AI疲劳、Robinhood与每个市场的加密化｜每周综述

Nothing said on Empire is a recommendation to buy or sell any investments or products.

Now we're ripping. Reporting live, ladies and gentlemen. We don't do anything live-streamed; it's all very heavily edited. We're missing Yano today, but Rob and I have things under control. Rob, what's going on, man?

We're letting the inmates run the asylum this week, which is great. I always love to see that happen, so just bear with us.

We have come a long way, my friend, since you were a guest and now you're running the show. You're left on my screen, you know.

You know what's funny? My face is still up a lot when they do the little clips, but I'm still not on the masthead. Maybe they just needed somebody with good bone structure, like a little face to put up front there.

1. CLARITY’s Falling Odds

Are you talking about the jawline here?

Yeah, the jawline.

Got it. Nice. All right, let's get straight to the jugular, I guess. We need to talk about CLARITY coming off last week. It seems like it was at an all-time high. There's a lot of optimism on the timeline. I think Brian Armstrong put out a couple of tweets. You're a CLARITY guy, so what's going on there?

I talked about this a couple of weeks ago when I was on in D.C., and I think you guys had Greg, the GC of Multicoin, on last week, which was a good podcast. Listen, Polymarket is, I think, down to 27%, and it's been trending down for a while. There was a sort of spurt of optimism on the timeline, which I think is understandable. People want to get it done, so they're excited about it. They're talking about it in public, and they're showing a lot of conviction and confidence because they think that helps get it done.

2. AI Capex and Semiconductor Volatility

If you talk to people in private, I think the reality is that there's a lot of bearishness around the expectation that it gets done. There are basically 2 open items. One is the ethics language, as we've talked about a bunch: What is the administration and the Trump family willing to do around carving out some of their ownership in these crypto-related projects? How are we going to litigate that in the future should there be complaints? Who would have the power to bring those complaints? Would it be the FBI or the state AGs?

That continues to be probably the major issue. There is 1 other issue around these developer protections with what's called BRCA, and how federal law enforcement feels it is able to use different types of tooling to potentially catch or bring bad actors to justice. That piece is also quite fraught at the moment.

Senator Cortez Masto, who's one of the leaders on the Democratic side and has been championing this, has had a perspective on that that both the Republicans and the rest of the industry likely would not be okay with. At least she put out some language a couple of days ago that said, "Hey, this is newly agreed-upon language with us, some of the people on the Republican side, and some of the people in the industry." That language included things that the industry and the Republicans would not have agreed with.

Those 2 topics right now feel potentially tough to surmount. We're still trying to get something done. I think there might be a vote next week. The vote might be along partisan lines, and then we end up in September. The longer this goes on, the less likely it is that we get to an agreement, but I'm still hopeful. A lot of people are still working really hard, but we're in a tough spot.

3. Content of the Week

This is not the Rob voice that I heard last week. Last week, you were like, "Let's [expletive] go." Right now, I'm looking at Polymarket. As we know, we like Polymarket. We're not sponsored or anything, but Rob and I like Polymarket. It's at an all-time low: 30%.

It felt like we were almost there. It jumped all the way to 46% last week, and then, like semis, man, it's just traded down.

Is that what you have in your mind? What is happening to my semis book? What is happening to my memory book?

4. Everything Trades Like Crypto

Listen, I'm just saying the market's up today. We'll get into the market and stuff. Boy, I thought—well, anyway, we're not going to get there just yet, but the market's up today. I thought I'd seen some volatility in crypto, and now the new trade for crypto bros is pivoting to memory and just getting absolutely whipsawed.

Here's the thing: If you weren't in crypto, you survived. Pivot to just trading all assets, and I think that's where the world is going.

On a more serious note, we're going to talk about Robinhood Chain. We're going to talk about tokenized stocks. Robinhood reported earnings this week, and we're going to talk about that, including their crypto numbers, which was a really interesting stat.

But zooming out, markets in general have been pretty volatile. You have semis doing 10% up 20% today. That's just—if there's 1 takeaway from all this conversation, it's don't use leverage, ladies and gentlemen. Charlie used to say—what is it? Ladies, liquor, and leverage. Of the 3, definitely avoid leverage. It's just really hard. I know you guys don't use much leverage.

No, we don't at all.

It's crazy. Should we talk about, obviously, the most important headline that was running around yesterday?

Before we get into situational awareness, which I know you want to talk about, I want to make 1 last point on CLARITY before we move on. You've seen it come out from both Chair Gensler and the chair on the SEC and CFTC side: They're ready to move forward with rulemaking and try to put a package together to help the industry have clarity on how to move forward, even if there isn't actual legislation.

That obviously creates a situation where it could be rolled back under different chairs. But I do think, regardless, the SEC and the CFTC are going to move forward with haste here. They're going to try to bring something that is good for innovation onshore, good for bringing back entrepreneurs, and good for both the industry and finance, as well as the companies that want to participate in this space. That's positive. Not all hope is lost if CLARITY does come along.

Are you buying at this level in Polymarket? Not financial advice, but, gun to your head?

No, I wouldn't buy. I think it's appropriately priced. I wouldn't buy the 30% on Poly, but do I think that crypto will be in a better spot in 6 or 9 months, regardless of whether it becomes law, with the leadership in the SEC and the CFTC? Absolutely.

Let's not forget that we're at a point where a conversation is happening. 2 years ago, 3 years ago, there was no conversation. Just the fact that we're talking about it is positive, in my opinion.

Obviously, the bill in whatever form—I still think there are different stakeholders here that care about certain things more than others—but a bill passing as it is, or somewhere close to where it is, would still be much more positive than a bill not passing. We talk about the signaling on the margin: Are enterprises going to slow down, and is it going to be an issue?

