Unchained - Crypto's Black Friday Was Its Largest Liquidation Ever. What the Hell Happened? - Ep. 922

Episode Date: October 12, 2025

When Trump tweeted about 100% tariffs on China, crypto collapsed — $19 billion in liquidations in just hours.  Altcoins plunged 95%, exchanges froze, and stablecoins depegged. In this special episode, Diogenes Casares, founder of Klyra Protocol, joins Laura to break down the chain reaction: what really caused the crash, whether insiders knew it was coming, and how infrastructure failures and extreme leverage turned a policy tweet into crypto’s Black Friday. Thank you to our sponsor, ⁠Aptos⁠! Guest: Diogenes Casares, founder of Klyra Protocol and advisor at Patagon Management Links: Diogenes’s article on X: "Black Friday: What Happened?" Jordi Alexander on "What happened? Stani Kulechov on Aave’s performance Binance co-founder’s statement  Timestamps: 💥 0:00 Introduction ⏱️ 1:00 What markets looked like in the hours before the crash 🕵️‍♂️ 3:25 Whether traders on Hyperliquid knew the tariff tweet was coming 📉 5:44 Why altcoins plunged up to 95% and how market makers amplified the move ⚙️ 7:56 How auto-deleveraging kicked in—and why it mattered 💣 13:07 How DATs created hidden leverage that made the system fragile 🏦 14:54 How perps DEXes and CEXes responded differently to the meltdown 🧩 18:09 Was it a coordinated attack—or just market mania? 🤯 25:56 What happened to smaller market makers when liquidity vanished 💥 29:44 How the USDe depeg on Binance triggered cascading liquidations 📊 32:16 Why Ethena “managed it well” and why exchanges don’t ADL their positions ⚠️ 34:04 What caused the USDe “depeg” 🔧 35:22 How infrastructure failures made price feeds unreliable 🚨 37:28 What perps exchanges need to change going forward ⏳ 43:56 Why Diogenes thinks this kind of crash will happen again—and worse 📜 50:20 The “extraordinary rights” LPs hold on exchanges 💱 49:02 How traders should decide where to trade after this 👀 52:19 A rumor about how much Jump lost Learn more about your ad choices. Visit megaphone.fm/adchoices

I think it's something like 10 times worse than the FTX collapse, which is pretty astounding if you think about it. I haven't seen anything this aggressive since then. The amount of money that all of these traders in theory were kind of not cheated out of, but were unable to realize this kind of mind boggling. The infrastructure in crypto just failed. Hi, everyone. Welcome to Unchained. You're a no high resource for all things crypto. I'm your host, Laura Shin. Aptos is the no-compromise infrastructure for global financial markets, fast, reliable, and 100 times more cost-efficient than other blockchains. See for yourself why Aptos is the chain of choice for institutions, users, and developers alike, at the Aptos Experience, October 15th and 16th in Brooklyn.

I'm here with Dio Casares, founder of Clear Protocol and advisor at Patagon Management. Welcome, Dio. Thanks, Laura. Good to be on. So obviously we're about 24 hours after a crazy set of events in the crypto market that all kind of happened in the span of, I don't know, like 20 minutes. You've read a great timeline of the events before and after this Black Friday crash in which we saw $19 billion in liquidation occur. So tell us the story of what led up to that moment and then what happened during the crash. Yeah. I mean, at the start, there was already kind of rumblings that something could happen.

On Thursday, China had announced any products being exported from China with a certain amount. It was literally like 0.1% containing like Chinese rare earth materials would need government, like direct government approval to be exported, which basically meant they were banning all exports of rare earths outside of China. That kind of was under the radar for a little bit, but it's largely seen as a response to the U.S. banning China from buying a lot of chips. And so this is kind of a very three-year-late tit-for-tat. And Trump has been very vocal about how he really thinks it's important to secure rare earths for the U.S. on that it's like a defense interest.

So basically at the end of, basically at market close in the U.S. US, Trump came out and said that as a result of this, the US would impose, I think it's 100% tariffs on China. That led, there was some interesting shenanigans and that 30 minutes before that there were these large hyperliquid whales that went short hundreds of millions of dollars worth of Bitcoin. And then there seemed to be similar activity on other exchanges, but you can't really see it. It's just hyperliquid where you can actually see it. And then the price obviously went from like 122,000 for BTC to as low as 104,000 within like a one hour, two hour timeframe. And I mean, yeah, it's unique in that it's a relative, it's something we've seen before,

tariff announcements, but the extent of the carnage is very unique. I think it's something like 10 times worse than the FTX collapse, which is pretty astounding if you think about it. Yeah, I saw a little kind of like chart on Twitter that showed that. But let's talk about that in a moment. I do want to hone in on the accounts on Hyperliquid that opened those shorts right before, those huge shorts right before the announcement. So first of all, those were new accounts. And when you say that you felt like similar activity might have been seen on

centralized exchanges? Like what, you know, clues are you? I mean, the price in that 30, in that 30 minute time frame, it might have just been the fact that, you know, these very large players were kind of telegraphing the fact that they wanted to go short. So then people were also going short on other exchanges. But the price dropped about 3% during the course of them going short. And to move the market 3% while at the time, I believe, the U.S. market was still open is not, the size that was exclusively on hyper liquid wasn't enough to move the market that much, if that makes sense. Oh, okay.

