We don't need ADL: A fourth solution
The crypto clearing trilemma: Why algorithmic optimization cannot escape socialized losses
The fundamental challenge in crypto derivatives clearing is not optimization—it's architecture. Tarun Chitra's December 2025 proof https://arxiv.org/abs/2512.01112 that no ADL policy can simultaneously satisfy solvency, fairness, and revenue confirms what the October 10 Hyperliquid cascade demonstrated empirically: even mathematically optimal socialization still penalizes winning traders. The critical insight for infrastructure investors is that three competing camps are debating how to distribute losses, while we advocate for a fourth approach—securitization—transfer risk to willing capital markets participants entirely. This distinction between socializing optimally versus transferring explicitly represents the largest alpha opportunity in derivatives infrastructure since the founding of modern clearinghouses.
Three camps argue over loss distribution while missing the structural flaw
The crypto derivatives clearing debate has crystallized into three distinct positions, each with prominent advocates and substantial technical merit, yet all sharing a common assumption: that counterparty risk must ultimately be absorbed by market participants rather than transferred to external capital.
Position 1: Current ADL mechanisms are "heuristics from 2015." The industry's dominant clearing mechanism—queue-based auto-deleveraging ranking positions by profit × leverage—was designed by BitMEX developers nearly a decade ago and has been copied across 95%+ of derivatives volume. "On October 10, 2025, when $19 billion in positions liquidated across exchanges in 24 hours, The Coinomist Hyperliquid's queue mechanism closed $2.1 billion in 12 minutes while overshooting optimal closure by 28× and imposing $653 million in unnecessary haircuts on profitable traders." The mechanism's "positional blindness"—inability to recognize hedging strategies—destroyed hedging utility in correlated positions, triggering cascade liquidations that explain why Hyperliquid lost 50% of open interest while competitors like Lighter recovered to pre-event levels. This means they can't be used for hedging, so can't underwrite RWA, they can't be used for Stat Arb as ADL exits one side at a time. They are fixed as gambling instruments.
Position 2: Optimized ADL through convex optimization. Chitra's arXiv paper (2512.01112v2) represents the first rigorous academic treatment of this problem, proving an impossibility theorem arXiv and proposing three algorithmic improvements: water-filling algorithms for fair haircut distribution, Risk-Aware Pro-Rata weighting by expected future deficit contribution, and Mirror Descent with incentive compatibility constraints achieving O(√T) regret bounds versus linear regret for static policies. Thankfully my Maths hasn't failed me. The technical elegance is clear separating severity optimization (total dollars impacted) from haircut allocation (per-trader percentages) enables tractable solutions. Yet even optimal ADL still socializes losses among traders; it merely does so more fairly. Ever that was the way we have a solution to optimise unfairness!!
Position 3: Traditional CCP infrastructure. Don Wilson's @drwconvexity October 2025 commentary articulated the case bluntly: "Auto-deleveraging isn't risk management—it's a failure mode." Cumberland's @Cumberland_DRW DRW argument is that crypto should import the FCM-CCP structure that didn't even trigger forced deleveraging during Lehman's collapse. The traditional model interposes FCMs as credit intermediaries, maintains mutualized default funds sized to Cover-2 standards, and uses variation margin gains haircutting only as an extreme backstop. Wilson's investments in ErisX (now Cboe Digital, a registered DCO) and Digital Asset (Canton Network) demonstrate commitment. The limitation: this approach requires trusted intermediaries, legal enforcement mechanisms, and clearing member identity verification incompatible with permissionless 24/7 markets. The downside no one in crypto has direct market access, traditional FCMs don't have infrastructure and are slow to evolve, tokenised collateral is poorly supported and treated separately from the portfolio. Essentially this is the tortoise in the hare and the tortoise but we want to enable the hare to finish not just back the tortoise.
Our position is Position 4: keep reading.
The October 10 cascade exposed ADL's fundamental limitations at scale
The empirical data from the largest liquidation event in crypto history provides concrete evidence of current mechanism failures. Bitcoin dropped from $126,000 to $102,000 in hours following Trump's tariff announcement, triggering cascading liquidations across venues.