I think GENIUS is a landmark bill, but CLARITY would just be so much more. It's a net positive for the space. There are people who are always on the margin and want a little bit more here and there, but as it is, if you were to pass it, I don't know if you would disagree with me here, but I think it would be much better for the industry if we got a bill passed. It provides legitimacy, and then you can do amendments and whatnot. It would obviously be much more detrimental if we didn't pass it.

100%. I hope it gets passed. We should continue to work to try to get it passed. A lot of people are working on it. But we're going to get rulemaking out of the SEC and the CFTC regardless. We'll be in a better spot regardless. I applaud everybody who's working on it. Hopefully, we can get something done.

There you go. Where do we go next?

We have to talk about your boy—

No, but if we get in there, we're opening a can of worms. I do think we have to talk—

How are you, by the way? I feel like you're a little bit of a mini Leopold here. Are you surviving? Did Goldman—did they liquidate you?

I got an email from Goldman today. It was not about that. It was about something else.

No, no, not at all. Coming out of the last pod, I was like, “Look, I’m unbothered, moisturized, in my lane.” I’ve always felt like these things are extremely volatile, and it’s similar to crypto. I’ve always felt like it’s very important to underwrite a thesis, but factor in, “Okay, what’s the volatility of these things?” Just bake that into it, and also have a certain price in mind.

I still think we’re nowhere near capturing the opportunity in AI and semiconductors in general. It was a big week because you had Meta and Microsoft report earnings, and everyone’s just looking at Google from last week. There were great discussions around Google’s numbers. What the market is really focused on is capex. There’s a great post by Satya Nadella around the ROI of AI and how they built it.

It was kind of surreal to see something like vibe coding just create an artifact that he was sharing in earnings and then tweeted. Everyone’s on the timeline, right? Meta, Google, and Microsoft have all maintained or increased their capex guidance, so I think the market is responding well to that. And the longevity—I guess this applies to your portfolio, USD.AI.

You remember, there’s always a wall of worry that you need to climb when it comes to crypto, AI, and investing in general. In AI, we’ve heard questions about the usefulness of these GPUs: If you’re going to invest all this money in building out hyperscalers, what’s the life? What’s the useful life of that?

I think one of the interesting things that I saw yesterday in Satya’s post is the ROI and the useful life. I think Meta actually said it: The useful life of these racks is much longer than what they anticipated, and I think longer than other people are anticipating. That obviously bodes well for everyone who’s behind deploying these.

But yeah, semiconductors, ladies and gentlemen, if you haven’t been paying attention, have been more volatile than crypto these days. Crypto is quite muted.

Just a trillion-dollar asset with 100–120 vol. It’s crazy.

Totally, totally, totally. The last thing I’ll say is that I think the Korean market is extremely volatile. It has had more circuit breakers—and there’s a circuit breaker, for anyone who doesn’t know, when there’s a pretty big move in the market, up or down. I think there have been more circuit breakers this year than in the entire history of the Korean stock market.

I think that’s relevant for crypto because you said earlier that crypto has foreshadowed what is happening in every other market. It’s not that the entire market is coming to crypto, or crypto is trying to mold itself into the other markets. I think that means the market is becoming more like crypto.

5. The Death of Long-Term Investing

What I want to get your take on is, in your conversations—because you come from TradFi, too—if you’re in an institutional fund, I understand retail, Robinhood, zero-day options, and that whole phenomenon. But if you’ve been paying attention to crypto, we all know it well. Let’s talk about institutions for a second. It also feels like institutions are becoming incredibly short-sighted and momentum-driven. If you’re running a hedge fund today, is it the most difficult time to run a hedge fund ever?

I have to tell you—listen, I’m not an active trader, although I’ve done a lot of investing in capital-markets infrastructure. This was bound to happen, obviously. The advent of AQR and everything that we’ve had since then has made trading much more quantitative, with much more machine learning being used when people think about their trading activity.

That’s why you see that the vast majority of trading activity today is automated. It’s ingesting a bunch of data and spitting out a bunch of data. It strikes me that what has happened is we’ve lost the long-term, buy-and-hold, fundamental investor because people cannot stomach the volatility that’s been brought into the market by more leverage.

I think Goldman said in the final half of 2026 that there was more leverage at their prime brokerage than they had seen since before the global financial crisis. Not only that, about 20% of that leverage was basically in just a few AI memory-chip names.

You have this bunching happening at the biggest names. You have leverage being put into the market in a way that we’ve seen in crypto. You have retail active trading becoming a much bigger part of the market, and it often feels like retail active traders are the marginal buyer.

A lot of the market makers, and generally a lot of the takers, are using many more quantitative strategies instead of just buying and holding long-term. You bring all that together, and the volatility in all asset classes is just exploding. It’s exploding in a way, and bunching in a way, that makes it look very much like crypto has in the past.

It does feel to me like—I know it’s not dying—but at least in the conversation, in what people are focused on and in our attention spans, the long-term fundamental investor is struggling more than ever. The LPs are also struggling to handle the volatility. This is something that crypto funds have heard for a long period of time: “We like the fact that tokens potentially give us earlier liquidity, but this is a venture book. It’s not supposed to go up by 50% and down by 50% in a quarter. That makes no sense to me.”

Now you’re seeing it across a bunch of different asset classes. People don’t know how to stomach the volatility yet, and it’s creating incentive structures that I think are quite perverse. Then we talk about Leopold, right? He supposedly sold his entire $14 billion of notional value in a single block trade because he was 4x levered, going into a period when memory stocks had fallen 50% in 2 months. It’s crazy.

