AI末日：加密货币准备好卷土重来了吗？

Are we at peak AI, or is this just another speed bump on the way to new highs?

When it comes to the whole AI trade, the charts just look like a standard altcoin that went parabolic and has now broken the trend.

There's a lot of people losing their minds with this drop. It was pretty heavy, but I'm still in the camp of holding for a few years. I don't think this is it.

Every 2 to 3 weeks, you have a chance to make $100K, but every day you have a chance to lose $100K. That's kind of what the market is like right now. Those chances are coming back now.

I think the big headline, and the place I wanted to start, is what seems to be a bit of an AI apocalypse, depending on who you talk to. There have been huge sell-offs, especially in some of the higher-beta memory names. The NASDAQ is holding up decently well, but there's this feeling that maybe we're at peak AI, or that this is just another speed bump on the way to new highs when it comes to the whole AI trade. I want to get your guys' general thoughts on the markets.

I know Jose probably has a lot of good thoughts on at least the memory aspect of it—the DRAM and HBM sell-off, and what's going on there, including who the big winners and losers are. Generally, when we recorded this show two weeks ago, I was looking at everything and thinking, “I don't know. Indices look like they're distributing a bit.” Crypto is obviously just doing crypto things, but I was in the camp that I didn't want to be taking any new longs, adding risk, or taking leveraged long trades on really anything at the moment.

Since then, we've sold off a bit. Some of that is probably due to further escalations between the U.S. and Iran and the conflicts going on there. We've had renewed tensions and strikes, or whatever you want to call them, which have spiked energy prices. Equities have started selling off again across the board.

Some AI names look pretty decent. A lot of them look like standard altcoins that went parabolic, broke the trend, and are now bleeding out toward something that more resembles fair value. TL;DR, I'm still relatively cautious. I think there's still some downside in the indices, but I think we're getting close to levels where I would be interested in buying equities.

I still think crypto looks pretty terrible. I think Bitcoin still needs some time. Like I said, 48K could be the bottom. We could still go there. Realistically, it's less of a price thing and more of a time thing for Bitcoin at this point. I'm still pretty cautious, but I'm curious to hear Jose's thoughts on the memory sell-off and what he thinks it means. That's my high-level take on the markets and where I'm positioned: I'm not really positioned. I just hold core positions. Jose, are your bags finally selling off a little bit?

They are. Memory is a hard one. With Kimi's new update last week, they introduced a new architecture that drastically reduces memory usage and especially moves things to HBM and away from commodity DRAM, which is what a lot of companies like Micron and others sell.

I do think memory is a hard place to spend time on. I've spent a lot of time trying to understand it over the last 2 weeks, and I ended up putting it in the too-hard bucket. There are just too many factors moving around, so it's hard to figure out what's going on. I still hold the positions that I bought. I think I'm about at break-even on them, or maybe up a little bit from where I got in.

I was really planning to buy a dip, and I think Bubble Boy actually saved me on that. Shout-out to Bubble Boy for some of his posts on the Kimi architecture. I think the more interesting area, or at least the area where I've been bidding more, is the neoclouds in terms of high-beta AI exposure. That's the more interesting bucket for me.

The majority of my exposure is still very boring: the Magnificent 7. I'm trying to add more beta with this drop to some of these neocloud names that I've been looking at for a while, like Nebius, CoreWeave, IREN, and a bunch of others. I just think it's an easier business to understand and an easier spot to be in, where data centers are going to be scarce for a while.

These companies, especially CoreWeave, are leveraged bets on data centers and AI demand. There's a lot of risk built into it, so it's not something you should size as a giant position. I think 65% of CoreWeave's revenue is tied to OpenAI, for instance. But I do think it's one of those businesses that benefits from the cost of intelligence going down in general.

Whether it's OpenAI, the big labs, or open-source models that start to gain share, I think the neoclouds do pretty well regardless. I'm pretty bullish on them, and I've started to buy some of those names here. The Magnificent 7 are generally pretty unbothered, actually.

For me, there's a lot of people losing their minds with this drop. It was pretty heavy, but I'm still in the camp of holding for a few years. I don't think this is it. I still think we have much higher to go as we climb the wall of worry, and seeing the reaction has made that more obvious to me.

I think there's a lot of people thinking it's over and a lot of people being carried out. In Korea, I think it was one out of 30 who got liquidated on the first day. I can only imagine what it is now.

The remaining.

Hopefully they all log out of their brokerages and start logging into Upbit.

Exactly. The equal-weight index, which I posted in the chat, Kevin, is making all-time highs, which is super interesting. It goes to show how this carnage and sell-off have really been about AI.

The bubble we live in—and AI names have been so hot that everybody's talking about them—has obscured the fact that a lot of stocks are still doing pretty well. The indexes aren't being absolutely slaughtered. The S&P has obviously done better than the NASDAQ, which has a higher concentration of tech and AI, but overall the stock market looks fine. It's just that the highest flyers have sold off a lot, and I don't know—this was bound to happen regardless of the underlying fundamentals. Things just can't keep going up.

SanDisk is still up 332% this year and has drawn down 60% to 70% recently. That's something we're kind of used to in crypto. I think that's the big thing with all of this, and why I teased the question: when you look at the performance, Micron and SanDisk are still in the top 4 in terms of year-to-date performance across the S&P.

What goes up violently often has pretty violent pullbacks. When you add in the leverage, the Korean stock market selling off and everyone making headlines about it makes sense. There are 2 stocks in there—SK Hynix and Samsung—that make up about 50% of that index. When you have those kinds of returns, you're naturally going to expect pullbacks. Then, when you add on the amount of leverage and the approval of those leveraged ETFs for Korean traders at the end of May, you knew something like this was bound to happen.

The question now is whether they're trading at multiples that warrant buying the dip, or whether, to Jose's point, it's a little tough to discern what the future of memory and the pricing power for these companies will be. Maybe it's better to buy layers deeper—something that's more picks-and-shovels, like the neoclouds, the Magnificent 7, and the big hyperscalers that have already installed compute bases and aren't as dependent on massive capital-expenditure and data-center builds that might not go live for another 1, 2, or 3 years.