Hyperliquid-specific data reveals the overshoot problem. True negative equity requiring ADL closure totaled approximately $23.2 million. Actual ADL execution reached $676.2 million—an overshoot of $653 million representing 28× the necessary intervention. The queue mechanism prioritized highest profit × leverage positions, which mathematically describes market maker inventory hedges. CoinDesk When one leg of a correlated hedge was ADL'd first, the remaining exposed position triggered follow-on liquidations. Selini Capital reported $50-70 million in losses across three sub-accounts; market depth collapsed 98% as liquidity providers withdrew.
Cross-exchange comparison highlights structural differences. Lighter—backed by Robot Ventures @robotventures , Founders Fund @foundersfund, and Ribbit Capital @RibbitCapital at a $1.5 billion valuation—did not activate ADL despite the same market conditions. Their LLP vault absorbed losses collectively (-5.35% drawdown), preserving trader positions at LP cost. Lighter went offline for hours due to centralized sequencer limitations (79.8× normal transaction volume), but the system maintained solvency without forcibly closing winning positions. The tradeoff: LPs bear losses versus profitable traders bearing losses. Though nothing here breeds confidence that this approach is always repeatable.
Open interest recovery divergence is the market's verdict. Post-October 10, Hyperliquid's open interest dropped from $13.8 billion to approximately $6.4 billion—a 50% decline—while Binance and Lighter recovered to pre-event levels. Market share in decentralized perpetuals fell from 71% to 38%. Jeff Yan's defence that ADL "made users hundreds of millions by closing profitable shorts at favourable prices" BitcoinEthereumNews.com misses the point: the mechanism's opacity and arbitrariness drove capital to venues offering more predictable risk.
Chitra's trilemma proof has genuine technical merit but confirms the architectural limitation
The impossibility theorem deserves serious engagement because it formalizes intuitions practitioners have held since 2016. Chitra proves that no ADL policy can simultaneously satisfy three desirable properties:
Solvency: Exchange maintains assets ≥ liabilities at all times
Revenue: Exchange maximizes long-term fee income and insurance fund preservation
Fairness: Losses distributed equitably among winning traders without disproportionate haircuts
As participation scales, a novel form of moral hazard grows asymptotically—losing traders have incentive to take more risk knowing winning traders will absorb losses. The trilemma cannot be "grown out of."
The proposed solutions are mathematically elegant. Water-filling algorithms minimize maximum individual haircuts through Schur-convex optimization. Risk-Aware Pro-Rata weights haircuts by each position's contribution to future expected deficit using perspective transforms for convex dominance. Mirror Descent treats ADL as a Stackelberg game between exchange (leader) and traders (followers), incorporating Follower and Leader Incentive Compatibility constraints to address the "waiting game" where traders delay exiting hoping prices recover.
Yet the trilemma persists regardless of optimization. Even optimal ADL still socializes losses among winning traders. The question isn't which algorithm to use—it's whether socialization is the right architecture at all. Chitra's algorithms are valuable for sizing tranches in a securitized structure, but they don't escape the fundamental problem that forcing profitable traders to absorb losses destroys capital formation incentives for sophisticated participants. This is where we were, Pascal was on-chain VaR, it was as optimised a solution to this solution as processing capacity allowed. On chain entirely though is to hard now, and also brings the issue that buyside adoption is not there on scale as they have no way to assess smart contract risk. Audits being relatively manipulable and valueless.
Investment conflict disclosure matters for positioning. Robot Ventures, where Chitra is Managing Partner alongside Compound founder Robert Leshner, invested in Lighter—a direct Hyperliquid competitor operating as an L2 on Ethereum versus Hyperliquid's L1. The $68 million November 2025 round at ~$1.5 billion valuation from Founders Fund and Ribbit Capital Fortune followed Robot's earlier participation. Gauntlet, Chitra's other venture, provides risk analysis services to DeFi protocols. The academic work is rigorous regardless, but investors should understand the competitive dynamics: optimal ADL research serves platforms implementing algorithmic clearing it still inherently suffers the limitation of pricing the different thresholds. Securitisation is a much more elegant solution to tail risk issues.
