Exorcism: DeFi Purges CeFi

_Decentralized_ protocols survive as designed -- while also mitigating the fallout from 3ac/Celsius/CeFi's irresponsible, "regulated" off-chain overleveraging

June 17, 2022

Celsius and 3ac were the second and third whales killed in the crypto bear market of 2022. Celsius’s highest reported AUM was $28.6 billion at the end of 2021, or almost 15% of DeFi’s peak AUM (~$250B). Although Celsius had a bigger retail profile, we believe Celsius was mostly a magnet for institutional DeFi flows, which require custodied assets for compliance reasons. (Retail has nothing but downside from third-party custody, but handling large amounts of other peoples’ money requires operational safeguards against the account holder stealing the money, i.e., custody.) Celsius’s implosion, while a failure of centralized management and risk control, also revealed severe weaknesses in the Curve-based cross-chain liquidity model.

3ac was a different kind of disaster. Although technically a “family office” with no outside capital, 3ac borrowed at least as much in outside loans as its AUM (reportedly $18B at peak) and was given extremely loose lending terms. 3AC’s dealings with centralized lenders obscured the extent of 3ac’s leverage and double-dealing. 3ac’s losses to the outside world will come from defaulted loans from these centralized counterparties, as 3ac’s liquidated assets are not close to covering their outstanding liabilities. Centralized counterparties with significant 3ac lending exposure seemingly include Voyager ($250-300M), Babel Finance ($100M+), and many VC treasuries who foolishly outsourced their treasury management to 3ac in return for 8-10% yields.

These two debacles have 2 things in common. One, reckless, centralized risk management. Two, because of the transnparency of Celsius’s and 3ac’s leveraged DeFi vaults on Aave/Maker/Compound, 3ac’s blowup was much smaller as a percentage of AUM (20-40% of what was probably an $8B-ish effective AUM) than the Archegos disaster (~100% of $10B AUM). The same goes for Celsius.

Celsius, 3ac, and stETH’s fatal Curve dependency

As with Terra, the Celsius liquidity crisis started on Curve — the Curve ETH-stETH pool, to be exact.

Curve is the go-to source of deep liquidity for asset pairs that should trade similarly. The two types of assets in this bucket are 1) stablecoin pairs, and 2) a native asset vs. its wrapped, cross-chain derivative (ETH vs Lido liquid-staked Beacon Chain ETH, or stETH).

The blowups of Celsius, 3ac, and ETH/stETH are all very closely related. Terra’s crisis spiraled, Curve LPs presumably withdrew massive liquidity from the ETH-stETH pool at the same time as Terra imploded, worried about the Lido Protocol’s overall Terra exposure (Lido is the largest stETH liquid staking provider), as well as Celsius’s exposure to Terra (it was always known that Celsius was a top holder of stETH and also a major player on Anchor). 3ac, it should be said, was also known to be a major stETH holder as well as a top LUNA holder.

Celsius, however, escaped the Terra crisis nearly unscathed: Celsius’s withdrawals from Terra were the immediate catalyst for Terra’s liquidity death spiral, but Celsius itself dodged the bullet.

However, stETH already began dumping on Curve (the first yellow spike below), probably due to concerns about Celsius and actual dumping by 3ac in the immediate wake of their Terra losses. This caused an immediate spike in Curve LP fees for the ETH-stETH pool (the first big yellow spike on the right).

However, the risk/reward for LP’ing on Curve turns very negative when volatility passes a certain threshold, causing the given Curve pool to death-spiral.

When a Curve pool reaches a certain level of imbalance between two assets (typically 80-90% in a 2-sided pool depending on the “a” amplification factor), the pool is unable to execute further large redemptions of the imbalanced asset (stETH). This causes trading activity to dry up in the pool—which reduces LP providers’ fees relative to the directional risk they take in offering to accumulate more of the out-of-favor asset (stETH). Which in turn causes LPs to withdraw liquidity from the pool … which exacerbates the illiquidity of the asset being withdrawn (stETH in this case), which incentivizes further withdrawals of the stETH being dumped, which in turn incentivizes more Curve LPs to withdraw liquidity.

stETH has crossed that event horizon. Why does that matter? Because the amount of stETH vaulted & borrowed against dwarfs the amount of ETH available on Curve to process a potential liquidation cascade.