Okay. So it would have had to indicate that there were other people that were making similar moves. Okay, got it. What do you make of the fact that we saw at least that on hyperliquid and we're, you know, it seems reasonable to suppose that there's similar activity on other exchanges. Like, what do you make of that? It's very likely a case someone knew something. Knowing hyperliquid in the history of whales putting large positions on crypto,

it could be an incredible story of some crazy whale deciding to go short. Because the Rare Earth Metals play was kind of telegraphed before, like it has been something that if you track Trump, he does care about for a while. And it would make sense for him to drop something like this right before market closed because he doesn't want it to nuke the market. He wants people to think about it over the weekend. But in general, the way it looks is that someone had was kind of inside the information loop there,

might have told someone else, and then they placed a trade to kind of bet on how the announcement would hurt the market or affect the market. Yeah. Okay. So now let's talk about the fact that so many alt coins, like seemingly nearly died, you know, you cited some of these in your little essay where, you know, Adam went to nearly zero. Athena dropped a lot and, you know, that's considered definitely one of the stronger, newer players that we've seen, Saul, which is definitely considered, you know, one of the

most legitimate coins dropped by 40%. So, you know, explain like why you think that happened. Yeah. I think an interesting thing here to also mention is that this almost exclusively happened on the perp order books. It didn't happen on the spot order books. And there's a reason for that. First of all, a lot of these, you'll notice this didn't happen with like XRP or some of these older kind of alts

that are still popular. A big reason for that, in my opinion, is most of the alts that were affected had like two to three market makers that were designated market makers for that token. According to what I've heard from a bunch of people, it became just absurdly hard for market. market makers to actually place orders.

And because there's only really at scale, like three companies placing bid and ask orders on the book, all of a sudden none of them can place orders. So the price just becomes, you know, no one's setting price. The price in theory can go to it almost nothing. So I think that's what you probably saw with Adam. You did. It looks like a smaller market maker also blew up. So like if you look at Resolve and some of these, I'm forgetting the exact name, but it's something like sciant or something like that.

It's like, or like, not cipher, but it's similar. They basically, their entire hyperliquid account got evaporated. And all the coins that they were market making basically almost went to zero. And so I think that's also part of it. But most of it, I think, was market makers not just not being able to trade. So what does that indicate? Is it like a skill issue or like why? It's not a in the case of like you know market makers going down it is a skill issue like the market maker is no longer in business. They lost all their money. That is they did not do their job.

In the case of most of the market makers though, in theory when you have we had a massive auto de-leveraging event. So the exchanges decided that they couldn't internalize any of this risk. And they would just start closing out a lot of these positions over the course of like 20 minutes. When that tends to happen, a lot of the times exchanges won't let market makers market make in the same way. Functionally because almost what's happening is like a reshuffling. Like ADL you can kind of think of if you've ever been in Vegas at a blackjack table. And the guy next to you. which you're talking about auto deleveraging.

Yeah. If the guy next to you blows up and goes bankrupt, you, if you're next to him, have to pay some chips because he still technically owes the bank some money. This is like a very aggressive equivalent of that. And the way it works is, say,

user one has $100 is short on 3x leverage, say. And user two has $100 and is long on 3x leverage. So you only have $100. So if the price goes down to 40%, or down like 30%, roughly, that's when the liquidation has to happen. So obviously you have, in this case it would be reversed. You have one trader that's down a lot of money,

one trader that's up a lot of money. They're inverse. But if there's no liquidity on the book, that means that the trader that's up a lot of money, if they don't close out the trade with the other person and just settle, suddenly this person is going to owe this guy more money than they have. So what ADL kind of does is when they're like, oh my God, there's no liquidity on the book. The price is moving too fast.

They basically shut down the market. And you have to think about the each exchange, right, is trying to be as safe as possible. So there's actually an incentive if the market's moving very fast to do ADL first. Because if you don't do ADL first, everyone else's price is going to be worse. So that's probably what likely happened. Suddenly market makers can't actually, like, it's not like a skill issue of like, oh, they weren't able to. It's like they couldn't really place bids or asks during that time frame. And I've heard at least in like kind of my own experience, a couple people have told me that that's basically what happened.

Like a couple of these top of bark market makers just suddenly couldn't place bids or asks for extended period. And so for those market makers, it's not their skill issue. It's the exchange skill issue. Is that how to think about that? I mean, it's not it's, they couldn't have placed orders because ADL was happening, basically. Oh, okay. If the exchange hadn't done, started an auto delveraging, there could have been a more serious issue. It's like a separate combo if the exchange deciding to do ADL instead of kind of stepping in and being like, hey, we're going to cover X amount of. of losses was the right choice. Like, lighter made kind of the decision that the LLP would take a hit instead of users, although they still had some auto-de-leveraging. But in terms of, like, was it the market maker's fault?