In 2 months? In a month.

Yeah, in a month.

My current portfolio, earnings, everything—the entire thing—is just down 50%. Today’s obviously a bright day there. I remember—we never experienced it, I think—but I’ve heard from a couple of fund managers that a lot of LPs were like, “Just don’t send me quarterly statements, please. It will make my life extremely difficult when I go to a committee and do a portfolio review.”

“In a perfect world, just send me something in 4 years, 5 years, 10 years. I don’t want to see the intraday vol.” It just tells you a lot about how difficult it is to underwrite crypto and other high-volatility asset classes.

Goldman has some interesting stats where they’re saying the momentum index is worse than in 2008, just to put it in perspective. We’ve had a massive deleveraging. All the different banks are obviously commenting. JPMorgan said that 90% of all the leverage in the system has been rinsed out.

A lot of this conversation we’ve had in crypto is that we look at portfolios—

Exactly like we were people tweeting about crypto.

So, anyway, this is a word of encouragement. We’ve always said here—and I’ve felt it—the lines between crypto and TradFi are increasingly blurring.

Yeah.

If you’re a trader or an investor in crypto, you’ve had to be a venture capitalist and a high-frequency trader simultaneously. I do think the world is just becoming more like that.

Ken Griffin said something interesting. He was probably way ahead of his time in being super focused on short-term, tight-risk books and pods. Now he’s saying the opposite: He’s more interested in long-term fundamental investing. He’s always, in my opinion, been ahead of the curve, and for someone like him to say that tells you a lot about the opportunities.

Again, don’t necessarily use that much leverage, or any at all. Expect a lot of vol. If you want to get creative or crafty—

HODL. Figure out—you’ve got to HODL. HODL your micro.

Or figure out a way to make money on vol. That’s what I think a lot about. But it is crazy when you have Meta down 9–10% today.

6. Why Market-Neutral Funds Win

The best-performing liquid funds in crypto, if you take the whole last 5 years, are all delta-neutral funds. All of them.

Delta-neutral.

7. Robinhood Chain’s Early Traction

Yeah, because of the volatility. All of those guys are now delta-neutral in traditional markets.

It’s funny you mention that. I was having lunch with these guys who started in London and, I heard, managed close to $4 billion. They’re all market-neutral. I was like, “Whoa, I’ve never—”

Heard of this? Oh, is this QRT?

No, no. I think the guy used to work at J.P. Morgan and Goldman in prop trading.

He started doing crypto and then slowly totally ditched TradFi strategies and focused on arbitrage opportunities—market-neutral in crypto. They did say this year has been a bit more tough than prior years. 2022, 2023, and 2024 had been much better; this year has been a bit more difficult.

But I was blown away. To clarify, they do have a—like we used to have, or still have, when I was there—a stablecoin arbitrage fund. There were marginal deviations between DAI and USDC, and Curve has come around. But I asked them how scalable this strategy was before knowing the AUM. They were like, “No, we have $4 billion.” I was like, “Wow.”

No, no, again, it just tells you that, to your point, the crypto market makers—just arb funds. I mean, Goldman, I think, has a pretty big strategy, right? A couple of the TradFi players. Citadel’s probably doing a lot. Jump’s doing a lot here.

It feels like, to your point, market-neutral funds and lawyers are the best-performing people in crypto. Everyone else is struggling.

8. Institutional AI Fatigue

I will say that maybe it’s worth noting here on AI versus crypto LPs. I’ve been on the West Coast this week, and we’ve had a bunch of LP meetings this week. One of the things that I’ve started to hear, probably for the first time, is a little bit of AI fatigue from the LPs—big institutional endowments and funds of funds.

Three or 4 months ago, there was still an insatiable appetite to get exposure to these AI funds, or to the generalist funds that are doing a lot of AI and are very good. It’s unsurprising, I guess, that what’s happening in the public market is going to affect that. But we’re starting to hear a little bit of, “Oh, maybe this got ahead of itself. There was way too much deployment in the first half of the year and late last year.”

I heard one LP who is not that exposed to crypto at all—only in a couple of managers—compare it to the 2021 crypto cycle and just say it simply: valuations are way too high. A lot of these companies are just software companies that have changed the business model. It’s not SaaS anymore, but it’s still just software, and there’s no reason it deserves a 100x multiple.

The OpenRouter $10 billion thing, when it’s a $140–$150 million revenue business, is insane to me, but that’s a—

Did they confirm $10 billion?

That was just a rumor, so I don’t know what the actual deal is.

More on that, yeah. I don’t know if they would exactly be revenue or cash flow.

9. Ondo’s Private Chain Bet

Yeah. It’s interesting because you might have heard me say on the podcast earlier this year or late last year that I feel like there’s exactly what you said. I do think there’s more nuance to the discussion around how AI is bucketed, like crypto was back in the day. Crypto was a massive bucket, but you can make the case that there are names that are very undervalued and underpriced, and there’s stuff that is extremely overvalued.

I think AI is in that camp, too. You can slice it into hyperscalers, neoclouds, memory names, training and inference. There are many different parts of the stack, and everyone’s trying to figure out what the bottleneck is.

Gavin Baker is worth following. I think he has some of the best takes. He said there are pockets of AI that are extremely overvalued and hard to justify, especially in the private markets.

I understand that researchers are now rock stars and are getting paid a billion dollars to join Meta and stuff. But if you’re a traditional venture fund, I struggle with why you’re investing in it. You see these rounds that close at $3 billion, then jump to $8 billion, and then go to $16 billion. What has changed in those 3 months other than you missing out on investing in the first round? We saw that a lot in crypto, right?