If I am reading Lido’s public risk dashboard correctly, there are approximately 80k stETH at a “B-” risk rating, meaning that if stETH drops 20-33%, that stETH will face a margin call. However, there’s also 100k stETH facing a margin call at a 7-15% further deviation from ETH. and another 1.1 million stETH facing liquidation from a 33-43% drop from current prices.

The latter liquidation scenario may sound far-fetched, except that if ETH dropped, say, 20% from here and a stETH liquidation cascade started, the illiquidity in stETH would cause stETH’s discount to ETH to blow out very quickly. A 25% drop in ETH could cause enough of a liquidation cascade to push the stETH/ETH discount to .80 or less, tripping the 35% net drop of stETH value that would trigger total liquidation of “B-” stETH, and commence liquidation of B-rated stETH … dumping more and more stETH into a more and more illiquid air-pocket of a market.

A fire-sale liquidation of “B” stETH at around $500 USD equivalent (if ETH traded at $800 and stETH traded at a .65 discount to ETH), however, would probably mark a major low in stETH, a clearance event of the biggest supply overhang, and a potential generational buying opportunity.

To some extent, Curve has been doing the heavy lifting through this crisis that no other exchange can, and deserves to be applauded. But investors need to realize that in a highly volatile environment, once a Curve pool passes the 80% event horizon of imbalance, only a dramatic macroeconomic/flows recovery can rescue it; otherwise, deteriorating LP risk/reward will incentivize LPs to withdraw liquidity from those pools, potentially start dumping the imbalanced token on Uniswap v3 or elsewhere, and incentivize further de-risking of the imbalanced token on Curve. If your investment depends on cross-chain access or Curve liquidity to survive, you should probably reconsider it. If you are a Curve LP, you should be extremely paranoid once a 2-sided Curve pool crosses even a 65% imbalance.

Celsius at least appears to have stabilized (for now). While Celsius appears to have have nefariously found enough additional liquidity to lower its vault liquidation thresholds and pay down some of its loans, the fact that it muscled its customers to give them more money probably means the end of their brand, and a slow death instead of a fast one. Or, crypto prices could drop another 30% from current levels, and completely liquidate Celsius’s remaining capital (at terrible losses to the Celsius “community”).

DeFi’s dependency on Curve-style leveraged liquidity, however, is coming under increasing strain.

3ac’s centralized leverage

Nobody knows how bad the damage is from 3ac’s collapse. What’s clear thus far is that 3ac borrowed billions from a who’s who of centralized exchanges as well as 3ac VC-invested treasuries. 3ac’s downfall very closely resembles the Archegos disaster on Wall Street, when Bill Hwang leveraged multiple prime brokers’ greed and willful ignorance into allowing him to source insane leverage (10x on an equity book, probably analogous to 1-2x on a crypto / crypto VC book). Archegos resulted in $10bn of prime broker net losses vs. Archegos’s peak AUM of $10 billion. If initial estimates are that the crypto ecosystem sustained “only” $1.5 billion of bad debt (somewhat offset by VC investments), then the crypto ecosystem actually avoided a much bigger potential disaster.

Because 3ac is 10 figures in net debt, 3ac’s large VC portfolio will be savagely liquidated as soon as their tokens unlock. This will be particularly bearish for the major ecosystems that 3ac invested in (AVAX, NEAR, SOL, DOT, and web3 gaming in general) as these tokens will be indiscriminately sold. Devs will frontrun these liquidations both in their personal token trading, and in their careers. Like Terra v1 developers, they won’t feel bound to protocols that are forced to dump tokens to pay off a ‘bad debt’ that seemingly came out of nowhere.

How did 3ac borrow so much money?