Almost certainly it wasn't the market maker's fault. The other thing is that with the amount of liquidations happening, no one has, like, I think Wintermutes what everyone thinks about when it comes to, like, massive size and yet, today said, oh, yeah, we're used to doing, you know, 15 billion in transaction volume a day. And there was 20 billion in just liquidations, right, within two hours. So just people were also just overwhelmed. You just didn't have inventory. You didn't have, how do you price stuff when all of a sudden also, you know, major prices

across exchanges were different by like 10 to 20 percent at times. Like it was just chaotic. It's like kind of, I don't think any one market makes. or would have been okay in this situation. Like, this is what happens when this level of de-leveraging happens. And so, like, why is it that it became so overwhelmed? Like, I don't know if you saw Jordy Alexander's kind of longer tweet, but he basically said that this type of cascade he felt like was because everybody was, like, way too leveraged

and nobody was kind of, like, basically, yeah, people, he, like, the way that he, like, the way that he described it was just like people just were chasing money. And he had this funny example. He was like, in recent months, I've been hearing more and more retarded thesees for buying coins. Did you know that CZ's gardener's dog is called asterisk? Time to bet on that shit. It's BSC season. So like, do you feel like that's just what it was? Like everybody got over their ski or not literally everybody, but a lot of people got way over their skis or like, why do you think this happened? Yeah, I mean, I think we had some indicators. Altcoin open interest on perps was absurdly high. It was close to being similar in size to BTC open interest, which just doesn't happen.

So that's, I think, one of the main reasons why, and a lot of people were also, there were a lot of DATs coming out. And a lot of the DATs, people assumed we're going to bring a large amount of cash to just start buying these tokens. And so a lot of people were incredibly levered thinking, oh, there's no marginal seller right now. It's going to go up, you know, 10, 20%. I should be, you know, 10x, 15x long, or even on large size 5x long, Avax, you know, any of these kinds of tokens that there's dats or rumors of new debts. And so there was a bunch of people chasing a catalyst that didn't happen yet. And then on the other side, you had, you know, a recent new all-time high happened. And then people go and kind of look at previous all-time highs.

And they're like, well, you know, 80% of the times that there was a previous all-time high, the chart looks like this. So therefore, I should go along with my whole net worth. And obviously, that's not how statistics and finance works. So I think it was a combination of there just weren't a lot of people actually, buying BTC a little bit more, but especially altcoins. Most of the price was just kind of this incestuous, more and more derivative and levering up. And then as soon as there was a serious decline, people just weren't ready for it. Especially on a Friday, the liquidity is already not that great.

Yeah, that makes sense. I actually just wanted to hone in on a difference that you made between the perplexes and the centralized exchange. you know, it seemed like obviously like more of the liquidations are happening in Perps than on Spot, as you mentioned. But like did you feel like there was a difference between the decentralized Perp's dexes and centralized ones? I mean, very publicly, hyperliquid had no significant downtime, which was quite impressive. Lighter had downtime, but it was after most of the chaos happened. The main thing, most, almost every single exchange, I have yet to hear of any exchange, including lighter, that didn't use some form of auto D leveraging.

Almost every single exchange did it. There was just no mathematical way for them to avoid some form of bad debt scenario. I think the main thing that we saw that was different is like, you know, the ADL kind of setup was advantageous for HLP depositors, oftentimes at the direct detriment to traders that were short because obviously their PNL was just like capped. And basically there's an interesting dynamic of a bunch of the HLP liquidity providers are up a lot. I mean, I think the pool made 40 million. And then the LLP liquidity providers are down, but also in theory users that were short are in a better place, although then it shut down for a couple hours.

So it kind of balances out. I think that's the big difference in how exchanges managed it. Backpack got in a decent amount of trouble because some people were depositing to add leverage or to like kind of top up their account, but ADL was already happening. And people were already getting liquidated. So some people that were adding to their account,

that would just add to their account, that would just add to the account balance, and then they were getting liquidated. And so in their mind, they were like, wait, all this money is just leaving me right now. But apparently that was like, they were bad with the comms initially, but apparently that was an issue on the technical side,

and they're going to be able to fix that. And users are going to be fine. But, I mean, I think the main question was kind of, how do you manage ADL? I don't think people realize, This was the most extensive use of ADL I've ever seen. Like there were cases also of people going long at the bottom of the market. And then the market went straight back up again.

And then those people also got AD held because just all the market makers were like, what is happening? We have not, we have not seen something like this. So there was not enough liquidity. And so ADL would happen on the upside two. So some people were short and they were right. And then it went down. And then here on the short, they were closed out. And then it went all the way down here.

And then they were, went long. And then they were right. And then it was closed out here. And then the price went up here. So the amount of money that all of these traders in theory were kind of not cheated out of, but were unable to realize this kind of mind boggling. Yeah. Yeah.

I saw some other tweets about that. You know, Binance ended up like releasing a statement apologizing to people because of the disruptions on its platform and saying that they're going to compensate some people. So like what do you think happened there? Was it just the. ADL or like what were the disruptions? There's like different ways that you can take this.

