Vitalik Buterin posted a research proposal on June 1 that asked one question: what if options-based DeFi replaced debt and liquidations entirely? The response from builders was not theoretical.

Days after the post went live on the Ethereum Research forum, at least one developer had already deployed a working implementation on Base and completed a full cross-party settlement cycle on-chain.

The Idea Behind Dropping Debt

The proposal centers on two ERC-20 tokens, P and N, minted from one ETH with a strike price and a maturity date. Before maturity, combining P and N recombines back into one ETH. At maturity, an oracle resolves the index price, and each token receives its ETH share based on where the price landed relative to the strike.

No forced liquidations. No real-time oracle dependency during the life of the position. Settlement runs as a pure asset transfer, not a price lookup. That last part is where the design breaks from how most DeFi works today.

As Vitalik Buterin posted on X:

“Looks like the options thing is happening already! Though I do strongly urge that if any of these get on mainnet quickly, we formally verify it first.”

He named Vyper and Fricoben of LFG Labs specifically as teams who could help with that formal verification step. The urgency in his wording is worth noting.

Settlement Without a Price Feed

Researcher mmchougule posted in the Ethereum Research thread confirming a live implementation on Base using a WETH/USDC pair. The vault address is public. The exercise transaction hash is on-chain.

Four parties ran the lifecycle. Party A minted P and N from WETH and sent N to Party B. Party B exercised N, paying USDC and receiving WETH in return. Party C settled the vault. Party A then redeemed P and collected the USDC that had been paid in.

The key step, per mmchougule’s account in the thread:

“N was burned, USDC moved into the vault, and WETH moved out to the N holder atomically.”

Settlement asked nothing about the ETH/USD rate. It just moved whatever assets the vault held. That is a clean break from how liquidation-based systems work, where a live price feed triggers automated decisions in real time.

The tradeoff mmchougule flagged is direct: removing the settlement oracle pushes responsibility onto whoever holds the N token. Price discovery moves off the protocol and onto the exerciser’s incentive. A keeper was prototyped to auto-exercise in-the-money positions, but that stays in execution infrastructure, not protocol guarantees.

Three Open Questions Nobody Has Answered

mmchougule closed the reply with three unresolved issues: how P/N secondary liquidity should form, how rolling and rebalancing can happen with minimal slippage, and whether exercise can be made reliable at protocol level without pulling oracle dependency back in.

None of those have answers yet. Still, Buterin’s own response to the thread treated it as progress, not a prototype to shelve.

Czar102, posting on X, acknowledged the attention and encouraged others to read the related essay. As Czar102 posted on X:

“Thanks for mentioning my essay! Hope it was interesting, recommend a read to anyone who hasn’t seen it.”

A separate testnet implementation referenced in the thread via Firefly is running at testnet.cleave.market. That puts at least two live builds running in parallel less than a week after the original forum post.

The rebalancing cost problem is still open. An ETH holder running a dollar-tracking position would face gradual exposure drift as prices move toward the strike, not an instant wipeout. Different risk profile. Not necessarily a safer one for anyone who misses a rebalancing window. Buterin described this drift as smoother than liquidation, but smoothness and safety are not the same thing.

Formal verification is the condition Buterin attached to any of this reaching mainnet. Whether Vyper or LFG Labs picks that up first will likely set the pace for what ships and when.