  • Like any degen Wall Street firm, 3ac traded on their name, to the hilt. Greedy, volume-driven lenders lined up from Singapore to Wall Street to oblige.

  • 3ac relied on the ambiguity of centralized and decentralized sources of liquidity.

  • 3ac muscled its ecosystem relationships to “strongly urge” weak investees to lend their treasuries back to 3ac for 8-10% guaranteed yields. So, an unknown number did (although many refused). The protocols that did, are toast.

In crypto, unfortunately, “on-chain transparency” stops at the doors of centralized institutions. 3ac, much more than Celsius, was a bubble of overconfidence in centralized crypto lenders who are paying dearly for their collective lack of transparency.

Furthermore, it was 3ac’s margin calls on decentralized, vaulted collateral which blew the whistle on the severity of 3ac’s situation.

Meanwhile — decentralized protocols actually worked

Throughout the DeFi death march of 2022, DeFi’s decentralized lenders and exchanges have hummed with grim, liquid efficiency. Maker’s algostable is surviving its second existential macroeconomic crisis at greater scale than the last. Frax’s algostable, though it has not scaled in terms of users, has worked admirably well. No major dexes have reported any execution issues. Curve has struggled, but in a predictable way under unpredictable stress: its model is still moving mountains within expected parameters.

Just as the magnitude of Terra’s failure created aftershocks for weeks, 3ac’s death will cause more dead whales to float to the surface. However, the surviving Class of 2022 is taking shape, and it has a very decentralized look. The centralized protocols, the cults of personality, the crypto mega-funds whose public egos were even fatter than their bankrolls, the “safely custodied” crypto-assets — good riddance to them all. The “safe, custodied, regulated crypto” (Celsius, Voyager, very likely many others that we haven’t heard about yet) has in fact been the least safe corner of crypto, outside of Terra. It has also used the cover of the law to, in some cases, extort its users for more money, or renege on written legal agreements with them.

We are still learning many lessons from this year’s blowups. The baggage and false security of centralization will not be one of them.

The lesson here is that certain hyper-centralized protocols (I’m thinking of one in particular, managed by a certain His Excellency) and stablecoins with asset-liability mismatches and potential cross-chain illiquidity vulnerabilities (I’m thinking of a particular very large stablecoin, 60% of which sits on His Excellency’s blockchain) offer, for the moment, a particularly unattractive risk/reward. (NFA.)

Staking at a Glance

Global value staked plunged in line with the broader crypto market. Interestingly enough, the population of global stakers registered one of its largest-ever week-on-week metrics. Ethereum 2.0 benefited significantly, as buyers presumably started snapping up stETH at a 5-6% discount to ETH.

Avalanche—despite the tsunami of negative news around 3ac, web3 gaming, and ETH alts in general—saw very large staking inflows relative to its overall market cap (+1% of its fully diluted market cap). DOT, another favorite ETH-alt of 3ac, saw the largest outflows

Global Staking Research

Cryptocurrency prices plunge as global market cap falls below $1 trillion

Crypto prices were down across the board on Monday in tandem with broader financial markets as the global cryptocurrency market cap fell below $1 trillion. Bitcoin (BTC) and ether (ETH) led the downturn, reaching the lowest levels in at least 17 months. BTC's price was down over 12% in the past 24 hours, trading at $24,040 at the time of writing, according to Coinbase data via TradingView. Leaving bitcoin at its lowest price since December 2020.

Celsius Freeze on Withdrawals Craters Crypto Market

Cypto lending platform Celsius has frozen accounts of its 1.7 million users, further undermining trust in the digital assets space on a day with double-digit losses for Bitcoin and Ether and just weeks after the Terra collapse.

Tron's algorithmic stablecoin slips further from its dollar peg

Tron's algorithmic stablecoin, called Decentralized USD and known by its ticker USDD, slipped further from its peg to the dollar on Tuesday in a sign of ongoing stress in crypto markets. USDD dipped as low as $0.974 this morning before recovering to $0.979 at the time of writing, according to CoinGecko data.