You can, if you're a little bit more conspiracy theory inclined, you can say, because the timing of the Binance and the by bit shutdowns of service, some of the APIs was pretty, was almost one to one. Like it happened roughly at the same time. So you could argue that, you know, this was like a coordinated attack if you're more conspiracy minded. What it probably was, to be honest, and what a lot of people forget is these exchanges have very large internal market making desks. And the internal market making desks function very similar to like an HLP or an LLP in that they're often the market maker of last resort on these platforms.

And suddenly with the same volatility that was happening, these internal market making desks found themselves with absurd exposure and on the wrong side. And so probably that caused Binance and Bybitt to panic and be like, oh, my God, we're sitting on theoretically, you know, billions of dollars in exposure. And we have no idea where this is going, you know, across asset classes as well. We need to push ADL now. That's probably what caused them to react that way. The other thing, the last kind of, there's kind of three ways that this could actually have gone. The other thing is it just could have been the APIs got overwhelmed. And there's kind of one on there's different types of APIs.

So the internal market Macare APIs basically always work. And then external ones were just getting absolutely flooded. So no one could really place orders. So there's kind of three ways explanations. And it could also be some bizarre mix of all three of them. Yeah, I was just about to say it could be all three. Yeah. I mean, I think like what's so interesting to me is like crypto has seen so many.

crashes before. So like what and and you know, I saw a tweet. I can't remember who said this, but they were just like the fundamentals are nowhere near as bad as they were like at the time of Tara Luna or, you know, 3AC or like FTX even. And that to me also is what makes it feel like, whoa, like how did such a bad crash? Have it like, do you remember during the COVID one where it was like, like, shoot, what was it? It was something like the blockchain itself, like, was stumbling. I don't remember all the details. I remember I think Kyle Simani came on the show and explained it all.

And I just remember my mind was like blown. But the point is that this, you know, what happened here definitely felt more like something structural, like about the exchanges or maybe simply. Because the other thing that I was thinking about was just like, we're at this time, we're at this like moment in crypto history where like the Perps thing is really getting underway. You know, it's not even just, you know, some of the stuff that we talked about. Like obviously hyperliquic kind of like created this, you know, just like suddenly everybody wants to get in on perps, right? So it's like, you know, Astor, the hyperliquid Astro thing that's going on with Seiz and Jeff. But then like Lighters getting in the game, they just raise money, you know. And then we've got. like phantoms launching perps and there's just kind of this like perps mania and like just the fact that you know like finance has been the number one exchange now for eight years and they've been super dominant in that like nobody has ever even come close to challenging them and you know like maybe like hyperliquid is the first thing but but it's just a different beast so um that's to me like that's why this one feels so much more mysterious or

like incongruent. It feels incongruent with the rest of the market. I think that's right. I think the big thing, you have kind of two big problems. And that first of all, you're right. The market share of perps has become incredibly fragmented. There's also what I call kind of, there's a phantom OI, if that makes sense, in the sense that there's a lot of people that are farming, you know, exchange one and they have a long. and on exchange two, they have a short, and then they kind of repeat that process. And it's not actually meant to be real open interest. It's just meant to be a farm. And so that inflates a lot of numbers, but it still sits on the exchanges balance sheet.

The other thing that I think is important to think about is because of all this competition you're mentioning, you have a severe freeloader problem in the sense that now none of the exchanges really, you don't have like a heat map of where all the liquidations will happen on every exchange. Most of the sexes keep it entirely private. Lighter to my knowledge, you can't really find out. Hyperliquid, you can find out what people can obfuscate it a bit. So exchanges are dealing with imperfect information as they like finance before. They knew, you know, this guy is going to get liquidated when the price happens here. We have 80% of the market. We can kind of map this out. Everyone's dealing now with like very imperfect information and they don't. And it's scary.

You know, if, you know, it's like, imagine getting a lawsuit. you can only see 20% of the words that are said. You would probably assume the worst. So in this environment, it's like very hard. It's very mercenary because the exchanges immediately when something like this happens. As an exchange operator, you just want to get this all off your book. You just want to get it all over with. Because otherwise, also, if you think about it, whoever does ADL last,

assuming like relatively similar liquidity, executes at the worst price. for a lot of the traders that are getting liquidated. So I think that's mainly the thing to explain all of this. The other thing is, I mean, I guess another thing is, you know, rate cuts are expected. So a lot and a lot of people are like used to now being long like three to five X on a lot of these tokens into rate cuts. So when prices suddenly went down more than 50% on a lot of these alt coins within 20 minutes, you can't even add margin on some of these exchanges. No one was really prepared for how to manage that.

Yeah. Yeah. Clearly a lot of people were caught flat-footed, unfortunately. All right. So in a moment, we're going to talk a little bit more about different sectors of the market and how they performed. But first, we're going to take a quick word from the sponsors who make this show possible. Over 3.4 billion transactions processed without disruption. More than $1 billion in stable coins circulating.

