The $150 billion in crypto liquidations recorded across 2025 — averaging $400-500 million per day according to CoinGlass data — exposed a structural fault line in decentralized finance. On October 10, 2025 alone, a single flash crash wiped out $19 billion in leveraged positions within hours, nine...
"P + N = 1. Hence, there is no possibility of liquidation." — Vitalik Buterin, Ethereum Co-Founder, ETHResearch Forum Post (June 1, 2026)
The $150 billion in crypto liquidations recorded across 2025 — averaging $400-500 million per day according to CoinGlass data — exposed a structural fault line in decentralized finance. On October 10, 2025 alone, a single flash crash wiped out $19 billion in leveraged positions within hours, nine times larger than any prior single-day event. The forced-selling cascade that followed pushed average token declines to 47%, exceeding the May 2021 crash.
On June 1, 2026, Ethereum co-founder Vitalik Buterin published a research proposal on the ETHResearch forum titled "Building index-tracking assets on top of options instead of debt." The paper argues DeFi should replace collateralized debt positions (CDPs) — the architecture underpinning protocols managing tens of billions in TVL — with options-based structures that make liquidation mathematically impossible. Within 10 days, the proposal moved from theory to working code, with multiple teams shipping implementations and at least one protocol, Cleave, launching on testnet.
The proposal addresses two systemic vulnerabilities simultaneously: the cascading forced-sell dynamics triggered by liquidation mechanisms during market stress, and the real-time oracle dependency that has contributed to hundreds of millions in exploit losses. Whether the architecture can scale beyond testnet remains unproven, but the speed of builder response signals material demand for alternatives to the CDP model that has defined DeFi lending since MakerDAO's 2017 launch.
DeFi's current lending architecture relies on overcollateralized debt. Users deposit assets — typically ETH or stablecoins — as collateral, borrow against them, and face automatic liquidation if collateral value drops below a protocol-defined threshold. The mechanism is designed to keep protocols solvent, but its second-order effects during market stress are severe.
According to CoinGlass data reported by Bitcoinist, nominal crypto liquidations exceeded $150 billion across 2025. The single largest event, the October 10 flash crash, liquidated $19 billion in hours after geopolitical news triggered a global risk-off move. An estimated 85-90% of wiped positions were long. CoinDesk Research calculated average token declines of approximately 47% across their tracked universe during the event.
Aave, the largest DeFi lending protocol by TVL (approximately $76.6 billion as of May 2026 according to Eco.com data), processed $180 million in liquidation sales during the October crash without system failure. The protocol functioned as designed. But "functioning as designed" means forced selling at the worst possible moment — the exact dynamic Buterin's proposal seeks to eliminate.
The problem compounds across the system. When liquidations trigger asset sales, those sales push prices lower, triggering more liquidations. In Q3 2025, crypto-collateralized borrowing reached a record $73.6 billion, with DeFi lending alone at $41 billion — creating the conditions for the October cascade.
Buterin's June 1 paper makes a structural argument: the problem is not how liquidations are executed, but that they exist at all.
Current synthetic assets and stablecoins built on CDP models require three components: overcollateralization, real-time price oracles, and automated liquidation bots. Each introduces fragility. Overcollateralization is capital-inefficient. Real-time oracles are manipulation targets — price oracle attacks accounted for $52 million in losses across 37 incidents in 2024, according to CertiK data. Liquidation bots create deterministic selling pressure during crashes.
Buterin's alternative: use options contracts as the base primitive instead of debt. The core insight is that options, by construction, cannot be liquidated because both sides of the contract are fully backed at all times.
As Buterin wrote in the proposal: "What if we use options as the base of DeFi, instead of CDPs and liquidations?"
The design connects to Buterin's broader thesis, articulated in a February 2026 post, about replacing fiat-pegged stablecoins with AI-powered price index baskets. The June paper addresses the on-chain mechanism needed to implement that architecture without introducing the liquidation risk embedded in CDP models.
The mechanism is straightforward in concept. One ETH is split into two paired claims — labeled P (protected) and N (leveraged) — tied to a strike price and an expiry date. The two positions always sum to exactly one ETH: P + N = 1.
At expiry, an oracle resolves the index value and distributes the underlying ETH between both position holders using a fixed formula:
Because the two payoffs are complementary and the system is fully backed — one ETH in, one ETH out — there is no scenario requiring liquidation. No collateral ratio to maintain. No threshold that triggers forced selling.
The design trades sudden, binary liquidation events for what Buterin describes as "quadratic drift" in index exposure. As prices approach the strike, user exposure decays gradually rather than snapping to zero. Buterin suggested accepting "one to four percent annual standard deviation" as reasonable, noting fiat currencies experience similar fluctuations relative to one another.
Users must actively rebalance positions before expiry to maintain desired exposure — rolling into new options contracts at updated strikes and dates. This is the primary operational requirement replacing the passive monitoring of CDP collateral ratios.
The proposal's second structural contribution may prove more significant than eliminating liquidations: it removes the dependency on real-time price oracles.
CDP-based protocols require price feeds that are accurate, instantaneous, and resistant to manipulation — three properties that are difficult to guarantee simultaneously on-chain. Oracle manipulation attacks, while declining from their 2022 peak, remain a persistent attack vector. According to Immunefi data, oracle and reward manipulation losses fell from approximately $400 million cumulatively in prior years to roughly $70 million in 2025, but the vulnerability class persists.