Chainlink Keepers and Chainlink VRF Go Live on Avalanche

After integrating Chainlink Price Feeds in July 2021 and giving developers access to high-quality price data to power a wide variety of DeFi applications, Chainlink announced that Chainlink Keepers and Chainlink VRF are now live on the Avalanche Primary Network, adding further resourcing for these apps.

Staking Assets

NEAR Foundation and Brave announce partnership to integrate Aurora into Brave Wallet

Aurora will become a preloaded blockchain available in the multi-chain Brave Wallet (in addition to Solana and Ethereum). The integration with NEAR’s EVM comes on the heels of last month’s announcement of a full integration with the Solana blockchain, and shows Brave’s commitment to increasing the pace of development toward an open, multi-chain experience in Brave Wallet.

Ethereum’s Ropsten Testnet Has Completed Its Merge

The Ethereum blockchain’s first dress rehearsal for its upcoming Merge was successfully completed Wednesday. The Ropsten testnet successfully merged its proof-of-work execution layer with the Beacon Chain proof-of-stake consensus chain – a process identical to the one that the main Ethereum network will undergo in just a few months.

After Exploit, Osmosis Relaunched

On June 8, Osmosis chain halted after a bug led to the theft of $5 million from liquidity pools. Now, a week later, Osmosis validators plan to restart the chain Sunday at 16:00 UTC with a new thoroughly tested version. The first 5 blocks will be "Epoch blocks" providing rewards from during downtime.

Binance Labs Makes Strategic Investment in PancakeSwap

Binance Labs, the venture capital and incubation arm of Binance, has made an investment in PancakeSwap’s utility and governance token- $CAKE. PancakeSwap is a decentralized exchange built on BNB Chain. The investment is part of Binance Labs’ mission to facilitate the next wave of global blockchain adoption by providing technology development, marketing and community support, and enterprise solutions for Web3 transformation.

Maiar Exchange: Elrond Incident and Recovery Report

Elrond publishes report about the exploit of Maiar Exchange. On June 5, a lot of suspicious activity started being reported on the Elrond mainnet, including large transfers of EGLD and large swaps on the Maiar Exchange. This report covers all the details of this exploit.

Staking Providers

Lido staked ether selloff continues as larger holder unloads 56,000 stETH

The Lido staked ether (stETH) token discount has deepened further following a large selloff of the token by a major holder. On-chain data from Etherscan show wallet addresses, labeled as Three Arrows Capital in a tweet from PeckShield, selling over 56,000 stETH for ether (ETH). One transaction involved a 17,780 stETH swap for 16,625 ETH which was converted to the dai (DAI) stablecoin for $20 million.

InfStones Nears Unicorn Status with $66 MM Funding Round Led By SoftBank Vision Fund 2 and GGV Capital

InfStones, a leading ****global ****blockchain infrastructure provider announced a new funding round of $66M led by SoftBank Vision Fund 2 and GGV Capital, with participation from INCE Capital, 10T Fund, SNZ Holding, and A&T Capital. The funds will be used for enterprise-wide expansion and to accelerate growth.

🔔 New Asset Integration is Live: Spool DAO

Stake Spool to get staking rewards and 80% of protocol revenue.

Check our their Staking Rewards profile here!

Staking Rewards is Hiring Researchers and Engineers

We are looking for proactive, intelligent, and determined people to join our team across all functions. Positions are fully remote.

We are looking to hire 2-3 more Researchers, to join our team of 3 on a mission to become the unprecedented experts on staking data and economics. Apply here.

We are looking to hire Software Engineers with Web3 experience to develop the data aggregation pipeline for Staking Rewards. Apply here.

We are looking for a People Manager, who will own all matters HR at Staking Rewards. Apply here.

Furthermore, we are looking for a Head of Marketing to lead our marketing efforts. Apply here.

If you know anyone who would be a great fit, please ping us. We offer a generous referral bonus

Reply

or to participate.