That feels very hot to me. I have my own opinions on who’s going to accrue value. I think you had mentioned that it was rumored that Leopold, who has a position in Anthropic, had sold a lot of that. There’s a big discussion about open-source models.

**Rob**

That also sounds like crypto. That’s also a crypto thing.

**Santiago Roel Santos**

Crypto. Where does value accrue, right? We’re going to talk about Robinhood Chain and Ethereum yet again. Ladies and gentlemen, we can’t escape that.

I do wonder. I think the AI trade is still very much intact. There’s probably more rotation within AI and where you put your money behind. I’m not sure Anthropic at a billion is attractive, even though I’ve heard the ARR just keeps crushing.

I see way more risk in being long Anthropic. If open-source models are progressing quite a bit, it does raise the question: are you still going to be able to get that much profit margin on your token spend? I don’t know. But all of this is starting to rhyme with crypto.

**Rob**

10. AI’s Token Economics Problem

We have a portfolio company that’s still in stealth, and I think it’s going to come out of stealth next week, which has a router product that is similar to OpenRouter. OpenRouter is probably its only other competitor, but it has much better benchmarking in terms of actually optimizing spend. If you want to optimize your token spend.

One of the things they’ve started to see is that they serve a lot of agentic businesses—people who are building agents to run their businesses fully, both for internal use cases and potentially consumer or direct-to-end-customer use cases. You’ve seen Coinbase come out and talk publicly about how they built an internal router for their own context, which helps them optimize costs. You saw Ramp come out and launch a product. It’s sort of a more closed-loop product, but it’s trying to do a little bit of the same thing.

It’s very clear to me that, with how much better and cheaper the open-source models are getting, and how the vast majority of workflows don’t actually require whatever the frontier model is, token spend is going to start to become a much bigger topic of conversation. Being able to charge people for the frontier models is a real premium. That’s probably going away in the vast majority of industries within the next 12 to 18 months.

**Santiago Roel Santos**

11. peaq Ad

Robots and machines will outnumber humans on chain and Peak is how they get there. Peak OS enables any machine to do business on any chain. It turns machines into autonomous actors and [music] liquid assets with ease, giving robots access to capital and the ability to compound their [music] value. That means funds, allocators, and institutions can underwrite, finance, trade, and route capital to machines the same way they handle any other [music] asset. Millions of new potential consumers for any Web3 service from compute to storage to anything a machine may need without a human in the loop. Head to peak.xyz [music] or click the link in the show notes to find out more.

12. AI Token Costs and Model Routing

It reminds me of the NFT craze, right? You’re paying $1,000 to settle a transaction.

**Rob**

Yeah, TBD. Obviously, a lot is moving quite quickly. But when you guys made that investment in the router company, I’m curious if you’re able to share how much growth that company is seeing and how much demand there is, because we hear some companies that are really focused on token spend and capping that, making sure the productivity is there, and others that are just token-maxing. What’s the defensibility of a routing product like that?

**Santiago Roel Santos**

I don’t want to get too deep into it. It’s one of their products in a suite of products, which should provide more of a moat because they do a bunch of stuff around enabling gigantic businesses generally, including payments, which is where some of the crypto angle comes in.

The growth has been a lot in the last few months. They’ve got some initial pieces of data. For one of their clients who runs a ton of spend, they’ve reduced costs by 80%. That is an incredible number. It’s only 1 benchmark right now, so it’s not to say that the benchmarks you’re seeing for some of these other people are nearly as drastic as that.

But it’s clear to me that that’s where the market’s going, and both Anthropic and OpenAI are going to have to figure out net-new business models. The CapEx spend that goes into training, and all of the work that you need to do to launch one of these frontier models, is really incredible.

How do the economics work? I think that’s the question everybody is still asking themselves. But from an Anthropic standpoint, to your point, when that goes public, I don’t know anybody who thinks it goes public for less than 1.2 trillion. There’s still a ton of demand there.

**Rob**

Yeah, yeah, yeah. It just feels the most exposed in this shake-up, in this world where there’s increased scrutiny for token spending. At that valuation, I’d rather own other stuff if I’m going long AI, but I’d rather own other stuff.

**Santiago Roel Santos**

I’m not coping. I could have said $200 billion, maybe $100 billion, but obviously the ARR number keeps growing. As long as that continues to happen, I don’t think it’s crazily valued, to your point. We might be in a world where their ARR is, what, $150 billion?

**Rob**

Yeah, I think there were some numbers that said they were close to—as of July—over, close to $150 billion. I think I saw a tweet about it.

**Santiago Roel Santos**

Yeah. That’s crazy.

**Rob**

But we’ll see, right? I don’t know.

**Santiago Roel Santos**

And I think this is, to put it in perspective, the fastest-growing ARR in the history of mankind.

**Rob**

Oh, yeah. It’s incredible.

**Santiago Roel Santos**

By a country mile, it’s just incredible.

**Rob**

The fastest-growing companies we used to see were companies that might grow 8 or 9 times in a year, and they would do that for 1 year, and then it would slow down—companies like Ramp and Rain and stuff like that. We’ve seen tremendous amounts of growth in these companies, where they’ve put big multiples on their business in a short period of time.

But what we’re seeing on the AI side from the foundational labs is so far and away unlike anything anybody’s ever seen, which is part of the reason you get this mania and also part of the reason that you might have a very big correction. But that doesn’t mean that the long trade, if you have the right long-term viewpoint, isn’t the right one. You just have to stomach the volatility in between.

**Santiago Roel Santos**

Yeah. All right. Should we pivot for a brief second to Robinhood Chain?