Over $720 million in real-world assets tokenized on-chain. All delivered with sub-second finality, block times under 100 milliseconds, and fees less than a tenth of a cent, making Aptos more than 100 times cheaper than other leading blockchains. Built by the team behind the DM project at Meta, Aptos is what blockchain performance looks like when it's built for global financial markets. Discover why global institutions like BlackRock, Franklin Templeton, and NBC Universal are building on Aptos and see for yourself at the Aptos experience October 15th to 16th in Brooklyn. Back to my conversation with Dio. So I did want to, you know, zoom in on the market maker comments that we were talking about earlier. It sort of feels like, you know,

there was wreckage yesterday. And now a bunch of people are probably kind of sorting through where they are. And you kind of hinted like you felt like maybe some smaller market makers maybe didn't make it. So like what do you think will happen from here? Like how do you think we'll figure out kind of like who blew up if anybody and where everybody stands and how long or how, you know, all of it will work its way out until, you know, we kind of what where everything comes to light? Yeah. I mean, I think it'll probably take a while. I think there's a big factor to consider is the lenders to these companies. I think probably every lender to any market maker right now, especially the unsecured lenders, are probably calling them up right now and being like, hey, we want our capital back now. Maybe some of the larger market makers are fine and they have good enough relationships to where it doesn't really matter and their books large enough to where it doesn't really matter. I think for a lot of the smaller, especially prop market makers, it could be over for a lot of them, especially some of these smaller ones that actually do have designated market maker agreements.

I think those are the easiest ones to spot because basically you can look at the charts that they manage and you can see, oh, you know, this almost went to zero. That's probably a sign that the market maker just basically died. I do think that like you have to caveat that though with like the infrastructure in crypto just failed like if anyone that was trying to use rabbi or debank during the chaos knows that like basically it was unusable for like at one point a couple hours and anyone doing a lot of transactions on Solana unless you were buddies with some of the nodes a lot of people weren't able to get transactions to go through either. So I think it's going to be hard. it's also hard because I don't think the market is going to, it's not like a Terraluna thing, where to your point earlier, you know, fundamentals don't seem that bad unless we're all not seeing something. So the market's probably going to go back up. So I think that it's quite interesting

because a lot of people will probably be able to hide kind of some skeletons in their closet throughout this and we'll only find out maybe like next year, which is kind of scary. Yeah, I mean, honestly, it's basically like, you know, how we didn't figure out anything about FTCS until November. And then, you know, Tara Luna, I think was like early May. So what is that? That's literally six months. Exactly six months. Oh, actually, actually, it was six months and one day because Tara Luna was May 7th.

And FTCX was November 8th. So interesting. Okay, so that puts us at, what is that? That's April 10th. So we'll see what happens come April 10th. I did want to also bring up DFI because there were a bunch of tweets about how kind of DFI worked properly, especially from the likes of AVE where it actually did work properly. They talked about how they had processed $180 million in the like,

quidation so it's a new record. So, you know, there was that. There were, you know, some instances where things, I don't know if you would say they went awry or not, like what happened with Athena. So just maybe, yeah, talk about like how you see it broadly and then you can pick out whatever examples. I mean, an interesting thing here is probably, you know, there was a promotion going on that you could post USDE as collateral. And then you could, it was basically like, you could get 12% basically. And then you could use that to run like other trades and earn more. So there are a lot of people running this kind of looped with a decent amount of exposure.

But at the same time, the way per markets work is that they have, you can have a collateral asset, but it's different from it being the quote asset. So the quote asset for most derivatives markets are, is either USDC or USDT. So these markets, to my knowledge, were USDT. So that meant as USDA depanked, it kind of forced a cascade of liquidations. And because they had say like, because if you consider the fact that it was being used as margin for a lot of these traders, a trader's position gets liquidated. Not only does it get liquidated, then that USDA they had as margin gets sold to cover the USDT. So suddenly there's an absurd amount of buy pressure that the market just can absorb.

because also whenever these things happened, like to a point before, there was $20 billion in liquidations. Also, now there's probably hundreds of millions of dollars and sell pressure on all of these collateral assets. So there just wasn't liquidity. If the, so the price, I think, went down to a low of roughly 62 cents on the dollar, which is a 38% decline and like almost a 50% return if you had bought it. if oracles had pulled the price from centralized exchanges, which also, by the way, all the centralized exchanges had totally different drawdowns and minimum prices.

But if Defi had, that would have wiped out probably close to $4 or $5 billion in liquidations across different platforms, which is orders of magnitude higher than what actually happened. And in this case, it wasn't necessary at all. It would have been a foolish thing to do that. But I think the parading around of like, oh, this didn't happen because basically we hard-coded an Oracle is probably, like, sure, there's, like, truth to that. But you also have to admit that hard-coding an Oracle is probably not the most secure and safe thing to do long-term.

Yeah, for sure. Other way. That's a big thing also to take away from this is how well Athena managed it. Apparently, Shoku on Twitter said that basically Athena has a setup so that they don't get ADLed. And I talked to some Athena guys and they said, since they rebalance so much, it functionally provides a level of liquidity to the exchanges. So the exchanges don't ADL their positions. Oh, wait, wait, wait. Just explain that in a little bit more detail.