How are you thinking about it personally, Kevin? I know you're relatively close to the AI trade, but you're not in the weeds like somebody such as PD[?]. How do you balance that for yourself in terms of taking a higher-level Magnificent 7 bet versus something more targeted on memory, for example?

Yeah, it's a good question.

I try to make it as simple as possible if I can, and so I've done a lot of digging and thinking about where value winds up accruing within the stack, as I think we all have. I've gotten, especially talking to people like Z, more bullish on the neocloud bet, especially over the last few months. I think the demand that's coming from AI, and therefore the demand for compute, is only going to accelerate.

Every time we see one of these big open-source models come out and potentially threaten open source versus closed source, and maybe there's a bit more of a threat to some of the frontier models, the trend toward more intelligence for cheaper, or intelligence per unit of compute—as long as that is continuing to decline in the aggregate—I do subscribe to that thesis that it opens up a ton of new use cases.

Even when I talk to people outside of what our day-to-day looks like—I think all of us do that—we try to talk to people outside of our day-to-day. I've got friends who aren't in AI. They're not investors; they're doing what I'll call normal jobs compared to what we do. When you talk to them about what the capabilities are, it's almost like lightbulb moments for them because they're not following this stuff day to day. They're not even using a lot of these tools. A lot of them are still using Claude and ChatGPT as chatbots or search engines.

There's this one side of the barbell where there's all this agentic build going on and agentic systems. Then on the other, you have the vast majority of people who still aren't really up the curve yet. As simple as it sounds, I just think that the potential demand for this is so much higher than it is today, and the pie is just going to get so much bigger that it's hard not to almost blindly ape the Mag 7 names and the neoclouds.

Maybe take a little bit of something like Micron or SanDisk as higher-beta sentiment plays. I've got a little bit of both. Maybe they're not my strongest convictions, but I don't want to be on the sidelines because I think all of these pieces wind up doing well to varying degrees. I'd rather own the basket than try to be super concentrated, especially as all this stuff is moving really quickly.

I've been asking this question to Jose several times over the last couple of months, maybe even going back to last year. I've asked other people because, again, it's just not the area I'm an expert in by any means. I think for me, I landed on a similar approach.

I'm pretty much just carving out an X amount of my portfolio that I want to go to AI-related stuff. Within that, I just build up a basket. If there's an ETF that exists that covers one part of the stack, great. If there isn't, I'll make it myself with single names.

I'll weight that based on memory. It's not going to be a super-high weight in my AI allocation, but something like hyperscalers later on or neoclouds or whatever. I'm basically building it and taking a basket approach across the entire stack, trying to weight it appropriately and rebalance.

I haven't been able to do any of that because I haven't really built this yet, but it's in the process, and maybe I'll share it when I'm finally done with it. That's my approach as a noob who doesn't really have an informed take beyond what you would read about any of the parts of the stack.

It also makes me feel okay watching this selloff happen without having any real exposure. It's like, “All right, finally. Finally, I don't have to see these charts go up every day for a month and a half.”

It's a little cathartic.

It sucks to see other people win, doesn't it?

Being long more broadly U.S. tech for the better part of a decade, I think that's still an area that, as we just talked about, I remain overtly bullish on. You kind of look for these moments or pockets in time where, if you get large deleveragings, liquidations, or forced-selling events, but your thesis is still intact, that's the kind of golden ticket for a buy-the-dip opportunity. I think we're going through some of that now.

I'm still digging through it. I think the market overreaction to the Kimmy 3 release and all of that—that's still the camp I sit in. Again, I could very well be proven wrong, but I feel very comfortable and still highly convicted in those bags.

1. Can Crypto Benefit?

One of the questions, though, is that the AI trade over the last few years has almost been narrative-hopping, in a sense. We saw this within crypto, and then you saw it move away from crypto and go toward all things AI. You can kind of see that in some of these charts that we show.

If you look back at early to mid-2024, that was when a lot of the AI trade was starting to ramp up. Then you had a bit of a selloff, but you had these big flows into semiconductor ETFs. Semiconductors started to come back into vogue, and then you saw this consolidation. Often, there was consolidation within Bitcoin, things calmed down a little bit, and you started to see BTC run up through the end of 2025.

Then that whole trend reversed. Semiconductors took off, the AI trade went nuts, and BTC fell 40–50%. The question now, and one of the questions I want to ask you guys, is: Do you think a bear market or a continued pullback in the AI trade is a catalyst for, or bullish for, Bitcoin and crypto through the second half of the year?

If this pullback were sustained, would that be a catalyst for Bitcoin and crypto to make a comeback, or are the two unrelated? Do you actually think Bitcoin is the double whammy, where if we get a selloff in U.S. equities and the broader AI trade, correlations will go back toward 1—but to the downside, as we've seen at times in the past? How are you guys thinking about capital flows?

In the past, if you asked me this a little while ago, I'd have been like, “You're insane to think that if AI goes down, crypto is going to do well.” But, again, going back to that equal weight and going back to the actual index values, they're still near all-time highs. It's pretty clear that a lot of the carnage right now is basically in most of the speculative names.

If the stock market kind of chops here and then starts going up, but it's not so speculation-focused—if the S&P is at new all-time highs by the end of the year, but the memory stocks are flat or just up a little from now—then I think you could definitely see a lot of speculative capital come back into crypto.

That money would be saying, “Okay, that speculative trade is over. Everything is still good, the economy is still good, and the whole stock market is still good, but that speculative pocket is over. Let's find the new speculative pocket.”

Obviously, if you were to get a broad-based selloff in the S&P or whatever, that's a different story. In no way do I expect crypto to do well under that scenario. That's the way I would view it.

Yeah, I think that tracks with how I think about it. I do wonder, in the event of a deleveraging event or whatever, and equities forcing things down pretty violently, how badly that would mess crypto up.

Initially, I would have been like, “All right, correlations go to 1. People on high leverage will get liquidated, just like every other Thursday. No big deal.” But I think because the AI trade has been going on for so long, and so much of it has been made available to crypto people on Hyperliquid and Binance, you could pretty much trade all of these names on crypto rails.

