Famous DeFi Hacks: The Biggest Exploits in DeFi History and What They Teach Us
In short: DeFi’s biggest hacks include The DAO in 2016, which led to Ethereum’s hard fork; cross-chain bridge thefts in 2022 such as Ronin (about US$625 million), Wormhole (about US$320 million) and Nomad (about US$190 million); and protocol exploits such as Euler Finance in 2023 (about US$197 million, mostly returned). Large exploits continued in 2025, including Cetus (about US$223 million) and Balancer (about US$128 million). The key points:
- Bridges, which hold large pools of assets, have been the most lucrative targets.
- Many exploits hit audited protocols; audits reduce risk but do not remove it.
- Key management and governance failures cause losses as well as code bugs.
- Recoveries happen, but are the exception rather than the rule.
Ethereum’s developer guide to smart contract security describes the vulnerability types behind many of these hacks, and the ATO’s guidance on loss or theft of crypto assets covers how losses may be treated for tax. Moneysmart’s overview of crypto assets covers the risks of holding crypto, and ASIC’s guidance on digital assets explains how crypto services are regulated in Australia.
The history of DeFi hacks is simultaneously a chronicle of the most expensive software bugs in human history and the most valuable security education library available for DeFi investors. The total losses from DeFi exploits, bridge hacks, and smart contract vulnerabilities since 2016 run to many billions of US dollars, with individual incidents ranging from relatively small protocol exploits to single events causing losses of more than US$600 million. Each major exploit has taught the DeFi security community something important: about the specific vulnerability class exploited, about the economic design failures that made the attack profitable, about the response mechanisms that protocols use to manage crises, and about the governance structures that determine whether affected users are compensated. For Australian DeFi investors who study these case studies, the accumulated lessons provide a practical framework for evaluating the security posture of DeFi protocols today, understanding which risks have been addressed by the ecosystem and which remain open.
The DAO Hack 2016: The Original DeFi Exploit
The DAO hack of June 2016 is the foundational case study in DeFi security history, occurring before “DeFi” was even a recognised term. The DAO (Decentralised Autonomous Organisation) was a venture capital fund built on Ethereum, which had raised the equivalent of about US$150 million in ETH (at 2016 prices) from thousands of participants, making it the largest crowdfunding campaign in history at the time. A smart contract reentrancy vulnerability in The DAO’s withdrawal function allowed an attacker to recursively drain ETH before the contract could update the victim’s balance: the attacker called the withdraw function, received ETH, and before the DAO contract could record the withdrawal, called the function again, repeating this cycle to drain approximately 3.6 million ETH (roughly US$50-60 million at the time, worth far more at later ETH prices).
The response to the DAO hack led to the most controversial event in Ethereum history: a hard fork that reversed the theft. The Ethereum community was faced with a binary choice: leave the blockchain immutable (with the DAO attacker keeping the stolen ETH, which was then locked in a child DAO by the smart contract’s terms for 28 days) or execute a hard fork that rewrote Ethereum’s transaction history to return the stolen funds to the original investors. The community (with developer and miner majority) voted for the hard fork, which was executed in July 2016, returning the stolen funds and effectively undoing the theft at the protocol level. A minority who rejected the fork continued the original chain as “Ethereum Classic” (ETC), which still exists today. The DAO hack therefore resulted in both the largest smart contract exploit in history (at the time) and the establishment of an important principle: even in the blockchain ecosystem, sufficiently large and motivated communities can intervene in apparently immutable history. This intervention remains controversial: it demonstrated that “code is law” has practical limits when the human cost of letting code run is high enough.
The reentrancy vulnerability that enabled the DAO hack became one of the most widely studied vulnerability classes in smart contract security. The pattern (external call before state update) is now checked by every DeFi smart contract audit, caught by automated analysis tools, and addressed by standard patterns (the “checks-effects-interactions” pattern, reentrancy guard modifiers) that are now standard practice in smart contract development. Modern DeFi protocols built by competent developers rarely exhibit simple reentrancy vulnerabilities, though novel variations continue to appear in less carefully developed protocols. The DAO hack’s most lasting contribution to DeFi security was not the exploit itself but the ecosystem-wide adoption of reentrancy protection patterns that followed.
How The DAO Shaped DeFi Governance
The DAO hack’s structural context is important for understanding subsequent DeFi governance design. The DAO had no upgrade mechanism, no pause function, and no emergency response capability: once the attack began, there was nothing the development team could do within the protocol to stop it. Subsequent DeFi protocol designs incorporated emergency pause mechanisms, upgrade timelocks, and governance structures with faster response paths for security emergencies, directly in response to the DAO vulnerability response crisis. For Australian DeFi investors evaluating current protocols, the presence of circuit breakers, pause functions, and responsive security governance (with appropriate timelock protections to prevent abuse) reflects the lessons learned from The DAO’s inability to respond to the attack in progress.
