Summary: B.Protocol is a liquidation-rights redesign. The useful part is simple: it decides who gets first crack at liquidations, where standby capital sits, and whether the surplus goes to keepers or backstoppers.
What it does:
Wraps or integrates with lending markets so liquidation handling can be rerouted through a dedicated backstop layer rather than left entirely to open keeper races
In V1, lets backstop liquidity providers obtain liquidation priority and share proceeds with borrowers and lenders through a profit-sharing mechanism
In V2/B.AMM, pools user funds into backstop pools that can repay bad debt during liquidations and then rebalance seized collateral back into the deposited asset
Parks idle backstop capital in yield-bearing protocols until liquidation demand appears, rather than leaving all liquidation capital inert
Uses an external price feed plus a Curve-inspired invariant adaptation to sell seized collateral with imbalance-aware discounts instead of relying only on open-market DEX execution
Positions itself as infrastructure for lending markets, synthetic assets, stablecoin systems, and long-tail/low-liquidity collateral markets where liquidation depth is the bottleneck
Key claims:
The official docs frame B.Protocol as a user-based backstop liquidity protocol whose purpose is to make lending platforms more capital efficient and stable by handling liquidations at scale with the Backstop Automated Market Maker, which is the clearest reason to catalog it as liquidation infrastructure rather than as another yield app.
The V1 materials are still analytically useful because they expose a distinct earlier control surface: auctioned liquidation franchises, user-rating-driven jar distributions, and explicit MEV/gas-war recapture from miners back to platform users.
The V2/B.AMM materials matter because they shift the question from who wins the liquidation transaction to where standby liquidation capital sits and how seized collateral is unwound after the debt is repaid.
The B.AMM summary explicitly says liquidations of large debt can be handled with smaller capital requirements because the system rebalances collateral back to the original asset instead of requiring the backstop to warehouse whatever collateral it seized indefinitely.
The docs also say idle user deposits are expected to sit in yield-bearing protocols until needed for liquidation, which means B.Protocol is simultaneously a liquidation-control layer and a capital-routing layer.
The rebalance path is not generic AMM swapping. Official docs say the system uses an oracle-guided market price plus an imbalance discount adapted from the Curve stable-swap invariant, which makes post-liquidation inventory management the core reusable mechanism.
B.Protocol is especially useful in the corpus because it exposes a clean historical progression inside one project: first replace keeper gas wars with liquidation-priority auctions, then replace specialist liquidator balance sheets with user-funded backstop pools plus automated collateral unwinds.
The main comparison question is not whether B.Protocol is better liquidations, but whether liquidation rights should be allocated by open mempool races, franchise-like priority auctions, pooled backstop capital, liquidation-range AMMs, or some hybrid of those layers.
Whitepaper: A B.AMM whitepaper is linked from the official V2 docs via IPFS and GitHub, but the strongest readable primary materials in this pass were the official general, V1, and V2 documentation pages plus the whitepaper repository README; see ../whitepapers/b-protocol-primary-sources-2026-05-14.md.