I do think a lot of people in crypto have generally stopped trading altcoins and moved over there. Honestly, if you're a trader, why are you going to trade shitcoins right now when you have Micron moving 10–15% in a day? You wouldn't. If you're in crypto as a speculative trader and you're not married to the trenches for some reason, I would think you're probably out there trading these things.

If they get washed out, or if you get a huge further deleveraging from these levels in those names, I think it hits crypto speculation relatively hard—or whatever's left of it—because I think most of the speculators have probably gone over there to trade. I know I've traded that stuff a lot more than anything in crypto over the last several months, and I'm not even particularly active on the trading side of things.

The more violent a selloff, the worse it hurts crypto because I think it wipes out a chunk of the small risk appetite that's still around.

So, I don't know. I just don't want to see a violent washout. But it kind of looks like you get something again. And then if you get a TACO thing, who knows, right?

I think the volatility point is a good one.

There's no reason to trade it.

2. AI Infrastructure vs. Apps

Bitcoin is trading like 1% a day, with 2% ranges a day, but SK Hynix, Micron, DRAM, or literally any of these names are trading 15% on the way up and the way down—15% a day, 10% a day. These are trillion-dollar names. Micron wasn't a trillion-dollar company trading 10% or 15% a day. It's moving more than the market cap of ETH in a day. There's no reason to trade altcoins, and you can clearly see it: you get a big watch list of altcoins and scroll through them, and once you get past the first 5 or 10 names that everybody's familiar with at this point, there's just no activity on them.

Jose, I'm curious. Do you agree with this tweet on the screen that AI is like the crypto trade? Everyone's going to buy the infrastructure layer, but then you actually want to invest in the applications, kind of how crypto has gone?

Yeah, I think that's true. I think that's true about most new technology. Every time, basically, the infrastructure layer starts off as the high-tech layer, right? Every business that is now commoditized and boring was once the frontier—telecoms, for example. There are many examples throughout history.

At the beginning, you can see that eventually it'll be commoditized, but at the beginning it's just really hard. With neoclouds right now, you can imagine that the cloud business will be commoditized eventually, but right now it's actually really hard to serve these models at the speed and cost that people want you to. So there's margin there.

It's also the easiest way to get paid, I think, in these upswings: to buy infrastructure. So far, AI infrastructure has outperformed everything. If you chart AI application revenue outside of OpenAI and Anthropic, it's tiny—definitely single-digit billions—and the infrastructure revenue is hundreds of billions at this point, across the neoclouds, data centers, clouds, and everything like that.

Obviously, OpenAI and Anthropic have inflected a lot, but eventually you do need applications to come in and do this. But it's one of those situations where you can be right and just miss out on most of the trade, right?

I do think the OpenAI and Anthropic trade is interesting. One of the reasons I was interested in Anthropic early on was hearing people make a lot of the same arguments that they made about L1s: “This thing is super-commoditized. There's going to be no margin in it,” and all this kind of thing. I do think this eventually happens, but I don't think it happens yet.

It's just harder to bet on the right applications, right? There are a few that are doing pretty well—the Lovables, the Higgsfields. It's always the problem with the video models.

But it's just harder, you know.

Yeah, I think one of the other things about the application layer that makes this tough, both from a public-markets perspective but more so from a private-markets perspective, is that right now it's where, obviously, a lot of the value of these companies sits.

The cost structure is different from underwriting a traditional SaaS business, right? To add new users, the marginal cost to service a new user under a SaaS business was pretty negligible. Here, you've actually got the more usage your application gets—which is a bullish signal for any type of startup or more mature company—the cost of actually serving each incremental user is a real thing.

That goes back to our conversation around open source getting better, opening up new use cases, making it more viable, or increasing the margin, let's say, at the application layer for these companies to be able to build around. But the infrastructure itself that powers an AI application company is just a very different cost structure, which, at least in my opinion, makes underwriting the application layer still difficult.

I mean, the infrastructure layer is just—I know this is going to be big. I don't know where the value will accrue, but I know that models will need to be served, right? They'll need memory or whatever, if that was your trade. They'll need increasing memory as the context windows go up and you do more agentic flows, and agents work on tasks for longer.

Those trades are much easier to make than trying to figure out what precise application is going to do well. Is it video models, world models, or coding models? Which is actually going to do well?

3. Hyperliquid & Tokenized Stocks

Speaking of stocks and crypto, and those market correlations, tokenized stocks and RWAs are obviously becoming a huge theme within the crypto universe. How do you think this affects things? One of the things I've been a little surprised about—and Jason, maybe you've got more insight into this—is that I would have expected, given how nuts tokenized stocks are and getting that exposure, especially for trading individual equity names, that the fundamentals around something like HYPE would still obviously be bullish.

I'm surprised at the price action more recently. Is that surprising to you, that we're at 55, or is that not even a concern?

I'm not worried about it. If you have a 1-week, 1-month, or 3-month time horizon, you might be worried about it. But I think it's relatively clear what's going on here.

The trend in HIP-3 markets, just like the trend in wanting to trade stocks on-chain with perps, is fairly clear. We've seen—there have been some tweets I could probably send you in the chat really quickly—talking about what's next for Hyperliquid. Obviously, how do they break into U.S. markets? There's this feature that allows HIP-3 market deployers to have an allowlist, or whitelist, for addresses that can trade those markets, which paves the very obvious way for a KYC-type market for U.S. institutions and all of the trading volume and stuff that brings.

Then it opens the question of liquidity fragmentation. I believe there's something that market makers can do to bridge the gap between the permissioned and permissionless markets. Anyway, the trend is very clear for Hyperliquid, or just RWA perps in general. You could maybe lump Lighter into this as well.

If you have a 1- to 3-month time horizon, you're probably getting a little nervous. I could see this coming down to the mid-40s, or even a little bit lower, if we get a violent type of sell-off across equities that spills over into crypto and you get correlations to 1 and a kind of cascade.

But I'm honestly looking to buy if it comes lower. If we get below 50, I'll probably start buying spot again with the cash that I raised a couple of weeks ago. I still think it's one of the only 3 compelling things in crypto to actually own that has a lot of secular tailwinds going for it.

A lot of people have been saying that CLARITY is bearish for it, but I don't know.

What's bearish for it?