So ADL works by prioritizing people that are percentage up a lot. And U-P-NL, so unrealized P&L, is quite high. So if you continually settle your position, one, oftentimes if you're Athena, you're often settling on the side the exchange needs orders for, like they're closing shorts, right? And basically that's like a net long, or not a net long. They're closing shorts in the profit. The exchange wants to close these shorts that are in the profit. So it makes sense for the exchanges to create an intent.

incentive structure wherein if they constantly rebalance their positions, especially as UPNL gets higher and higher, they don't do ADL because in cases like this, Athena actually limits, in theory, the amount of ADL that will be needed. They're functionally providing a level of liquidity instead of having to close someone else out. Athena is willingly closing a position, right? So I think it makes a lot of sense that they were able to get that deal. But there's also probably a decent amount of, I know for a fact there's a couple prop firms that don't have a deal like that,

that definitely did feel the impact of ADL. And there's probably a couple of smaller protocols that try to kind of mimic a theme that that were in trouble. Although from what I've seen at least resolve and neutral seem to be okay. Okay. And wait, so I guess one thing in that I don't understand is so like why did it DPEG? Not that. USDA? Yeah. USDA, yeah, sure. So one, obviously, you have nothing's clear, people don't read. So like, it's not clear that USDA is like in theory

the biggest large stable coin that could collapse. So probably you have some people freaking out about that. But the big thing. That's why they call it the synthetic dollar. Yeah. But the big thing I think was you could post USDA as collateral. And so as soon as these liquidations started to happen. Oh, right. The USDA was forced to be sold. And most of the time, this is an on like, USDA lives on chain. I don't know anyone that uses USDA that's actually done a swap on a centralized exchange. You go on chain and you swap on Athena's site and you get the best price. So all of a sudden, all of these orders are going through centralized order books during a time of market crisis to kind of cover the perpetual positions. And that's what caused the DPEC. It wasn't like a run on the bank,

really. It was just a systemic thing when it came to that people had USDA collateral. They had to sell it to cover the USDT debt in these prepositions. That created too much USDA pressure and then that created its own cascade. Got it. Okay. That makes sense. And I just wanted to go back to something that you briefly mentioned. You talked about like some of these wallets did not have the correct on-chain prices or something like that. Like what happened there? Yeah. So basically because of the amount of activity, this tends to happen. It happened with exchanges as well, but like Rabby, a lot of service providers like RPCs, et cetera,

just went down. So I know a lot of people that are kind of in the sniper community in Solana, for example, some people were able to sell Solana 5 to 10% above what the actual market price was in all the centralized exchanges, because no one could get a transaction through and because of how fast it was.

So there was a lot of situations like that. There was a lot of liquidators. There's a lot of aggressive liquidations because liquidations function like an auction. And then you like get handed over whatever's left over. A lot of people were liquidated and functionally given nothing back. Because since there wasn't a lot of competition, a lot of people couldn't even place bids. And also the market was crazy volatile. There just wasn't an incentive for people to give any money back or like bid to bid higher on that liquidator auction.

So it was a lot of people, like if you use debank, which a lot of people do for pulling data if you don't have your own system, you had no idea what the hell prices were for a long time. And for a lot of like other services and a lot of other RPCs, you were getting, you weren't getting any results if you were trying to pull on chain prices. If you didn't literally have like your own node or access to a node. Okay, that is insane. And when you say that, how long did that last? It's probably service dependent, but I know for a fact that at least D-Bank was down for at least like an hour or two. And D-Bank and Rabi are kind of the same, like they're the same kind of company, I guess. And so if D-Banks down, basically Robbie was down at the same time. Okay. So when you look at kind of everything that happened, you know, I know you have like extensive.

experience as a trader. I know you think a lot about like exchange infrastructure or things like that. Like what do you feel like the industry needs to improve on? Like what are all the different areas? And I mean, it could be technological. It could be economic. Like there's just so many different, you know, ways that things could be improved. But yeah, like just in the first 24 hours, what lessons can you think of that the industry should take forward to improve itself? I mean, I think you should have right now people have become very tribal about this whole thing you know the the hyperliquid people are like oh look at how much money the hlp made isn't that fantastic and you know we had no downtime whatsoever but you know they had extensive ADL whereas on the lighter side it's like okay we had 5%

drawdown but the lLP is still up you know 237% on the year uh and we didn't really have ADL on the downside. We had a little bit of ADL on the upside. And I think that like an interesting question to be posed to like all the exchanges is they all make, uh, billions of dollars a year, uh, on perps platforms, um, on like, or derivatives products in general, I guess. And given that there's more competition now, people are going to remember which platforms, excuse my French, but I'm struggling to find a better term, screwed them over, versus, versus the platforms that they felt gave them better execution. And given that it's becoming so competitive,

I think that a very healthy thing for this space is like, you know, Bibet, I think, has made enough money to pay back the $1.5 billion that they lost already, right? Obviously, these are like crazy numbers in theory, but if you actually add them all together and you assume the worst case, in theory for a lot of these positions,

the net exposure for a lot of these exchanges is not, that that. So I think there should be an interesting conversation about like at what point should ADL be used versus just using it immediately. And what role do internal market makers play and how harmful can they be when you basically prioritize your own internal market maker over these other market makers that might have, you know, deals for specific tokens, et cetera. are suddenly locked out of the market.