**Rob**

Is that your biggest holding? Is that your single biggest holding?

**Santiago Roel Santos**

Not at all. I have some exposure, but no. I used to own it quite a bit; now, it’s definitely not that. Guess what it is. It’s—

**Rob**

AI-related. It’s not Robinhood. To put it in perspective, they launched a couple of weeks ago, and there’s been a ton of activity there. There’s a tweet by Frank, and I’ll just go through it: Less than 1 month after launch, Robinhood Chain has become the largest network by tokenized stockholders, surpassing Solana, BNB, Ethereum, and Base. Distribution matters.

There’s some nuance to that, which is that the way you count users is tricky. We’ve gone into this discussion around crypto: a wallet is not necessarily a user; a user can have multiple wallets, and how you count them is different.

**Santiago Roel Santos**

Also, by the way, active accounts have actually trended down after that initial bump, so they’ve been down the last week or 2.

**Rob**

But it’s pretty impressive, I have to say. I don’t know if you have any particular take on that. I just found it pretty interesting to see that a lot of the attention is there now.

**Santiago Roel Santos**

Listen, I think very clearly, number 1, the Robinhood Wallet product is good. Anybody who hasn’t used the Robinhood Wallet product—it’s different from Robinhood itself—it’s a good product. Obviously, they do a very good job of building consumer products, things that are easy to use and that people like to use.

They’ve done, I think, a really good job on distribution and a really good job on marketing. Instead of building their own closed-loop, net-new protocols, they brought in Lighter, Morpho, Ethena, and Uniswap to come in and be kind of the cornerstone, but still distributed them the same way they’ve distributed the other products. Obviously, that’s the smart move. It’s been really, really fun to see.

I expect we’ll see, after this initial wave of excitement, a little bit of a trough, and then we’ll have to see what happens over time. I’d love to see them bring the Robinhood Wallet product together with the Robinhood core app in the same way that Coinbase eventually did that. That would be a humongous unlock for the chain.

It’s very clear that they see this in the same way Stripe and some others have embraced crypto: as a core growth engine for them in the future. I’m excited to see what continues to happen. I do think people get a little too excited, probably a little too quickly, about the initial thing. There were a bunch of people trying to front-run the meme coins, but the same thing has happened on other chains. That’s all fine and dandy, but once this thing settles in, we have to see how growth goes over time. The same thing happened to Base, by the way.

**Rob**

13. Prediction Markets Eclipse Crypto

We’re going to talk about publicly traded companies that have a crypto business—not just Circle, Coinbase, and Securitize. Really, Robinhood, in my opinion, is the most forward-leaning traditional company moving into crypto across prediction markets, its own chain, and so on. I’ll spend a minute on their earnings just for people: They reported $1.3 billion in earnings, growing 32% year over year, so pretty strong. They beat consensus estimates and whatnot.

I think the average trading per account has come down quite a bit. Maybe there’s some seasonality to that. Prediction markets did $156 million—

**Santiago Roel Santos**

Bigger than crypto.

**Rob**

Yeah.

**Santiago Roel Santos**

No, right.

**Rob**

Exactly.

**Santiago Roel Santos**

Crypto did $100 million. And so that volume, starting from a low base, is 10 times year over year. Like—

**Rob**

The $156 million is revenue, correct?

**Santiago Roel Santos**

Revenue, yes. Not volume, obviously.

Yeah, I mean, that’s to be expected with the World Cup. I’ll be interested to see what that is in July and August.

Fair point. World Cup was the biggest contributor there, I think. Again, interesting—we’ve talked about this before. Robinhood drove, I think, over half of Kalshi’s volume. Then they bought NYX and—

They essentially did a— they bought NYX, or the old LedgerX business. And they did a joint venture with Susquehanna to launch a thing called RothEra. RothEra is their prediction-market exchange, but they started to move some of their Kalshi volume to RHDX. The expectation—

Right.

The expectation is they’ll continue that over time. I’m sure they’ll still do the best pricing, but you’d expect more and more over time that they want to own that full stack.

Yeah. Exactly. Not to get too much into it, but I feel like that’s pretty indicative of how traditional firms are going to start implementing things that are working in crypto. If you have distribution, you launch your own chain, you launch your own prediction markets, and you route volume away from Kalshi or Polymarket, you know, in-house.

I don’t know if you saw this, but the CME right now is suing the CFTC over perps. Then they came out in their earnings call, and they were actually somewhat positive on perps, which very clearly tells you they’re talking out of both sides of their mouth because they’re doing something as well.

They’ve been very upset with the prediction-market exchanges. They came out and said yesterday that the CME is probably going to launch sports-linked prediction-market contracts, or options contracts. It tells you that all of the incumbents—you try to get in front of the innovation; when you can’t get in front of it, when you can’t stop it, you embrace it. That’s what we’re going to see from everybody.

Yeah, but look, it’s interesting. Robinhood—just another point—it’s been 4 weeks: $12 billion in DEX volume and 100 million transactions. They’re reporting a holder count of 328,000. Maybe we’ll revise that number, obviously, but they’re the number 1 chain by RWA holders. That’s Frank’s tweet.

Some people might contest that, but it’s still impressive in the sense that they’ve sort of leapfrogged every other crypto-native operation out there with their distribution. Which, to our earlier point, means the Robinhood user is the prime user for prediction markets and just on-chain activity.

My view is Robinhood is probably the most important company to watch because it will tell you if flows are coming into crypto, the sentiment—if it’s true that there’s AI fatigue and people now want to trade crypto or NFTs or whatever, you’re likely going to see it first on Robinhood. Not Robinhood earnings—Robinhood Chain and the activity there. Before, it would have been BNB Chain, which I think we haven’t talked about or don’t talk about too much.