But, like, CLARITY. I don't know if that's true. What do you think about them? It's clear that they're going to have a KYC-gated version of it. I doubt they would put KYC on the whole thing, but I'm curious—if they had KYC for Americans, right?

Yeah.

Would that only be U.S. liquidity? Probably. But if it's U.S. and there's enough liquidity in the U.S., that's not that big of a deal if they're isolated. It would probably be segregated from the rest of the exchange, I would think, so that the KYC doesn't trade against the non-KYC.

Yeah, again, nobody knows exactly how it would work yet. But there would have to be segregated books. I think there's a way that U.S. liquidity can't trade against non-U.S. liquidity, but I don't know if that bars market makers from doing both markets. You know what I mean? They could market-make both the permissioned and permissionless markets and bridge that gap in a sense.

There are probably workarounds. This is also clearly the reason why TradeXYZ is a separate entity from Hyperliquid, right? The KYC permissioned markets kind of fall on the HIP-3 deployer versus the Hyperliquid core exchange and business.

I think that's clearly by design. You had those threads going out: What if the TradeXYZ team leaves Hyperliquid because they have all this increasing leverage due to the success of HIP-3 markets? But, again, I don't think it's a zero-probability event. I think it's a pretty low single-digit probability. It's a pretty symbiotic relationship between the 2 teams.

Is it easier to build an exchange like Hyperliquid and bootstrap the liquidity, or is it easier to bring a TradeXYZ-type business in-house at Hyperliquid? What's easier to do if push came to shove? It's much easier for Hyperliquid to deploy the markets than for TradeXYZ to leave and get the liquidity to come over, right? They might as well just deploy on other exchanges like Lighter and stuff, which I think they can do anyway.

Yeah, I don't know. I'm a buyer of HYPE. If it goes 10–15% lower, like I said, I'm going to start buying.

So you have cash?

Yeah, I've just sold some bad things that I bought that went down. I harvested some losses, and I'm sitting on some cash. I have Bitcoin, HYPE, and some ETFs.

Are you guys up this year, by the way? I'm curious. I was wondering this before.

Without HYPE, no. But HYPE's a huge part of my portfolio. With HYPE, yeah, but without HYPE, I would be down for sure.

Let's see. My crypto portfolio is down year to date.

Same.

My stock portfolio is definitely up year to date, though. I just own the indexes. That's how I like to live: if crypto is doing badly, my real money is doing well. But I'm still pretty optimistic for the back half of the year in crypto.

What are you optimistic about?

I wish you guys were going to be more bullish on crypto here. Bitcoin and stuff seems to be holding up pretty well considering the stock market.

No clue.

But yeah, I don't know. I think there's a lot of interesting stuff going on. Onchain is kind of back at different levels. Again, we talked about this before: you're seeing more and more things doing well.

There was a funny tweet the other day: in crypto, every 2–3 weeks you have a chance to make $100K, but every day you have a chance to lose $100K. That's kind of what the market's like right now, which is true. But those chances are coming back now, right? You're getting more of those experiments and speculative things coming back.

I don't know. I've been really bullish on MetaDAO, and they seem—

Have you participated in any of the markets yet?

In the ICOs?

No, the futarchy markets. Have you?

Oh, no. I don't trade the futarchy markets, but I've done some of the ICOs.

Which ones, if you don't mind?

Jurassic.

Throw up that tweet, Kevin. They just got a Skull coming onchain.

What did I do? I did Cars the Hot Wheels.

That one's kind of cool. Yeah, I kind of like that.

Then I've done some other ones. I forget. I haven't done a ton.

Have they all used the Colosseum stamp, or was it just the—

No, no, no. The only one I think that's done the stamp is Credible Finance.

Okay. I guess if they had raised prior, they would use that. So I don't know if the other teams had—

Yeah, a lot of them—for instance, Jurassic and the Cars the Hot Wheels thing—were both low raises. The thing that's really nice about MetaDAO is that founders are incentivized to raise at a lower valuation versus what we've seen so much in crypto, where you raise at a high valuation and then it just drips out, right? Because founders get paid, their tokens unlock based on multiples of the ICO performance.

So you want to balance it. You don't want to do too low of a valuation because then you're giving everything away, but you don't want to do too high, right? The Hot Wheels thing raised at a $625K valuation. Jurassic raised at maybe a $1 million valuation. Credible is more of a legitimate business; it was like a $5–10 million valuation or something. I don't know, but they're all good valuations, right?

Then the founders get paid if, basically, investors get paid, and so everyone is aligned. I really like that aspect of it. They're getting a good variety of stuff, man. Credible is a good payment company. They have some of the neobank stuff and privacy things. They're now doing this Jurassic thing. Hot Wheels, an order-book DEX, just raised there the other day, again at a $300K-something valuation.

It's nice to be able to buy stuff at cheap valuations, and you don't need it to be some crazy billion-dollar unicorn outcome—

Even if it goes to a $10–20 million market cap, you can get a really nice return on it—

You know, especially if you're someone with less capital and you're trying to turn a few hundred bucks into something. I don't know.

They have some good momentum recently.

I mean, they got an Upbit listing today, which is why it pumped. At first, I was really excited when I saw it pumped a lot, and then I saw it was because of Upbit, and it made me less excited. If you look at all the Upbit listings recently, they all seem to have retraced after the listing.

The price action before it was kind of grindy higher, which generally means you probably had people positioning and taking profits, too.

Yeah, if you look at the GEODNET one, that one definitely grinded up a lot. So I don't know.

Going to Jose's point about not being too bullish on crypto right now, Kevin, if you have your TradingView chart, one thing that I always do for a gut check, just to make sure I'm not on the wrong side of a trade, for example, is pull up Guppy. Do you know what Guppy is, Kevin? I've shown this to you.

I think I got it saved.

Yeah, you could use—

Is it this one?

Yeah, this one.

Fine. Yeah, they're all basically the same [expletive].

Pretty much, it just gives you a bunch of moving averages: a bunch of short-term moving averages and a bunch of long-term moving averages. Now just zoom out so you can see the whole chart.

When the short-term moving averages are below the long-term moving averages, you don't really want to be long. There are a bunch, right? You can separate the light blues—the blue ones—from the red ones.

The blue ones versus the red ones.