I think there's like a lot of conversations that people should have around that that aren't kind of, you know, hyperliquid did this. It's positive. Lighter did this. This is the right way. Should be like more design focused. And so what would be an example of like an improved, you know, way to set that up? I mean, I don't know if it's like improved as the right word because I don't necessarily

know the answer, right? Should you prioritize, in my opinion, capital and crypto is, relatively cheap. So I don't think the idea of like prioritizing capital on its own, like sure, the HLP made an absurd amount of money in a very short time frame. Is making that amount of money worth running over functioning a lot of market makers that were on the platform? Because suddenly, you know, sure, they had full uptime, but their short positions were

closed out. and also I think that's a larger discussion on the hyperliquid side of like almost all the money goes to buybacks if you didn't use all the money for buybacks you could have money to just have the HLP be protocol run and then at that point you don't even have to have this like mercenary capital problem it's like okay it's making like 15 30% a year that's not fantastic compared to defy yields maybe but it's like a sunk cost for the business and the event something like this happens, we can take a hit and we can allow traders to know that their position is going to be safe versus on other exchanges, people know that their position is not going to be safe. But that's still a question of like give and take. I don't know actually if that's the right decision. I just think that it's a conversation that should be had because I think people kind of before now were like ADL is very theoretical, but it's no longer very theoretical.

Okay. And so I just want to make sure I understood. So basically you're saying that instead of doing so many buybacks, they could just put more money in the HLP. And then that would allow them to have fewer auto to leveraging. So the users wouldn't necessarily like lose money. And but then. It's basically a question of like more broadly outside of the of hyperliquid specifically. how much responsibility should exchanges take for prioritizing user trades being respected versus avoiding some kind of downside for the exchange, given they're such profitable companies. Like kind of the system right now is like the exchanges make billions of dollars a year. And as soon as the system's unhealthy on its own, they kind of are like, oh, like, you know, this is unhealthy, we're going to do ADL. in practice maybe you should have a more proactive way of like hey we made you know three

uh three billion dollars this year a billion dollars is going to be set up for situations like this um and then if that because i think another big thing is like people don't fully understand when adl happens or why it happens versus if you have a very clear policy of like if we're having to provide a billion dollars worth of liquidity and basically no one else is providing liquidity after that point an auto de leveraging event happens. But I and then there's the other side of like well if you telegraph that then people are going to freak out even more. But but that's kind of what I mean is like all of this ADL was kind of a very theoretical

thing that people didn't really think they would have to be concerned about. I think probably there will be a much larger derivatives crash. in crypto in the future. At that point, it's just going to happen, right? If we think that the space is going to grow more, it's going to happen. There's going to be more exchanges. It's going to be more, you know, people are going to offer 100x, 200x, whatever. And, you know, perps are going to become legal in the U.S.

in six to 12 months. So what's that going to do when, you know, every Robin Hood trader can go long on like three to five X? And when that, when the, When this kind of situation happens again, I guess the question should be how much responsibility should the exchange directly take? How much responsibility in like defy system should like LPs take? Because the LPs have extraordinary rights as LPs in the HLP and in the HLP and in the and in the LLP. sorry, I haven't slept.

In the lighter community? Yeah, as in, no, as in they get exclusive rights over, you know, they have kind of priority and how orders are filled. HLP gets basically all the liquidation flow. So in HLP's example, for example, sorry, I'm repeating myself. With HLP, all the liquidation fees went to the HLP. So in theory, they were up like, 40 million and more.

But they still did ADL, even though hyperliquids, HLP is supposed to, you know, the kind of tradeoff with these vaults is you are the provider of liquidity of last resort. Like you get all of these advantages in order to provide this liquidity.

And now we have a situation where that liquidity is really needed. And, you know, the ADL is still happening. And, you know, they get the kind of the good side of like all these liquidations go through and they get all the fees, but the system just refused to make orders at the same time. So then they still had to do ADL. So maybe that's the right setup because you need capital

and because the capital is provided by outside investors. But that's, I think, the conversation that people have to have. Oh, sorry. I mean, honestly, like when you paint that scenario, my first thought is what's going to happen is for any exchange that offers that, the customers on that particular exchange will probably feel more emboldened, I think, to like, yeah, make riskier trades. Or, I don't know, emboldened maybe isn't the right word, but like, it will lead, it will lead to different behaviors on different exchanges. So it almost like, it's like by telegraphing that we will not, you know, like auto de-leverage you at the same, you know, threshold that another exchange might.

It, I feel like people will, you know, account for that. Yes. I think that all of due leveraging, though, it affects the people that are on the side that the exchange actually wants to be adding positions, if that makes sense. Like, the people that are going to get liquidated are going to get liquidated. That's just happening. The auto-de-leverging side is the people that are in the profit now are being forced to basically take profit. Oh, I see what you're saying.

