—I haven’t seen this before, but the ICE announcement with HYPE is super bullish. That’s probably the most bullish thing for perps, which have always struggled to come into the US.

11. ICE's Sprecher on Hyperliquid

There’s a new thing that just dropped. Jeffrey Sprecher, who started ICE—let me read you this quote. It goes, “I think, stepping back, we heard about Hyperliquid, gosh, about a year ago. We really just heard about its tech features. We were able to track them down in Singapore, and I actually had a meeting with them well over a year ago to talk about working together. But it was really, you know, I was invited to a press conference in the White House, and President Trump actually asked me to stand next to him when he started talking about the crypto bill. As he was doing it, I looked over and saw his script and saw he was going to talk about Hyperliquid. In that moment, I said, ‘Okay, the US wants this platform to come and be regulated in this country.’”

“So we redoubled our efforts to bring both founder groups together, both Trade.xyz and Hyperliquid, to work with us. These markets, as I mentioned, we believe are complementary.”

I don’t know what the lesson is, but there’s a lesson there.

Well, the lesson is: try to see what the president has written down on his legal pad.

And then get behind whatever that is.

There’s a lesson there.

You can’t say there weren’t signs. I mean, the thing that—

Also strikes me, if you’re trying to draw the line out, it’s very clear. Let’s think about it in perspective, right? The midterms are next week. It’s very clear that MAGA is going to get shellacked in the midterms. It’s only a question of how much.

And right now, they’re at 94% to win the House, and—

Roughly 70% to win the Senate.

And when that happens, everything they’ve been saying is that the first thing we’re going to do is go after Trump’s corruption. At this point, it’s one of the highest-salience issues for both Democrats and Republicans, which is weird, because normally it’s the economy. Normally corruption doesn’t rate, but right now corruption rates. So you’re seeing it in a lot of the campaigns for folks in the midterms. This fake hair is getting in—it’s everywhere. It’s very unpleasant.

The fact is, for the next 2 years, we’re going to see a parade of news around Trump’s business interests, his family’s business interests, particularly around crypto, because that’s the easiest shorthand. There’s a lot of other stuff that has nothing to do with crypto. There’s Kimberly Guilfoyle, their sons investing in drone companies, and blah, blah, blah, but the crypto stuff is the easiest to explain.

If I—yeah, go ahead.

On that point, because you said something earlier which I agree with, I think the tides are turning. Right now, AI is top of mind for people because of data centers, and it’s easy—people always want to pick a scapegoat. There’s an affordability crisis, so people say, “What is that? It’s a data center. It’s raising my electricity bill.” It’s really not that; it’s just the war in Iran and all this.

Nonetheless, I think we’re entering an inflection in AI around usability, with consumer apps like Muse and others entering, and their adoption making it apparent to people that AI is a positive force in their lives. Getting an executive assistant, basically for free in the case of Instinct, is an example.

That’s sort of the—I see.

I don’t think it relies on the personal assistant. It’s not the personal assistant. I think if biotech can have massive breakthroughs and AI starts curing cancer, it’s not like if AI starts booking my flight for me, then I all of a sudden—

Don’t underestimate the small things, dude. The wonderful experience you might get with that. Yes, yes. Obviously, healthcare is going to take much longer. We’re underwriting it right now. It’s going to come for sure, but it’s a very concrete and simple thing where your quality of life is unequivocally better when you have a personal assistant like Muse. I mean, just for a sense of scale, Muse right now has about 2.5 million daily active users. It has a 30% retention rate after 7 days, meaning that—

70% of people stop using it very quickly.

They have, like—

Most people, they have 10 million downloads.

Yeah, yeah.

Not 10 million users.

Most people do not need a personal assistant. They don’t know what to do with it. Most people don’t get that much email. They get a lot of spam, you know. Having seen enough of my friends try to use these AI agents, I realize that, as somebody who’s very busy and has a lot going on in my life, I’m so atypical. I’m like, these are so amazing. They absolutely change—

This is exactly it. I have 6 flights in the next 2 and a half weeks. My friend, who I was trying to get on this, isn’t flying this year. He’s like, “I don’t need it to book a flight.” He’s like, “I’m not going to fly this year.”

Yeah. Rob did say this: most Americans don’t get on a flight.

12. Content of the Week

All right, content of the week. I’ve got to jump; I can’t be too late for this call. Content of the week, Haseeb. Now that you’ve had a week to prep, you know we do this. What do you have for us?

I did. I just saw “You Can See Everything.” Kind of horrorish. Incredible film. I have a lot of frustration with it, but it was a very good film.

On the Halloween theme, I don’t like horror, but I did see the new Stanford Prison Experiment on Netflix.

Oh, interesting. So they did a Stanford Prison Experiment.

Well, people might remember the Stanford prison experiment: They put a bunch of normal people in groups of guards and prisoners. It’s a fascinating psychological experiment about authority and whatnot. They recreate that with some flavor. It’s interesting. I was expecting a bit more out of it, but, you know…

Sounds very MrBeast. Okay.

Kind of. Yeah, yeah. But it’s tame. Anyway.

I have a podcast recommendation that isn’t Invest Like the Best and isn’t related to David Senra. Those are both banned podcasts; we’re mentioning them too much. It’s Ty Brown on Money Stuff, talking about gambling proxies, the DAT craze, and non-cross-collateralized lending. It’s a very wide-ranging conversation, about 40 minutes. It’s good money stuff.

I’ve got to jump. Folks, this is great. Hib, are you leaving us? Is this the transfer?

I am, unfortunately, leaving you. We’re going to get the real Rob back on the show. I wasn’t cutting it.

Oh, we’ve got—ladies and gentlemen, we’ve got so much pain. His hair grew out, and it was gray.

I can’t do this anymore, you guys. You’re stressing me out too much.

I see. All right, fellas. Pleasure. Go enjoy the weekend.

See everyone. See you guys.