Yeah. The red ones are long-term moving averages, representing the long-term trend, and the short ones are the short-term moving averages, representing the short-term trend. Generally, when all of them cross through from the top or the bottom, it's a trend shift.

For all of the coins—and HYPE is not looking good—if you go look at equity markets, for example, you're starting to get that shift. When I see these big, long-term things that take weeks for momentum to play out and roll over and shift, I've generally found that not getting in the way of it is better. Just wait. I don't need to buy the bottom. It's not an ego test for me.

I'd much rather wait until the structure is either crossing back up or bottoming and not pulling away, so that the acceleration of the sell-off gets less and less.

This in here?

Yeah, you buy the [expletive] out of that.

Right.

You buy the [expletive] out of that. You don't buy the [expletive] out of that, at least I know.

This is BTC on the daily, right? I look at this as one of two things. I don't need to buy this here. I think buying this above the reds—above the recent highs—I would much rather let this show me that it has good momentum than trust that it's going to develop decent momentum.

This is just a standard thing I use all the time: don't get caught with your pants down on the wrong side of a higher-level trend shift. A lot of the things I look at are showing that, and I'm just like, you know what? I don't really need to be long here.

That's what I was looking at over the last couple of weeks, with equity indices kind of rolling over, especially the Qs. The Qs look pretty shitty. Like, really shitty. ETH has actually been looking the best out of a lot of coins right now.

I'm still not going to buy it because I just can't. It's too much. But I'm saying it doesn't—

We get smacked usually.

Yeah. But if you go with this kind of filter on a chart, there aren't many things that are like, “Oh, yeah, this looks [expletive] good.” This is a nice retest of a strong trend. Most of them are like, “Oh, we've retested the trend several times over the last month, and now we're finally not really bouncing from it. We're actually kind of rolling over.”

Just scrolling through the list, trying to find things that—

I mean, DXY looks great.

Yeah, the dollar looks good.

That's why I'm not really bullish on crypto, because almost all of the names in crypto that I want to buy look like they're getting smoked and probably aren't going to do too great in the short term. I'll just look to buy when there's some kind of structural change. The fundamentals look good for HYPE, so that's why I'm not worried.

The Multicoin tweet was so funny to me.

“We’re not selling,” but—

It was just a way to shill Zcash. Honestly, it was chef’s kiss. It was such great bag work. I actually think they might have sold some. That’s what I saw people saying the next day, but I didn’t verify that. I don’t know.

But it’s true, though. People should be able to sell without—man, this is why crypto is kind of lame. Everybody pocket-watches each other.

You can have better wallet hygiene, though.

Yeah, you can. It’s not that hard to make sure not every wallet is tagged, especially on something like HYPE, where you’re mostly buying it on the market.

It’s kind of—

I saw Alex Svanevik from Nansen get into this big argument about whether transparency is a good or bad thing. I was like, “This is the dumbest argument,” because this guy runs an on-chain analytics company and this guy’s an investor. Why is this argument happening on my timeline? Obviously, they’re never going to agree on anything.

4. Zcash & Ironwood

It was funny, though, because he shilled Zcash and Zama after that, right? This is why we need privacy.

Speaking of Zcash, do you want to pull up the Orchard stats?

Yeah, the Ironwood stats, I mean.

Zcash launched Ironwood as a reminder. The Orchard pool was the one that had this bug found back at the beginning of June—I think it was June 3 or 4. That’s when Zcash dumped 70% over the next day.

They made this Ironwood pool, and Ironwood is the exact same technology as the Orchard pool—the Orchard pool that was patched—but it’s also formally verified. Formally verified basically means that the math checks out. It doesn’t mean that there can’t be a bug, but the bug would have to be outside the circuit.

I’m not sure I understand exactly how to explain it all, but basically, the circuit and all the math check out. I don’t want to make it seem like formal verification means exploits are impossible, though. It means the odds are very, very low. You can look more into what “formally verified” means later.

They launched this new pool, and it went live yesterday. Orchard is now withdraw-only, so you can’t put money into Orchard; you can only take it out. You can either put it into a transparent address, which you could always do, or you can move it into Ironwood.

The more Zcash that moves from Orchard to Ironwood, the more that supply-verified pie chart increases. Basically, the way Zcash works is that if there were an infinite-mint exploit in Orchard, you could only take out the amount that’s in Orchard. You can’t run Orchard to a negative balance. So the more Zcash that moves from Orchard to Ironwood, the less a potential exploiter would be able to take out of Orchard.

It went live yesterday, and it’s been about 24 hours now. About 4.4% of the supply has moved. It’s pretty slow, right? I take that as a good sign. If you wanted to get your last bit out before the drain, you’d probably be trying to move a bit more, right?

I think a lot of people are just comfortable that there wasn’t an exploit. The Orchard pool will probably still have at least 500,000 Zcash in it by the end of the year. You can see Sapling, which is an older pool, has 581,000 Zcash.

Over time, people will become more comfortable, so this is probably just going to be a slow drip. You’ll probably see a few whales make some big movements at some point. I’m sure the team, early Zcash whales, or whoever has a lot in the shielded pool, right? Maybe they’re just sitting in there for now, or they’re going to do it over a really long time to make it more privacy-preserving.

It’s a cool tracker, and it’s good to see that go live. They also have their new pool, Tachyon, which is the most advanced pool they have. That’s coming out this year, and it’s also formally verified. Tachyon is more scalable and has far fewer lines of code.

The fewer lines of code you have in the circuit and the pool, the better. You also have all the best AI models auditing this, along with formal verification of everything. Again, that doesn’t mean it’s impossible for there to be a bug, but you’re making the odds lower and lower.

Is there a reason you would use Ironwood over the one that comes out later, or no?

Well, Orchard was around for a while, and so if you’re just—

If you’re in Orchard, would you just wait for the one that comes out later if it’s probably more secure and has fewer lines of code?

No. If you’re in Orchard, Ironwood is the same technology, but it’s also formally verified.

And if there were Zcash fraudulently minted in Orchard, you would want to get out before—

ASAP. Yeah.

Because then you’re left holding the bag. Realistically, all of Zcash—even if you’re transparent—if that happened, you’d lose it anyway, right?