Traditional CCP advocates correctly diagnose the problem but prescribe incompatible solutions
Don Wilson's argument deserves more engagement than crypto-native critics typically provide. DRW traded through the 2008 crisis and knows that traditional clearing infrastructure weathered Lehman's collapse—the largest counterparty default in history—without triggering forced deleveraging of winning positions. The FCM layer provides credit judgment and capital cushioning; batch margining allows time for margin calls; mutualized default funds spread risk across members who influence CCP governance. Cumberland The structural differences Wilson emphasizes are real. CCPs employ default waterfalls with defined loss allocation: defaulter's initial margin → defaulter's default fund contribution → CCP "skin in the game" → mutualized default fund → unfunded commitments → end-of-waterfall mechanisms (VMGH, assessments). The Cover-2 standard requires funds sufficient to cover simultaneous default of two largest members. Members have governance voice over risk policies. Legal systems provide dispute resolution.
But crypto's permissionless architecture is incompatible with FCM intermediation. Traditional clearing requires identity verification, membership agreements, capital requirements, and legal enforcement mechanisms. These prerequisites make sense under the world of traditional finance, but many crypto entities don't have USD, don't have compatible corporate structures. There is no discount FCM intermediation game in town either. This works for the largest market makers but not for the vast majority of their counterparties. "Paper" drives the market, we can't look at the volume these solutions can cover but the volume of "Paper" these solutions can cover. I would argue that is diminishingly small. It is the reason CME is so big in stats in crypto and so small in market presence. Wilson's vision requires either abandoning permissionless markets, on-chain activities or creating parallel infrastructure serving only identified institutional participants—neither of which addresses the $23 trillion annual derivatives volume occurring on existing platforms. Personally I think this conflicts with the goals of Canton, that want to bring that accessibility. We see it the other way around, bringing market like FX Swaps IM/VM into blockchain settlement not crypto into legacy FIS platforms.
The relevant question is what crypto-native infrastructure achieves CCP-equivalent risk transfer. Canton Network @CantonNetwork, which DRW invested in via Tradeweb's $135 million round, Tradeweb Markets demonstrates institutional blockchain infrastructure for settlement and collateral mobility. But Canton addresses different problems: repo financing, cross-margining, and atomic settlement between identified counterparties. The gap is tail risk transfer from permissionless derivatives markets to willing capital markets investors.
Securitization represents the fourth way: explicit transfer versus implicit socialization
The structural innovation that neither ADL optimization nor traditional CCP importation provides is transferring tail risk to external capital markets participants who explicitly choose to bear it in exchange for compensation. This is how traditional derivatives markets handle concentration risk through credit default swaps, how reinsurance markets handle catastrophic losses through CAT bonds, and how CLO markets provide $1.4 trillion in credit intermediation without forcing loan originators to absorb defaults.
The CLO precedent demonstrates that proper structuring survives extreme stress. AAA-rated CLO tranches have a lifetime default rate of zero percent across all rated issuances since 1994. VanEck Oaktree Capital During the Global Financial Crisis, only 0.88% of ~$500 billion in U.S. CLOs experienced any default versus CDO meltdowns. Oaktree Capital The difference: CLOs are backed by diversified senior secured corporate loans with 60-70% recovery rates, actively managed with concentration limits, and structured with meaningful subordination (35%+ credit enhancement for AAA tranches).
The CAT bond model demonstrates tail risk transfer at market prices. The $43 billion catastrophe bond market transfers 1-in-100 to 1-in-250 year event risk from insurers to capital markets at explicit premiums averaging ~5.4% over risk-free rates. Federal Reserve Bank of Chicago Positive annual returns in 21 of 22 years. Morningstar Near-zero correlation with financial markets. ScienceDirect Trigger mechanisms can be indemnity (actual losses), parametric (index-based), or modeled. Marinpost The structure: sponsor transfers specified tail risk to SPV → SPV issues bonds with proceeds held in collateral trust → investors receive collateral return + risk premium → if trigger occurs, principal transfers to sponsor.
Applied to crypto derivatives clearing, the architecture would tokenize clearing fund exposure into tranches. Senior tranches (attachment: 10-15%, spread: SOFR+100-200bps) attract traditional institutional investors seeking uncorrelated yield. Mezzanine tranches (attachment: 5-10%, spread: SOFR+400-800bps) serve hedge funds and crypto-native capital. Equity/first-loss tranches (attachment: 0-5%, target IRR: 15-25%+) absorb expected tail losses and could be held by the clearing platform itself or sophisticated investors. Junior tranche pricing is market set and the art in the middle of this.