So you can argue that it's actually the opposite because as soon as ADL happens, suddenly everyone's like, wait a second, even if I place a position with how the APIs are working, I could place a position. In theory, my system says I'm up, you know, 100K, I'm moving all of these different things around. and then I find out that I got ADLed and made 10K. Like suddenly you have no way of knowing as the person on the right side of the trade, the side of the trade that the exchange wants in theory, you have no way of knowing your trade is going to be respected. So I think that's true to an extent in the sense that like you could have like a weird situation where a bunch of people just think that like they're right about a trade or something like that

and they're like, okay, so I'm going to put heavy leverage on this. I'm not, this isn't going to happen. And then they're wrong. But I think it's actually quite positive in general. Because in theory, if you have the right participants, it'll, it'll ensure people that even if they place a trade and it might take a bit to fill, their trade's not going to be basically closed out for them. Because otherwise, if you're in an incredibly volatile, market and you're also dealing with the fact that you don't you have no idea if your trade will actually be kept open or closed it's like impossible to trade so it's like that's why it's kind of a

nuclear button to use um but everyone used it today yeah okay well i did want to also ask you so at the end of your piece you kind of basically said that you feel like people really should consider the choice of venue where they trade and it's like you know, a consequential choice. So, like, what are the different factors that people you think should look at or like, what do you look at? You know, what's important? I mean, I think it's impressive. There's a bunch of different factors here. On the just the tax side, it's like, it's objectively very impressive that hyperliquid had no serious downside. So if you're someone that trades very, very actively or systematically, that's probably a big sign to you because

Binance and Bybit basically weren't a good platform for you because you literally couldn't do any of your strategies because all of these APIs were closed. So it's going to add more juice to the hyperliquid side and the hyperliquid finance floors, which is interesting. Yeah. Yeah. It definitely is. Also, you know, ByBit came out and was like, hey, there were scenarios where people didn't have access.

that's not okay and we're going to refund users. Binus, I don't think did that. Lighter also came out and said, we're going to refund some users. Actually, Binance did, I think. Did they? I didn't see it. Yeah.

So, you know, Richard Ting didn't, he didn't in his. I think Haye did. I think she, yeah, I saw news article saying that at least. Okay, then that's great. But like, I think in these moments, it's important to notice, like, who, which it depends because there's so many different axes of like what matters. But for each trader, if you just care about, like say you were aggressively short after the Trump news, you didn't really care about trading while all this mayhem was happening.

You probably should have been on lighter because they didn't close out your position. If you're a more systematic person, you probably should be on hyperliquid because if you're on finance and by a bit, you got blown out. Like I think all of these, this is also like a very natural kind of thing that happens with exchanges in traditional finance and that it becomes kind of specialized for whoever's using that venue. And as more and more of these venues pop up, you might see venues that are basically entirely filled with market makers or institutions or retail. and then people kind of balancing out positions between these different exchanges. But as a user, you should just kind of know the risks. And in theory, I think we've, the perp meta, if you can call it a meta, has made a lot of people forget that like actually, you know,

exchanges are very different. They have different ways of calculating different things. They have different times when like ADL is decided to be done or to be used. So you should kind of try as much as you can. to understand the risks that your exchange has. Okay. Are there any other like winners and losers you want to call out from the past 24 hours? I'm not sure.

I heard that a jump might have lost a lot. But I don't know if that's confirmed. The person that told me is like market making related. And they said apparently there's like an API that suggests that jump lost like a billion dollars. which if true is pretty nuts because in theory they managed nine billion. Oh shit. But yeah, I mean, that's the only like concrete thing. Everyone else seems to be okay.

I mean, I think Salini got hit for a decent amount, but they're mostly prop, so it's fine. Winner Mead obviously lost, like obviously what market makers lost a lot of money. The magnitude of that is what we're going to find out April 10th, as you said. All right. Well, is there anything I didn't ask you that you feel like we should talk about before we wrap? I think you basically covered everything. Okay. All right. Well, it was super fun unpacking the whole thing with you. Oh, my God. That was wild. And kudos to you for doing this after you didn't sleep for the past 24 hours. Yeah. I mean, I'm kind of a sicko in that I enjoy this stuff. So it hasn't done that hard.

Running on adrenaline. I mean, yeah, and I've just been looking at charts and doing some, like, some of the USDA stuff I caught, but just looking at stuff and being like, wow, this is like properly chaotic. Like, it very much reminded me of the FTX collapse. I haven't seen anything this aggressive since then. I know. I know.

And, yeah, just to what we were saying earlier about how it sort of came out of nowhere, like, you know, the FTX thing, like you could kind of understand. because there was a buildup with like the FTT and the CZ tweet and Caroline and, you know, and so the fact that this was so much bigger and it truly just felt like it came out of nowhere. Grinidad, I'm not, you know, the person who's like following Trump's tweets or anything about like China and rare earths and whatever and maybe I should be. But so maybe somebody else was like, oh, I could have seen that coming. But yeah, I think now everyone is going to, we're going to see on Twitter,

everyone's going to become a rare earth medals. expert within the next like two days. Exactly. All right. Well, Dio, you know, good luck to you. Thanks for having me on. Yeah, it was great chatting. I'll catch you later. Thank you. Unchained is produced by Laura Shin with help from Juan Oranovich, Margaret Curia and Pam Majumdar. Thanks for listening.