So, yeah. The market clearly doesn’t think there was an exploit, based on how everything is moving and trading.

5. The Fed & AI Debt

But it’s interesting. Zcash has actually sold off a bit after this. I think people were positioning for it to really pump after Ironwood, and it hasn’t. It’s been a bit weak since.

Kev, slight change of topic before we leave. How do you feel about this Powell presser at the FOMC? It seems like it’s the first one in a while where there isn’t really a market consensus, based on Nick Timiraos’s posts—I think; I don’t know how you pronounce his last name. It’s the first meeting in a while where there isn’t really a clear expectation. There will be a surprise, so to speak, depending on what he does.

How do you feel about that? Do you think he’s going to raise rates or not? Do you think it’s a good thing that he’s not doing forward guidance going forward?

I’d obviously be surprised if he did a rate hike today. You’ve had a couple of people, like Citadel, coming out and saying they expect potentially a rate hike—a surprise to the upside.

I think Powell is caught between a rock and a hard place. You don’t want to look like a Trump puppet, but at the same time, in my objective opinion, I don’t think a rate hike makes a lot of sense at this point, even where we are in the cycle.

When you look at it, the economy is still running pretty strong. A lot of that, again, is being driven by AI, capex, and infrastructure buildouts. But underneath the surface, there are definitely cracks.

Even though equity markets are where they are, what happens with a rate hike? You’re potentially going to get some further tightening at the short end and the long end of the curve. Rates and yields are already kind of exploding to the upside, both in the U.S. and abroad.

Interest payments on our fiscal deficit are going to top $1 trillion. Higher rates make pushing out debt monetization more expensive, right? So rate hikes—especially when you’re doing it in the face of what I think will eventually be more deflationary pressures, during this inflationary period, with oil being as volatile as it is and energy prices going up—just don’t make sense.

The world was flush with oil before this. I think eventually we’ll get back to more modest oil prices. I don’t think that’s going to translate to long-term upside inflation expectations. So, to me, if you weigh the cost-benefit of raising rates, I just don’t see it.

That’s what I would do if I were in his seat, but I’m not in his seat. I can see why this is going to be contentious and why you’re going to see volatility around it later today. We’re recording this, and it’ll come out tomorrow, but we haven’t actually had the Fed come out yet. So we’ll see how well this ages.

I think what it means is that there’s going to be a lot more volatility around these things. These are actually going to become trading events again, especially because, to your point, there isn’t going to be forward guidance going forward.

I can see why that can be the right move, because the Fed—and Powell in particular—almost became a bit of a puppet himself, having to project too much to the market and answer too many questions from, I would say, uninformed people who were asking those questions.

I don’t mind not having forward guidance, but I think you have to take the lead from what the market is telling you. Right now, going into this rate hike, the market is split.

So, if you’re the Fed and you’re looking at this, I think it’s a hold situation. Then you hold until probably September, and September will probably be the one with more fireworks, because that’s the last real opportunity to do something before the midterms, right? If you stay on hold from now until September, the odds that you’re going to raise in October are pretty slim, in my opinion.

It’s a catch-22 of trying not to be a puppet for Trump and having your own independence, but at the same time, I just don’t think, objectively, it makes sense to raise here.

I’m in a similar camp. Let’s be real: Fed independence—yeah, they’ve been pseudo-independent for a long time. Going back to Yellen and Powell, everybody has to work with what they’re given and the administration they’re currently serving under, right?

On the forward-guidance thing, I think it’s probably a good thing. I liked his initial interview. I think it makes perfect sense. At a point, the previous Fed was becoming an input into its own decision-making process because of how consistently it was talking to markets and talking about specific inputs in markets. That essentially dilutes the signal, to a strong degree, of the inputs they use to make their decisions, right?

I think this goes back to something more resembling what it should be. Again, I don’t know if the market’s going to like that or if it’s going to be more volatile because of how it’s been conditioned over the last couple of Fed chairs consistently talking to markets, talking about specific inputs, and things like that.

I’ve always been of the view—even with equity reporting, quarterly earnings reports, and things like that—that this kind of forward guidance, talking to markets, occasionally really shifts the focus for investors toward short-term stuff rather than what’s actually good for a business. What’s good for a business over a quarter or two might not be the right decision over a year or two.

When you have a lot of these things, both at the high level with the Fed and at the low level with companies, it distorts a lot of market signals that otherwise have historically been really potent. So, I think this is the right move. I just don’t know how it’s going to play out.

And again, I’d round it out with this: you’re literally seeing global government bond yields. This isn’t just the U.S.; it’s an aggregate that’s U.S.-dominant, but you’ve got a bunch of other developed countries—the U.K., Europe, and so on. Those are now starting to all break out. You saw a little bit of a pullback this week, but directionally, the trend has been up.

It ties into a bunch of stuff we’ve been saying for years, which is that all these countries are running pretty sizable fiscal deficits. The U.S. isn’t the only one in this camp. Everything going on with defense spending and war, and even just trying to reshore manufacturing, is going to require a lot more capital that some of these countries simply don’t have or that their growth rates don’t support.

So, debt becomes the answer, and that’s starting to put pressure on the long end of the curve. One thing I’m glad I’m not sitting in is long-term Treasuries, if I’m being quite honest.

Wait, you don’t want to fund all of this? This is the chart I want to see Bitcoin have soon. I want breakouts this way. I don’t want this to be long-dated government bond yields or the dollar index. I want this in the bags.

What does this mean for what comes next? It seems like this should start to show up in inflation at some point as they try to make rates go down.

The counter is that if you cut rates or keep rates where they are, you’re allowing inflation to run more rampant if you believe inflation is going to be a big issue. I also don’t think this is a hammer-in-search-of-a-nail problem. I don’t think raising rates 25 basis points today does much around the energy-oil dynamic or inflation, or how that translates or reads through to inflation. I don’t think it really matters, to be honest.

Where this could wind up kicking us, to tie it full circle on our conversation, is the bear case longer term. This might be a 2027 or even 2028 issue, but we talked a couple of months ago about how a lot of the AI buildout is now getting shifted from equity financing and balance-sheet financing, with a lot of these hyperscalers having massive free cash flow, to being debt-financed.

