Decentralized finance lost $45 billion in total value locked during the first half of 2026, dropping 39% from $114.5 billion to $71.8 billion as of June 18, according to DeFiLlama data. Liquidation cascades — automated forced selling triggered when collateral values breach preset thresholds — amp...
"What if we use options as the base of DeFi, instead of CDPs and liquidations? Instead of extreme price movements creating a sharp and global 'you get liquidated' effect, instead your position smoothly rebalances over time." — Vitalik Buterin, Ethereum Co-Founder
Decentralized finance lost $45 billion in total value locked during the first half of 2026, dropping 39% from $114.5 billion to $71.8 billion as of June 18, according to DeFiLlama data. Liquidation cascades — automated forced selling triggered when collateral values breach preset thresholds — amplified each downturn. On June 2 alone, $1.8 billion in leveraged positions were liquidated across 272,000 traders. In March, a misconfigured oracle on Aave's CAPO system incorrectly priced wstETH, triggering $27 million in erroneous liquidations across 34 accounts without any external attack.
On June 1, Ethereum co-founder Vitalik Buterin published a research post titled "Building index-tracking assets on top of options instead of debt" on the ETHResearch forum. The proposal argues that DeFi's foundational primitive — collateralized debt positions (CDPs) — should be replaced with options-based constructions that eliminate forced liquidation entirely. Within 10 days, at least three independent teams shipped working prototypes, including one physically settled implementation on Base and a testnet exchange called Cleave. The speed of development from proposal to code is unusual, but the gap between testnet prototype and production-grade infrastructure remains wide.
DeFi lending protocols — primarily Aave ($12.1 billion TVL), Morpho ($10.6 billion total deposits), and legacy MakerDAO infrastructure — operate on collateralized debt positions. Users deposit crypto assets, borrow against them, and face liquidation if collateral ratios fall below protocol-defined thresholds. The mechanism is functional under normal conditions. Under stress, it produces cascading failures.
The data from 2026 quantifies the cost:
The liquidation mechanism operates as intended — it protects protocol solvency by force-selling undercollateralized positions. The systemic problem is that mass liquidation events depress asset prices further, triggering additional liquidations in a self-reinforcing cycle. Each cascade compounds the next. DeFi TVL has declined every month in 2026.
Historical precedent stretches back to MakerDAO's "Black Thursday" in March 2020, when cascading liquidations under extreme market conditions exposed fundamental design flaws in CDP-based systems. Six years later, the same structural vulnerability persists across a larger capital base.
Buterin's June 1 research post proposes replacing CDPs with a paired-option construction. One unit of ETH is split into two complementary claims, labeled P and N, that always sum back to one ETH. P represents the upside exposure above a given strike price. N represents the downside or cash-equivalent portion below that strike. Because the two payoffs are mathematically complementary, "there is no possibility of liquidation," according to Buterin's post.
The construction is conceptually closer to a covered call strategy than to a collateralized loan. A user who wants dollar-denominated stability holds the N claim. A user who wants leveraged upside exposure holds the P claim. Neither position can be liquidated because there is no debt — only option exposure that expires or rebalances at maturity.
Buterin frames this within a broader thesis he first articulated in a September 2025 blog post on "low-risk DeFi," positioning the options-based model as essential infrastructure for Ethereum's long-term economic sustainability. The June proposal extends that thesis into a concrete architectural specification.
The proposal also extends beyond single-asset exposure. Buterin suggests the framework could enable "personalized value baskets" — index-tracking assets tied to CPI-based measures or multi-asset indices rather than single fiat currencies. This would allow users to construct censorship-resistant alternatives to stablecoins without inheriting the liquidation mechanics of CDP-based stablecoins like DAI.
The P/N construction operates as follows:
Minting: A user deposits 1 ETH into a vault smart contract. The contract mints one P token and one N token. Combined, they represent a claim on exactly 1 ETH.
Payoff Structure: At a given strike price (e.g., $2,000), the N token is worth the lesser of the ETH price or $2,000. The P token captures any value above $2,000. If ETH trades at $2,500, N pays $2,000 and P pays $500. If ETH trades at $1,500, N pays $1,500 and P pays $0.
Settlement: Positions settle at maturity — a defined expiration date — rather than being continuously monitored against collateral ratios. This eliminates the need for real-time liquidation triggers.
Redemption: At any time, a holder of both one P and one N token can redeem them for 1 ETH. This arbitrage mechanism keeps the combined value of P + N anchored to the underlying asset.
The critical difference from CDPs: no collateral ratio is monitored. No liquidation bot network is required. No cascading sell pressure is generated by price declines. The downside for the P holder is that their token goes to zero if ETH falls below the strike — but this loss is absorbed smoothly, not amplified through forced selling.
Buterin's proposal carries a secondary structural implication: it reduces dependency on real-time oracle infrastructure. Current CDP-based protocols require continuous price feeds updated within seconds to trigger timely liquidations. According to Buterin, "real-time oracles can only rely on a small number of automated actors watching live price feeds, and they leave no room for dispute resolution or slow verification."
The options-based model only requires price data at settlement — once per maturity cycle. This opens the design space for "slower oracles" that Buterin describes as "options-based, slow-oracle-friendly DeFi." Prediction-market-style oracles with dispute resolution windows could replace Chainlink-style real-time feeds for these instruments.