Buterin's options-based system settles once, at maturity. This allows it to run on what he calls "slow oracles" — the dispute-based resolution mechanisms used by prediction markets like UMA's optimistic oracle. These systems allow extended dispute windows and human review, substantially reducing manipulation vulnerability compared to the instant price feeds that must resolve in a single block.
As Ethereum researcher Andy Guzman noted in the discussion thread, this connects to broader concerns about censorship resistance. A confidential USDC contract freeze incident in late May 2026, where a developer reported funds were "caught in a crossfire of another case," underscored the ecosystem's dependence on centrally-issued stablecoins — and the need for synthetic alternatives that don't inherit those censorship properties.
The proposal's most notable outcome is the speed of implementation.
Buterin published on June 1. By June 11, he posted: "Looks like the options thing is happening already!" Multiple teams had shipped code. According to CryptoTimes reporting:
Buterin cautioned that implementations should be "formally verified first" — mathematically proven to behave as specified — before handling real funds. He flagged oracle robustness as the next critical unsolved challenge.
The 10-day turnaround from research paper to testnet code is notable. It suggests the concept addresses a recognized pain point that builders had been waiting for a credible framework to address.
The proposal enters a market where on-chain options protocols already exist but remain marginal relative to lending and spot trading.
Panoptic operates a perpetual, oracle-free options protocol built on top of Uniswap V3/V4 liquidity pools, replicating call and put positions through LP dynamics. Opyn built options infrastructure on Ethereum. Premia offers a decentralized options AMM. According to DeFiLlama data from prior cycles, on-chain options protocols collectively held over $830 million in TVL at their 2022 peak, though figures have contracted since.
The distinction in Buterin's proposal is conceptual rather than competitive. Existing protocols use options as a trading product. Buterin proposes using options as the foundational layer for index-tracking synthetics and stablecoins — replacing the role CDPs currently serve. This is a structural reclassification: options as infrastructure, not product.
Whether existing options protocols can adapt to serve this infrastructure role, or whether new purpose-built protocols like Cleave will dominate, remains an open question.
The proposal acknowledges several unsolved problems:
Rebalancing costs. Users must roll positions before expiry. Each roll incurs slippage and transaction costs. Buterin and forum participants identified this as the primary competitive risk. If rolling costs exceed 1-4% annually, the system becomes uneconomical relative to CDP-based alternatives that charge borrowing rates in a similar range.
Liquidity fragmentation. Each strike price and expiry date creates a separate market. Without sufficient liquidity at each combination, spreads widen and the system becomes impractical. This is the same challenge that has limited on-chain options adoption historically.
User complexity. CDP-based protocols offer a familiar mental model: deposit, borrow, monitor. Options-based positions require understanding of strike prices, expiry dates, and rolling mechanics. The user experience gap is substantial.
Formal verification requirements. Buterin specifically called for mathematical proof of correctness before mainnet deployment. Formal verification of DeFi protocols remains resource-intensive and has been completed for only a small fraction of deployed contracts.
Scaling questions. The system works cleanly for ETH-denominated synthetics. Extending it to multi-collateral environments, cross-chain implementations, or real-world asset indices introduces additional complexity that the current proposal does not address.
The proposal has potential relevance for institutional DeFi adoption, though this remains speculative at the current stage.
Forced liquidation risk is cited by institutional risk committees as a material barrier to DeFi protocol deployment. According to analysis from P2P.org, liquidation cascades represent "the primary mechanism through which DeFi contagion spreads during market stress events." An architecture that eliminates this vector would reduce systemic collateral concentration risk — a metric institutional allocators track.
The slow-oracle design also aligns with institutional compliance requirements. Extended dispute windows and human-reviewable settlement align more closely with traditional financial settlement processes than the instant, irreversible liquidation mechanics of current DeFi lending.
However, institutional adoption depends on mainnet deployment, formal verification, sustained liquidity, and regulatory clarity — none of which currently exist for this architecture.
Buterin's proposal represents a structural critique of DeFi's nine-year-old CDP architecture, not an incremental improvement. The claim is that liquidation is not an unfortunate side effect to be optimized, but a design choice that can be replaced entirely.
The data supports the urgency. $150 billion in 2025 liquidations, a $19 billion single-day cascade, and persistent oracle exploitation losses collectively demonstrate that the current system's failure modes are expensive and recurring. The October 2025 flash crash alone erased more value in liquidations than the entire DeFi lending market held in TVL three years prior.
Whether the options-based alternative can achieve the liquidity depth, user adoption, and formal security guarantees necessary for mainnet deployment remains undemonstrated. The 10-day sprint from paper to testnet is encouraging but proves only that the concept is implementable, not that it is economically viable at scale.
What is clear: the economic value currently extracted through liquidation penalties, MEV from liquidation transactions, and oracle fee revenue is substantial. A shift to options-based infrastructure would redistribute that value — reducing costs for borrowers while eliminating revenue streams for liquidation bots and oracle networks. The redistribution of economic rents, rather than the elimination of technical risk, may ultimately determine whether this architecture gains adoption or remains an elegant theoretical alternative.