A lot of this is off-balance-sheet, project-based financing for tens of billions of dollars of data-center buildout. What this does is, if yields continue to push higher and financial conditions tighten, it has a ripple effect into broader credit markets. You’re starting to see this over the last couple of weeks with some of the big-tech and hyperscaler bond prices selling off, and with CDS buying.

That isn’t necessarily a big red flag in itself, because there are multiple potential reasons why CDS could be rising. But the general point is that if financial conditions get tighter, if credit becomes more expensive, and if debt financing gets more expensive, all of a sudden some of the math on these projects may not make the most sense. You start to eat into the potential margin because of interest costs and the cost of getting that capital.

That’s where I think it’s not a today thing, but it’s something to keep on the radar. If you’re looking for bubbles, the AI debt bubble is one that can become very real, and this only works to hinder that and potentially pop it.

Yeah, I kind of look at this in 1 way. Interest rates are high and creeping higher, and you have this inflation scare with the energy shock and everything going on. But to Kevin’s point, increasing rates by 0.25% probably doesn’t do much on that side of things.

What it does is make borrowing more expensive and increase the debt service on the existing $40 trillion of debt. The benefits you get from curbing—or maybe not even curbing—inflation pressures get outweighed and completely overshadowed by all the negatives on the flip side: debt-service costs going up and all these other things, which then continue to put pressure.

You’re just not in a great spot. At the end of the day, I think the Fed has to—when push comes to shove, inflation is what they’re going to let in. That’s the release valve for the debt. It’s not like they can raise rates very high. The math just doesn’t work at a certain level, which isn’t that far away.

That’s ultimately how it has to play out. It can’t play out any other way. The marginal benefit you get from a hike is completely washed out by everything else on the other side of the equation, which just brings the decision-making timeline even closer.

I don’t see how they hike. On the credibility thing, they’ll probably be like, “Oh, he’s a puppet,” but I just don’t see it right now. I don’t think it makes any sense.

I think the things to watch are less the Fed and more what happens in the oil-market complex or energy markets, and this whole thing we’ve been talking about for the majority of this conversation around the AI buildout. You have to track whether the economics there start to break down. I don’t think they have yet, but this is part of the bear case that people keep talking about: the debt-financing side, where rates obviously do eventually play a bigger role.

What is the CDS on Oracle? If you were to convert that into an implication, what odds is the market placing on their debt default?

That’s a good question. I’m not entirely sure, to be honest with you.

Yeah, that’d be cool to know. It’s obviously high.

It’s a lot higher than it was a couple of years ago.

Obviously a lot higher, right? But again, they’re just benchmarks. A lot of times, you’re looking at just what the change has been. It’s not to say Oracle’s necessarily going to default, but it’s just a risk gauge, I guess you could say.

6. FWA & Crypto Gambling

There’s a lot of hedging that goes on with this stuff, too. I think some of the CDS concerns, especially for the hyperscalers, are overblown. I don’t think they’re going to default on any of that type of debt, but debt’s becoming a bigger part of the story, and it’s something you definitely have to keep track of.

Anything else? Do we want to talk about FWA quickly?

Yeah, that’s a good place to end it. What do you make of going to FWA and just how the mechanics work, quickly?

I’m curious to get your guys’ take on the whole gacha market right now. To me, it seems—obviously, there are certain markets where I think it’s a good business model for those who are actually offering this.

I think there are certain things, like Pokémon cards and maybe some of the Hot Wheels stuff, that are tangible, at least, right? With FWA.

I don’t know. It just seems pretty gimmicky, but—

So—

Maybe I’m missing something.

So I’ll tell you my take. First of all, I own this.

I’ve been playing it kind of aggressively. I’m also not—

Interested in the game. Okay. I’m not interested in what—

What is the game?

Yeah. So I guess gacha and all this stuff in general is something I’m personally not interested in, but it seems to be picking up a lot of steam. A lot of people are interested in it, a lot of people are playing, and they like to do this gambling thing.

FWA is essentially this, but it allows individuals to be the house and be the liquidity providers, right? They’re pairing NFTs from different collections. You might as well open up the site. People can come in and pair one of their NFTs with ETH, and basically, the ETH that you pair it with is what you think is the fair value price of it.

Then people buy these tickets, like they do in a normal gacha game. Your EV is negative, so you’re normally going to draw shitty NFTs all the time, but there’s a Punk in there and a legendary Punk in there. You could win something worth 350 ETH, right? It’s very similar to the Pokémon card pack things and all of that.

I think what’s interesting here is that the token incentives are super heavy right now. When you draw one of these cards, you can either choose to keep the NFT, take 85% of the value in ETH, or take 85% of the value in their token, FWA. Every day, 2% of the supply goes to purchasers who played the game, and those run out around August 2nd, 3rd, or 4th or something.

The only way to buy the FWA token right now is to play the game. You can’t buy FWA on the open market, so all external buys are turned off right now. If you click FWA at the top, it’ll bring you into the buy-and-sell mechanism. You can see that on the buy side, you can’t do it; you can only sell.

The chart looks like a continuous honeypot and rug pull, right? It’s just DCA up, then somebody decides to sell a lot and it dumps. Then it’s DCA up again, because you can only buy by playing the game. It’s a very slow process. You keep drawing these things, and that’s how the distribution has been ongoing.

I listened to some stuff with the founder. He spoke recently and said that he’s probably going to turn external buys on when the emissions are done, so that’ll be around August 3rd or 4th or whatever it is. I don’t know if that’ll be bullish or bearish. It’s possible this thing all blows up then and people are like, “I’m going to hold for the external buys,” but then nobody wants to buy externally.

I think what they’re doing here, though, is that they started with NFTs and then started adding ERC-20 tokens into this. Now you can pair with meme coins and stuff. They’re going to start adding support for other kinds of collectibles, too, right? You can see this becoming a platform that has all of the different things—not just NFTs, but tokens as well.

Hopefully, they start getting Pokémon cards and stuff like that on-chain. My thing was that I hate using Ethereum, and this got me to use Ethereum, so that’s worth something to me. I’m straight-up gambling with this right now, and it’s possible that the FWA bag goes to zero.