Critical advantages over both ADL and mutualized default funds:
Explicit consent: Investors choose exposure for compensation; winning traders aren't forced
Market pricing: Tranche spreads reflect actual tail risk assessment versus algorithmic allocation
External capital: Brings new capital into clearing infrastructure versus redistributing existing capital
24/7 native: Smart contracts automate waterfall mechanics without FCM intermediation
Regulatory alignment: CFTC's December 2025 pilot permits BTC, ETH, USDC as margin collateral;
The competitive landscape reveals infrastructure gaps worth hundreds of billions
Understanding who's building what—and what's missing—is essential for infrastructure investment positioning.
Canton Network dominates institutional blockchain infrastructure. Digital Asset's platform processed the first fully on-chain U.S. Treasury financing against USDC via Tradeweb in August 2025. Traders Magazine The $135 million round led by DRW and Tradeweb Tradeweb Markets plus ~$50 million from BNY, Nasdaq, and S\u0026P Global demonstrates institutional commitment. Partners include Goldman Sachs, Citadel Securities, DTCC, and BNP Paribas. Tradeweb Markets Supporting $4.5 trillion+ in annual trade flows via Daml applications.Digital Asset But Canton addresses settlement and collateral mobility for identified counterparties—not tail risk transfer from permissionless markets. Regulated venues are expanding crypto derivatives aggressively. CME Group recorded $900+ billion in combined crypto futures/options volume in Q3 2025, with average daily open interest of $31.3 billion. CME Group Ainvest SGX launched Asia's first regulated bitcoin and ethereum perpetual futures in November 2025. Morningstar The TRADE Cboe launched Continuous Futures (10-year expiration with daily cash adjustments mimicking perpetual swap economics) in December 2025. Ainvest Crypto.com Derivatives North America holds the complete CFTC stack: DCM, FCM, and DCO licenses. FintechNewsSG The DeFi perpetuals market is consolidating rapidly. Hyperliquid holds approximately 38% market share post-October decline, generating $1.24 billion in annualized net income with 11 employees—capital efficiency unprecedented in financial infrastructure. Tech Startups Lighter's $1.5 billion valuation from Founders Fund and Ribbit Fortune represents conviction in the L2-on-Ethereum approach versus Hyperliquid's proprietary L1. Ostium raised $20 million from General Catalyst and Jump Trading for perpetuals infrastructure. Fortune Grvt raised $19 million for privacy-focused onchain derivatives. Cointelegraph
The gap is tokenized clearing fund structures. No project currently offers securitized tail risk transfer for crypto derivatives. The CFTC's December 2025 pilot creates regulatory pathway. Davis Wright Tremaine Tokenization standards exist (ERC-1155 for tranching, ERC-1400 for compliance controls). CAT bond and CLO precedents provide structural templates. The missing piece is execution: an issuer structure, tranche definitions, trigger mechanisms, and investor distribution.
Investment thesis: Securitization captures $60 trillion market opportunity that algorithms cannot
Chitra's cited figure of $60+ trillion in annual perpetual futures volume represents the addressable market for clearing infrastructure. arXiv Current ADL mechanisms socialize approximately 0.1-0.5% of this volume annually during stress events—$60-300 billion in potential tail risk transfer annually. At CLO-equivalent spreads (150-400bps across tranches), this implies $1-4 billion in annual premium income for securitized clearing structures.
Capital efficiency improvements justify institutional adoption. Portfolio margining on platforms like Bullish reduces capital requirements by approximately 40%. Ainvest Cross-margining between CME and DTCC is expanding. But these efficiency gains assume clearing infrastructure that doesn't forcibly close positions during volatility. Institutional allocators cite ADL exposure as the primary barrier to scaling crypto derivatives positions. Securitized structures that transfer tail risk to willing investors remove this constraint.
The 18-month window exists before traditional players adapt. CME's 24/7 trading plans (early 2026), CFTC's tokenized collateral pilot (operational H1 2026), and Canton Network's scaled deployment create convergence pressure. First-mover advantage in securitized clearing structures captures network effects before incumbents integrate equivalent capabilities into existing infrastructure.
Revenue models parallel traditional structured credit. Arrangement fees: 1-2% of issuance. Ongoing management: 25-50bps annually. Performance fees on equity tranches: 15-20% of returns above hurdle. Distribution revenue from tranche placement. Technology licensing for smart contract infrastructure (Pascal).