I mean, BNB still has those—Binance is still one of the biggest fintechs in the world in a lot of localities.

Yeah. So, anyway, one last point: They launched tokenized stocks, obviously not available to U.S. customers, so everyone else ex-U.S. can get exposure to tokenized stocks on the Robinhood Chain, right?

That’s also a metric that I’m really closely following because it is the true expression of what Hyperliquid has done well, with HIP-3 and stuff like that. Solana’s done it, too. I’m quite curious to see how much activity there is, and we’ve talked about tokenized stocks at length here.

That’s Robinhood. Similarly, there’s another great tweet. Because Jason’s not here, we’re going to reference The Block.

14. Ethereum L2 Activity Collapses

So, obviously, Robinhood Chain is built on Arbitrum, which is an L2, so it’s in the Ethereum ecosystem. But that’s a bright light and a lowlight for every other L2 out there. I think the total value locked, the TVL, in all other Ethereum L2s is back to roughly $5 billion. The last time we saw that was in 2023, so it’s a pretty big reversal. I think all the other L2s have found themselves in a difficult spot, to say the least.

What do you make of that?

I don’t know. I would like to see whether there’s been an actual rotation away from these, and those maps that show where the flow is. I don’t have it in front of me, but it’s not surprising. I think we’re at a point where there’s nothing really truly exciting happening.

You have the Ansem bull coin in Solana, which was the only thing that gave a little bit of life. But this chart talks about optimistic rollups, ZK-rollups, validiums, plasma, and state channels. It’s just not interesting for a retail user. You have days where there’s 10-plus; there’s super-high volatility in traditional markets.

Korea has always had very high crypto penetration. I think it’s the highest in the world. These guys are not trading crypto right now; they’re trading other stuff.

Yeah, I will say The Block’s numbers are a little bit odd, because I think it’s L2BEAT that also said the combined TVL for everything in the Ethereum ecosystem—not including some team-controlled tokens that they didn’t think were particularly real on Arbitrum—was around $33 billion this week as well.

I don’t know if you’ve spent any time thinking through, or if you looked into, how The Block is coming up with that methodology, but it is lower than what other sources are reporting.

Yeah, I will say that the trend is all in one direction. If you look at DeFiLlama or Blockworks, the trend is in one direction.

I think the other thing that we’ve talked about a lot—I don’t want to talk about it on this pod—is that it’s been pretty brutal this year in terms of hacks. You had Kelp and a number of others. There’s increased awareness that we’ve had more hacks this year.

I think the first half of this year is the highest in crypto history: 212 exploits and over $1 billion in funds lost. That’s 3 to 3.5 times more than last year. North Korea’s Lazarus is half of that, or more than half of that. Kelp and Drift were the biggest, right?

A lot of this has been social-engineering attacks and bad risk-management policies. I don’t want to say that smart contracts all of a sudden just became more vulnerable and fallible and all that. I think a lot of these are social-engineering attacks and bad risk-management policies.

Nonetheless, when you combine low yields on-chain, increased hacks, and AI becoming the momentum trade, I’m not surprised. I think those 3 things really explain less TVL. The basis trade is not as interesting anymore, right? The rates are—where are they at now? They used to be 10-plus; now they’re 3% or 4%.

It’s just not enticing.

It’s over 4% now, but they’re doing a lot of RWA stuff now. I mean, it’s still well above SOFR. But it is hard when Robinhood Chain right now is on USDG to give you 7%. Clearly, they’re just subsidizing.

X Money is giving you 6%.

Like 6%, yeah. Yeah.

Yeah, yeah. Are you guys—I know your partner has talked a lot about on-chain hacks and is pretty good at diagnosing that. Internally, as a firm, is this a major concern for you when you think about the viability of crypto and deploying capital and advising founders? Or is it just, “Look, guys, have good opsec and you’ll be okay”?

Sorry, I’m going to make one last point on the TVL stuff before we go into this. We have 1 portfolio company that runs a lot of these vaults for a lot of the institutions. We had the board meeting a couple of weeks ago, and TVL for the whole industry—for all products—was down a lot on the quarter.

But it does seem that when people talk about TVL numbers, about half of that number is actually just ETH and Bitcoin coming down, because a lot of these TVLs are in—

Yeah, it’s denominated that way. I think it’s worth keeping that in mind when you talk about these numbers.

Okay, so that’s the last point I’m making on that. On the point about security, listen—

Wait, wait, wait. Before we go there, have you guys looked at—to me, the most important metric is active users on-chain. I think a16z came out with a report last year, which you might have also heard me mention quite a bit. They slice it in their own way, but it’s less than 100 million active users on-chain. They give a range of between 40 million and 60 million.

I think that’s worth keeping tabs on, whether it’s a16z or whoever. To me, that feels like the most important metric. TVL, to your point, can just be related to price action, but users, in my opinion, are the most important metric, because I don’t think we’ve onboarded—

And I probably lost quite a bit of active users.

Yeah, well, I think the TVL and the users are also related to the question you asked about security. We actually saw that the vault products struggled a lot after April and after the Kelp hack. Then there was Stream Finance, there were a few others, and Drift. There was Ostium recently on the OP side.

People have started weighing the risk-reward of being on-chain a little bit differently. At the same time, I do expect active users to pick up with the usage and proliferation of a lot of tokenized assets being distributed by people like Robinhood or other neobanks, where they’re abstracting away the being-on-chain part.

You have these 2 countervailing forces: a lot of the crypto-native activity may be coming down, while more retail-oriented users coming through traditional distribution modes are starting to grow. It will be interesting to see how those interact.

