Cross-Protocol Flash Loan Exploit Drains $48M From Lending Protocol
A flash loan attack exploiting a reentrancy vulnerability in a lending protocol's donation mechanism extracted $48M in under 60 seconds — a post-mortem analysis.
A DeFi lending protocol was exploited for $48M on May 24, 2026, in an attack that combined a flash loan, a donation mechanism edge case, and a reentrancy vulnerability in the liquidation callback. The attacker repaid the flash loan and exited with $48M profit in a single transaction.
On-Chain Context
The attack vector centred on the protocol's "donate to reserve" function, which allowed users to increase a pool's reserves without minting shares. This created a discrepancy between the pool's share price and its actual asset value, which the attacker exploited via a sequence of leveraged positions and targeted liquidations.
Risk & Opportunity Assessment
Three root cause lessons: First, donation functions require the same security scrutiny as deposit functions. Second, flash loan access should be restricted during protocol state changes. Third, share price manipulation via direct asset transfers is a class of vulnerability that auditors should explicitly test for in all lending protocols.
"This development underscores the maturation of DeFi infrastructure — protocols are increasingly competing on execution quality rather than raw liquidity depth."
The broader market context remains constructive. Total value locked across DeFi stands at $148.2B, up 12.4% month-over-month, driven primarily by renewed institutional participation in structured yield products.
Comparative Protocol Analysis
When benchmarked against competitors, the divergence in execution strategies becomes clear. While some protocols have prioritised simplicity and gas efficiency, others are betting on composability and hook-based extensibility as the primary moat.
For DeFi participants, the actionable takeaway is to monitor on-chain flow data over the next 72 hours. Capital allocation shifts of this magnitude typically produce follow-on effects across correlated pools within three to five blocks of the initial transaction.
AI · Based on Rekt News
Defiliban Research
Senior Analyst