I don’t want people to think that I’m super convicted in this, but I think it’s one of these things that comes around and is kind of cool, right? We’ve seen a lot of things come around that get people’s attention. We saw it with Friend.tech, and there are a lot of examples of things that get the limelight for a few weeks and then die.

It’s entirely possible that happens here. Neil sent this to me on the weekend, and I was like, “This thing would have cooked in 2021.” When NFTs were the hot thing, this thing would have gone insane. But I think there is a real gacha meta now. Most of it is on Solana, but there is a real gacha meta, and people are genuinely interested in this.

A lot of people are bearish on this because it’s NFTs and everything, and people don’t care about that anymore. But playing in its favor is the fact that this gacha thing is doing well. This builder is pretty respected within Ethereum. I had one tweet about FWA, and tons of people I respect liked it, which is a good signal to me that a lot of smart people are interested in this.

So, we’ll see how it goes. I don’t know if the possibility is too high, but it’s there. The coolest thing about it for me was the distribution of the token, and that’s why I got interested in it. All right, Jose, take it easy. We’re going to end soon anyway.

I mean, yeah, usually when it’s hard to buy a token, it pumps a lot, right? When the friction is there, that’s what it was like in the early days with bridging to all those other chains, like Harmony and everything. If you went through the effort to get over there and get the token on the one marketplace DEX that existed, you were early enough.

Yeah. I played a few rounds and kept losing. I’m like, “Okay, fuck this. I’m just going to buy the token.” That’s when I realized you couldn’t. I was like, “Oh, you can’t even buy the token.”

So when the emissions run out, that’s when you can only claim the 85% in ETH going forward? Is that—

No, you can still claim it. You claim it in the token because it buys it in the pool. That’s why—

15 mil worth, you know. And it is.

Yeah. It’s not bad.

Yeah, and it’s fun. Maybe we could end it with this: I did a roll while we were talking.

Okay.

And it wasn’t great.

I hate the friendship bracelets. I don’t want to see another friendship bracelet in my life, man.

So this is the dynamic you’re talking about. It was 0.107 or something, right? I’m already in the hole on this.

Yeah.

If you go down, this is the dynamic you’re talking about. I can take 85% of this number right here back in ETH.

Why would I not just take FWA?

Yeah, I mean, if the game is more mature, I could see people taking ETH back, right? I’ll do this over here, and let’s—

The other way to play this, I guess, is that some people are taking everything back in ETH. Then, when the daily incentives come out for that 2% of the supply, they’re just selling the FWA token and hoping they make out ahead by doing that.

For the more adventurous folks, they’re just buying FWA and hoping that this thing ends up cooking once it becomes more mature. But we’ll see. You’ll basically know within a couple of weeks if this thing is going to last or not, because both the incentives and the external buys—the incentives will be done in about a week, and the external buys will probably be turned on in about a week. They’ll probably put some buybacks on in a week or so, too, right?

What they’ve got to do is get more markets. They’ve got to get more things that aren’t just NFTs. You need CS:GO skins, Pokémon cards, Magic: The Gathering cards. You need shit that people who aren’t on Crypto Twitter want.

For sure. One of the reasons I think it’s starting to take off is that, if you’re the issuer of these, you look at Courtyard or Collector Crypt. Those types of models are really good, lucrative business models if you’re the one who owns the issuance or owns the game itself. Being an LP seems like a pretty good go-to if you can actually see how much—I want to see how much that Punk has made on the main page.

So, in 6 days.

In 6 days, the Punk has made 90 ETH. It basically needs another month to break even, I guess. But the odds of pulling that Punk are insanely low. That Punk is probably going to make money; it would be really bad luck if they didn’t. But I don’t know at all about the LPs of a rare NFT.

Are the rolls provably fair? Is that audited?

Yeah. Apparently, there was an exploiter. They used Chainlink, I think. In the very early iteration, there was some sort of exploit, but they fixed it.

Yeah, because I’d be so pissed if I put that 350 ETH—

—up and then it just got yanked.

Pretty good, though, man. If you have 20 ETH back, I’d throw myself out my window.

Yeah, it would suck to list a rare NFT and then get it pulled on the first day.

Dude, but I mean, you can make good P&L because, with the 85%, you keep that 15%, right?

Yeah.

You’re essentially supposed to make a 15% yield because nobody keeps the NFT.

Yeah, unless it’s a good one.

The only way I would keep the NFT is if I pulled a Punk. If I pulled a floor Punk, I would probably keep it because then it’s easy to sell for the ETH. If I pulled a floor Ape, I don’t think I’m going to ape it all into the token. Why would I take an 85% haircut if it’s pretty liquid anyway and I can just go sell it on some NFT market?

Why, for 2 days, is that one only 0.3 ETH?

Yeah, I’m not sure. I’d imagine it’s prorated based on what percentage of the pool value the NFT represents.

Maybe it got pulled at one point. No.

It could have been pulled, though.

No, but if it got pulled, your P&L would be crazy negative.

It’s interesting, though.

It’s a cool idea.

How much do you roll with—as much money as you want?

Yeah.

You can buy 5 tickets at a time. They’re pretty much always 0.1 ETH.

Okay.

You’re literally just playing a fucking raffle.

You’re playing a very negative-EV raffle. The odds are very bad—

The odds are already bad, and then you’re buying the token at 85%. You’re basically expected to lose 21% of what you’re playing with. Maybe it’ll change over time and stuff. We’ll see.

Yeah. Unless FWA pumps, I’m down 50% off the bat on that one roll.

Yeah.

Straight up.

Anyway, I think it’s one of those things where a lot of people are talking about it. It’s pretty interesting. It’s kind of a true crypto-native project, right? It has that feel to it, and those are always the things that I find interesting. I think people like to shit on it. I think Path had a good tweet about that: You just can’t launch a new idea in 2026 without people tearing you apart, right? I think we should encourage people. Yeah, you can—

You can say that this thing is going to fail. That’s fine. But it’s cool. It’s good to get people building on-chain. Man, who uses EON Chain again? I don’t have money on ETH, and I started to play this, and now I do, right?

I’ve been buying—I’ve been pressing the buy button on ETH.

I’m like, I can’t believe somebody got me to press the buy button on ETH again.