Yeah, yeah. I do think applications benefit much more than Arbitrum and Ethereum, the infrastructure providers. Applications like Uniswap and others benefit much more. Ethena and some of the others, right?

Just shifting gears a little bit, what do you make of Ondo launching—I don’t know that much about it, but I did see that they launched this private chain. We’re back to being curious if you guys have looked into that and what your thoughts are on their design choice to do that versus being public.

Yeah, I haven’t looked that closely at it, but they announced, “Okay, we’re going to do an institutional L1,” and I think they’re saying it’s going to be private. A lot of the language sounds a lot like Canton, right? Basically, they want to compete with Canton.

Canton’s working really well. We’re going to double down on our effort that we announced earlier this year, and we’re going to be fully institutional. We’re private.

Yeah, so it’s like, “It’s going to be verifiable like a blockchain, but it’s going to be private. It’s also going to be non-custodial, but you won’t necessarily be able to run an independent validator.” They didn’t give a ton of details about it yet, so we’ll see what it actually ends up looking like.

What’s very clear is that for the Ondos and Cantons of the world, which are focused on how to serve the needs of larger institutions trying to use tokenized assets—and that are potentially worried not only about a lack of privacy, but also about open-source hack risk—there’s a much bigger conversation today around how to serve those people appropriately.

I have a perspective that this is a little bit like—I don’t know if you remember a decade ago, but we had the Hyperledger and Enterprise Ethereum and—

All those years.

Yeah, we—and, yeah, this feels a lot like that conversation. People thought that completely enterprise-private systems didn’t really work, but maybe something hybrid would work. Maybe that will someday, in the future, end up being public and permissionless. But it feels like you just need to pull these institutions along.

I don’t actually think a lot of this infrastructure that’s being talked about right now—which is sort of just a database that looks like a blockchain, but doesn’t have a lot of the capabilities or the ways in which blockchains are actually useful—is coordinating trust across a bunch of different counterparties.

I don't know. We'll see what happens.

Yeah, in fairness, it always goes back to execution, verification, and settlement. Those are the 3 functions of a blockchain. In fairness, I think if we had the Ondo team here—and if they want to come on, we're happy to have them on—their argument would be that we're keeping the settlement public, but the execution, and maybe the verification, are private because that's what institutions want. You can't have order flow being public, and it's a fair argument.

I think that was their whole messaging: it's still public to verify, I believe, but the execution itself is very much private, and that's what institutions want. We have to listen to that, and we're going to build in that direction. It would be interesting to see a discussion between Canton, Yubal[?], and Ondo, and maybe Mike Cagney or someone from Figure talk about what they're doing. Maybe Securitize too. I don't know.

Obviously, on the Figure side, they've tried to make Provenance completely open.

And permissionless, right? They just haven't been able to get other people to build there. Listen, I think what's very clear is that the direction of travel is tokenized assets. Everything's going to be tokenized. People are coming on-chain in one way or another, but whether that is Ethereum, Base, Avalanche, Monad, or Solana, or it's Canton and a private network, nobody knows.

There's a lot of competing conversations happening right now. There are also a lot of differences of opinion within the organizations themselves. There is an ideological fight happening here: what serves my needs versus what I think will serve my needs in the future, in a different regulatory environment.

Yeah, yeah. TBD. It goes above my pay grade. It sounds a lot like a dark pool with some sort of blockchain settlement, but they'll believe this architecture is a secure enclave and that these are trusted execution environments. It's a dark pool and that, but I don't know that much about it.

TEEs—I mean, there's a lot of work being done around TEEs right now. This isn't new, but I think there's more excitement around them than there has been in a while.

In contrast, you hear Ondo talk about how institutions want the ability for their order book not to be public and exposed. Then you have Robinhood saying, "Yeah, we're going to launch a fully public, permissionless chain," a free-for-all kind of thing. It's a different customer base: one is more retail, and the other is more institutional.

You're definitely seeing more willingness to take risks among the people who are direct-to-consumer than among the people who serve institutions, and the institutions that also serve institutions. It's very clear that Robinhood has innovated in a way that has hurt the Schwabs of the world and the traditional wirehouses. That doesn't mean they're not right about what the world will actually look like in the future.

15. Stablecoin Adoption and Crypto VC

Yeah. And, by the way, there's a world where both are equally successful. Robinhood is clearly monetizing distribution, retail, and activity there. They've done that with Citadel, with flow and everything like that. Now they're doing their own chain, and Ondo is a totally different animal with institutions. Interesting. What else is there on the agenda?

I mean, probably just the end. I don't think there's much else, so that's probably most of what's happened this week. The AI conversation has been dominating, both from the open-weight conversation and now what's going on with situational awareness.

The last thing that interested me was Visa announcing its stablecoin platform. It was essentially just a way for them to say, "We're going to do value-added services for institutions that want to use stablecoins as part of a wallet stack that we own, potentially doing direct settlement with us in stablecoins for your card product, and potentially doing non-card, non-network stablecoin transfers as well."

We've seen this growth in what I would call the forward-deployed model in stablecoins, in a way that you've seen work really well in AI. We have a portfolio company called Velocity that's doing quite a bit of this on the institutional side.

We're a year past GENIUS, and we're starting to see every institution say, "I think this is interesting. I need a strategy." That's been happening over the last year, but they're still a little unsure of how to put the pieces together. Visa understands that, and so I do think this forward-deployed model causes another wave of adoption—a quickening of growth in that market. I'm interested to see what happens because Visa has obviously been very forward-thinking on stablecoins for a long period of time. That was exciting to me.