Technical requirements for credible implementation
Infrastructure investors should evaluate clearing ventures against these criteria:
Trigger mechanism design determines viability. Indemnity triggers (actual clearing losses) create moral hazard if platform controls loss determination. Parametric triggers (BTC/ETH price moves >X% in Y hours) avoid manipulation but may not match actual losses. Index-based triggers (total liquidations across exchanges exceed threshold) provide external verification but depend on data availability. Hybrid approaches combining parametric activation with indemnity settlement merit consideration.
Collateral structure must satisfy regulatory requirements. CFTC guidance now permits BTC, ETH, and USDC as margin collateral under the December 2025 pilot. CoinDesk Commodity Futures Trading Commission Tokenized Treasuries and money market funds qualify under existing frameworks. Bloomberg Real-time settlement capabilities (vs. T+1) provide liquidity advantages. Segregation requirements and bankruptcy remoteness remain critical.
Waterfall mechanics require smart contract audit. Priority of payments: insurance fund → equity tranche → mezzanine → senior. Overcollateralization and interest coverage tests triggering automatic deleveraging if collateral deteriorates. Governance mechanisms for parameter adjustment. Liquidation procedures if triggers breach multiple tranches.
Sybil resistance prevents manipulation. Tokenized clearing fund shares must be non-transferable during risk periods or subject to lock-up. Governance tokens should be stake-weighted with slashing conditions. Identity verification for senior tranche investors (accredited/QIB standards) enables regulatory compliance without compromising permissionless trading.
Risk factors and structural considerations
Secondary market liquidity will develop slowly. Initial tranches should assume hold-to-maturity investment horizons. Senior tranches may attract repo financing against tokenized Treasury collateral. Mezzanine and equity tranches remain illiquid until trading venues develop.
Regulatory classification uncertainty persists. Tokenized clearing fund securities likely require registration or exemption (Reg D, Reg S, Rule 144A). CFTC/SEC jurisdictional overlap creates dual compliance burden. Cross-border treatment varies; Cayman SPV structures may avoid certain U.S. classifications while maintaining operational substance and proof of concept.
Correlation risk during systemic events. October 2025 demonstrated cross-asset contagion 20% stronger than 2018 trade war spillovers (per academic analysis). Clearing fund losses may correlate with broader portfolio drawdowns for institutional investors. Senior tranche sizing must account for tail correlation.
Model risk in tranche pricing. CLO and CAT bond pricing models assume historical loss distributions that may not apply to crypto markets with limited history. Parameter uncertainty demands risk premium multipliers (2-5× expected loss, per CAT bond precedent). Conservative attachment points (higher subordination) appropriate for initial structures.
The path forward requires synthesis, not advocacy
Chitra's optimization work, Wilson's infrastructure experience, and securitization precedents are complementary rather than competing. Risk-Aware Pro-Rata algorithms can size tranches based on expected deficit contribution. Traditional CCP waterfall structures provide tested priority-of-payments frameworks. Tokenization enables 24/7 native implementation without FCM intermediation.
The strategic insight is architectural, not algorithmic. Even perfect ADL still forces winning traders to absorb losses they didn't choose. Even perfect CCP infrastructure still mutualizes default risk among clearing members. Securitization transfers risk to investors who explicitly price and accept it. This is not a minor optimization—it's a fundamental shift in how derivatives infrastructure forms capital and allocates risk.
For high-end infrastructure investors, the opportunity is building the first credible tokenized clearing fund structure that achieves native crypto operations while providing explicit tail risk transfer to capital markets. The technical pieces exist. The regulatory pathway is opening. The market gap is clear. Execution is the remaining constraint. Opportunity For infrastructure investors, the opportunity is building the first credible tokenized clearing fund structure that achieves native crypto operations while providing explicit tail risk transfer to capital markets. The technical pieces exist. The regulatory pathway is opening. The market gap is clear.
This is what we're building at DRN.
The architecture is defined. The legal framework is structured. The technology foundation including from Pascal Protocol work on the risk aspect provides the risk calculation engine. What separates this from analysis is execution capital and the 18-month window before either traditional players adapt or well-capitalized competitors complete DCO registration.
The first entity that deploys securitized clearing correctly doesn't just capture market share—it establishes the structural position for how institutional capital enters crypto derivatives markets.
If you're an allocator focused on derivatives infrastructure or have institutional relationships that benefit from solving the credit intermediation gap in 24/7 markets, let's talk.