The Aave CAPO incident in March 2026 underscores the risk profile of real-time oracle dependency. The oracle reported a capped wstETH exchange rate of approximately 1.1939, well below the actual market rate of 1.228. The error was not an external attack — it was a misconfiguration in the protocol's own update pipeline. The safety constraint designed to prevent manipulation became the mechanism that caused incorrect pricing. Under an options-based settlement model, the brief pricing error would have had no effect because no real-time liquidation trigger would have existed.
Within 10 days of Buterin's post, multiple teams shipped working code.
Cleave launched as a testnet options exchange, positioning itself as "DeFi's missing third pillar" alongside Uniswap (spot) and Hyperliquid (perpetuals). Cleave splits ETH into an "upside half" and a "cash half" through one-click Earn and Boost products. Settlement relies on a median of three Uniswap time-weighted average prices (TWAPs), removing single-source oracle dependency. The protocol operates with no margin, no funding fees, and no liquidation mechanics. As of late June, it remains on testnet.
A Base prototype implemented the full P/N lifecycle using physically settled contracts: minting paired claims from WETH, transferring them, exercising against a USDC strike price, and settling the vault. The implementation settles by moving the underlying assets rather than reading a price feed, removing the oracle from the critical path entirely. Transaction hashes and contract addresses were published in the ETHResearch forum thread. A perpetual liquidity-backed variant was also noted as live.
On June 11, Buterin acknowledged the development pace, posting "Looks like the options thing is happening already!" He paired the endorsement with a warning: "I do strongly urge that if any of these get on mainnet quickly, we formally verify it first," tagging the Vyper team and formal verification developers.
The proposal has drawn substantive criticism.
Rebalancing slippage is the most cited practical concern. Options-based positions tied to indices require periodic rebalancing, and each rebalance executes a trade. On Ethereum mainnet, gas costs and slippage could erode returns, particularly for smaller positions. This is the inverse of the liquidation problem: instead of catastrophic one-time losses, users face continuous friction costs.
Loss of core lending utility: Ben Lilly of Brownstone Research argues the options structure eliminates a primary use case — borrowing against appreciating collateral. Current CDP users can "borrow against appreciating collateral while holding the underlying asset." The options model requires users to give up either upside (by holding N) or principal protection (by holding P). "Most borrowers simply want a button that gives them dollars and another button that gives them their collateral back," Lilly wrote, "not management of option strikes, expirations, and rolls."
Maturity rollover costs: Unlike CDP-based loans with indefinite terms, options expire. Borrowers must "roll into a fresh contract every cycle, which means slippage and no set-it-and-forget-it mode," according to Lilly. This introduces recurring costs and complexity that CDP-based lending avoids.
User experience gap: The mechanism requires familiarity with options terminology — strikes, expirations, payoff curves — that most DeFi users do not currently possess. Whether abstraction layers can sufficiently simplify the interaction remains unproven.
Scale and liquidity: No options-based DeFi product has yet demonstrated sufficient liquidity depth to handle the volumes currently managed by Aave ($12.1 billion TVL) or Morpho ($10.6 billion deposits). Moving from testnet to production-grade infrastructure capable of absorbing billions in deposits represents a qualitatively different engineering and economic challenge.
The proposal arrives at a structurally significant moment. DeFi TVL sits at $71.8 billion, down from $114.5 billion at the start of the year. Lending protocols hold approximately $36.5 billion in aggregate TVL — roughly half the sector's total. The lending category is both the largest capital pool in DeFi and the most exposed to liquidation cascade risk.
Simultaneously, institutional capital is entering DeFi lending infrastructure. Morpho closed a $175 million round in June, co-led by Paradigm, a16z Crypto, and Ribbit Capital, with participation from Apollo Funds, Circle Ventures, and VanEck. The round valued Morpho at $2 billion. Morpho's public metrics show $10.6 billion in total deposits and $3.7 billion in active loans as of June 22.
Aave, the largest lending protocol by TVL, launched V4 in early 2026, which crossed $200 million in deposits and $56 million in active loans within three months. Aave founder Stani Kulechov has stated that the protocol sees growing adoption from traditional finance players.
The institutional interest in DeFi lending — and the simultaneous structural fragility demonstrated by the CAPO oracle incident, the KelpDAO contagion, and recurring liquidation cascades — creates a tension that Buterin's proposal directly addresses. Whether the options-based architecture can scale to meet institutional requirements, or whether it remains a research curiosity, will likely be determined in the second half of 2026 as testnet implementations approach mainnet readiness.
Buterin's options-based DeFi proposal addresses a real structural vulnerability — one that has cost billions in forced liquidations, eroded TVL, and created systemic contagion pathways across the sector in 2026. The mathematical elegance of the P/N construction is sound: complementary option payoffs that always sum to the underlying asset value eliminate the liquidation trigger by design.
The gap between mathematical soundness and production infrastructure remains large. No implementation has been formally verified. No options-based DeFi product has demonstrated liquidity sufficient to absorb even 1% of current lending TVL. The user experience challenge of abstracting options mechanics into accessible interfaces is nontrivial. Rebalancing costs at scale are modeled but not empirically tested.
What makes the proposal significant is not any single implementation but the architectural framing: that DeFi's most economically destructive failure mode — cascading liquidations — may be a design choice rather than an inevitability. Whether the options-based alternative can scale from testnet prototype to production reality is an open question. That the question is now being asked, with code rather than whitepapers, marks a meaningful development in DeFi infrastructure design.