Yeah, definitely. There's obviously one last tidbit here—not to end on negative news—but I think it was also a record quarter, or the lowest amount of VC activity and new deals in crypto.

Yeah, although the amount of money put to work was still holding up well. It's just that the later-stage deals have happened.

Later-stage deals are getting more attention, including from more fintech-focused funds, but especially on the earlier side. It's pretty tough. I've had conversations with companies that have raised a Series A and are raising the B, and it's really, really tough out there.

It's not just in crypto. If you're not on the AI train, it's really tough. Let's not forget that software in general and other types of businesses are really tough. I think the only things getting funded in a very meaningful way and sucking the air out are defense, robotics, and AI-related businesses—anything involving frontier models, edge AI, chip design, and stuff like that. Those are massive rounds, and everything else has been very tough to fundraise against.

Listen, VC is a herding culture. People are all chasing the same things all the time. It's refreshing when I talk to 1 or 2 LPs who say, "Now is actually the right time to invest in crypto." I was like, "You're speaking to the choir here."

I think we are coming back a little bit, and so I expect—

How many of those have you heard?

I heard it from a very big allocator. I've heard it from a couple of very big allocators this week.

Fair enough.

To be very honest, the people I'm talking to are the champions—the crypto champions—at the firm. When they go back to their investment committee, what is the investment committee going to say? That's still up for debate.

It would be quite alarming. I understand new money not coming in, but if you're raising another fund and existing investors don't at least put in however much you distributed, that is a cause for concern. I don't have intel on that, but—

The biggest thing that gets asked every time is, "What are your existing investors doing?" If you have less than 80% to 85% from your existing investors, it's a conversation every time.

Correct. For people who don't know, if 85% of your investors aren't doubling down or at least putting money into your next fund—

At least the same dollar amount in your next fund, yeah.

Do they really focus on the dollar amount, or do they just care about whether they're in the fund?

They focus more on whether they're in the fund. It doesn't have to be the same dollar amount, but if the existing investors are saying, "I want to support you, but I'm coming down," then usually that means you have to bring the fund size down.

Right. Interesting. Do you feel that we've seen the Founders Fund of the world and Paradigm pivot a bit into frontier tech?

I mean, Founders Fund just did a nuclear deal this week.

Yeah, I know. They're just frontier tech, as I think of it—anything frontier: robotics, AI, small modular reactors, all this stuff. Is the idea that, for crypto to be an investable universe that can create venture-like returns, you just need to bring your fund size down until we see major wins?

I think if you're going to do crypto only, it depends on how you define that. If you define that as fintech and a lot of these fintechs that are using stablecoins, you can still deploy a $50 million fund like ours. I think if you're going to do a $1 billion or $2 billion fund, it's really hard to be very deep crypto-only right now.

Andreessen is obviously trying to do it with their $2.2 billion fund.

They already raised it.

Yeah, and they also do a lot of liquid in that fund.

So they buy and hold Bitcoin and even Solana and all these things, whereas in most of the smaller VC funds, they’re either more explicitly one or the other. I think fund size—we’ve seen it for almost everyone—came down. It came down for a little bit for Paradigm. It came down for Andreessen. It came down a little bit for Haun. I think it came down for Framework. It came down for Variant.

Still big for Haun. I mean, Haun just continues to—

A fund that I don’t know if—

Fundraise, and then, obviously, they were one of the first checks or one of the biggest shareholders in Arbitrum, which has been an incredible return for them.

Incredible return, yeah. That’s right. That’s right.

Good stuff. Shall we go to conviction of the week?

Let’s do it.

Nice. What do you have?

I’ve been watching the show Lucky on Apple TV, which stars Anya Taylor-Joy. I didn’t know anything about it. The one thing outside of investing that I spend a lot of time on is film, but not as much TV. I saw it one day while scrolling past and started watching it. It’s an interesting caper, a mobster show where she’s running from the police because she stole some money. I’ve been enjoying it during my week of travel, so I’m going with that.

Nice. I got 2: the Nadal documentary on Netflix. I love anything related to sports. I’m a terrible tennis player, but I love watching tennis. I had no idea he had a foot deformity issue from very early on, and that he was always working and struggling with it. It’s pretty amazing—mind over body—and really interesting.

It’s a couple of episodes, well worth the watch, in my opinion, if you like tennis or just like the mentality of the relationship he had with his coach, who was his uncle. It reminded me a lot of the movie Whiplash. I don’t know if you’ve seen the documentary. Yeah, Whiplash is—I love that movie. It takes two to tango, right? It’s not just the professor; the student clearly wants to get pushed, and I think you see that very much in the Nadal documentary. So, I really enjoyed that.

The other one is—I read Sam Walton’s autobiography. Really, really good. Walmart is obviously a phenomenal company, and what they’ve done is really interesting to see: that evolution and how he thought about building it and stuff. So, yeah, continuing on the autobiography train. I really enjoyed that one.

Cool. All right, July 30th. We’ll be in August the next time we record. Are you taking some time off? Are you going to be here, or am I going to have to do this solo?

I took some time off a couple of weeks ago. I’ll be around for August. I’m okay. I get so much joy out of working.

Truly. That’s the thing: no one ever retires.

That’s awesome. Okay, well, thanks everyone for listening. Have a great weekend. We’ll see you in August. It’s crazy.

Nothing said on Empire is a recommendation to buy or sell any investments or products. This podcast is for informational purposes only, and the views expressed by anyone on the show are solely their opinions, not financial advice or necessarily the views of Blockworks. Our hosts, guests, and the Blockworks team may hold positions in the companies, funds, or projects discussed.