The tax implications of the DAO hack for Australian participants provide an early example of the ATO’s approach to DeFi losses. Participants who lost ETH in the DAO hack and subsequently received it back through the hard fork faced a complex tax situation: was the theft itself a capital loss event? Was the recovery income or a reversal of the loss? For current DeFi exploits (where no hard fork recovery occurs and funds are permanently lost), losses from exploits may be claimable as capital losses once there is no reasonable prospect of recovery, provided you have the evidence the ATO asks for. The deductibility of DeFi exploit losses and the timing of the capital loss event require careful consideration with a crypto tax specialist in the aftermath of any significant DeFi loss event.
Major Bridge Hacks: Ronin, Wormhole, and Nomad
Cross-chain bridge hacks have been responsible for some of the largest single theft events in DeFi history, with the Ronin Bridge hack (March 2022, about US$625 million stolen) representing the largest single DeFi exploit. The Ronin Network is an Ethereum sidechain built for the Axie Infinity game; the Ronin Bridge allowed assets to flow between Ethereum mainnet and the Ronin sidechain. The attack compromised 5 of the 9 validator nodes in the Ronin multisig (4 through a social engineering attack on Sky Mavis, the Axie Infinity developer, and 1 through a legacy validator node that should have been revoked months earlier but was not), allowing the attacker to approve withdrawals of 173,600 ETH and 25.5 million USDC from the bridge. The security failure was fundamentally a governance and operational failure (insufficient validator decentralisation, poor key management, lack of monitoring that missed the withdrawals for 6 days) rather than a smart contract code bug.
The Wormhole bridge hack (February 2022, about US$320 million) exploited a signature verification vulnerability in Wormhole’s smart contracts on Solana that allowed the attacker to mint 120,000 Wrapped Ethereum (wETH) on Solana without providing any ETH as backing collateral. The vulnerability was in the guardian signature verification: the attacker found a way to pass a spoofed guardian account that the contract mistakenly accepted as a valid verification, allowing unauthorised minting of bridged assets. Wormhole’s parent company Jump Crypto subsequently replenished the stolen ETH from its own capital to make affected users whole (a crisis response that cost hundreds of millions of dollars and demonstrated the financial commitment of institutional backers to protocol integrity). The Wormhole hack highlighted that bridge smart contracts, which often hold billions of dollars of assets, are extremely high-value targets that require extraordinary security rigour, and that the most critical infrastructure in DeFi warrants the most intensive security investment.
The Nomad Bridge hack (August 2022, about US$190 million) was distinctive for its chaotic crowd-sourced nature: rather than a single sophisticated attacker executing a planned exploit, a message routine vulnerability in Nomad’s contracts was discovered and initially exploited by one attacker, then immediately exploited by hundreds of opportunistic copycats who recognised the transaction pattern and submitted identical transactions to drain remaining bridge funds. The vulnerability (a routine upgrade that accidentally initialised a storage value in a way that caused the bridge to accept any message as valid) was so straightforward that non-technical opportunists could copy the exploit transaction with minimal modification, leading to the bridge being drained by a crowd of exploiters within hours. For Australian DeFi investors, the Nomad hack illustrates that even sophisticated bridge smart contract vulnerabilities can be discovered and replicated by unsophisticated actors once the initial exploit transaction is public on-chain.
Bridge Lessons for DeFi Users
The security lessons from bridge hacks are directly applicable to Australian DeFi investors who use cross-chain bridges: prefer native rollup bridges (controlled by the L2 protocol’s own contracts rather than a separate bridge operator) over third-party bridges for moving assets between Ethereum mainnet and Layer 2 networks; limit the amount transferred through any single bridge at once (bridge the minimum needed for the specific DeFi activity rather than all capital simultaneously); check the bridge’s audit status and TVL trend before using it (rapidly declining TVL may indicate market concern about the bridge’s security); and be aware that assets in transit on a bridge are at elevated risk during the bridge window. The total losses from bridge hacks (about US$2 billion in 2022 alone) reflect the specific risk concentration that bridges represent as intermediaries holding enormous sums between chains.
For Australian DeFi investors evaluating the tax treatment of bridge hack losses, the same principles apply as for other DeFi exploit losses: losses from bridge hacks that are clearly irrecoverable represent capital losses in the year the loss is confirmed. The ATO’s crypto data matching capabilities extend to identifying on-chain bridge interactions, so bridge activity should be included in ATO reporting regardless of whether the exchange or counterparty provides transaction records. Bridge transaction records are permanently on the public blockchain and accessible to the ATO’s blockchain analytics tools, making accurate voluntary DeFi tax disclosure both legally required and practically necessary to avoid underreporting penalties.
Protocol Exploits: Euler Finance, Cream Finance, and Compound
The Euler Finance hack (March 2023, about US$197 million) is notable both for its scale and its unusual resolution. Euler Finance, a DeFi lending protocol on Ethereum, was exploited through a combination of flash loan and a vulnerability in the donate() function that allowed an attacker to create a large bad debt position while simultaneously holding a large donation claim against the protocol, and use the resulting imbalance to drain approximately US$197 million in DAI, WBTC, stETH, and USDC from the lending pools. The unusual part of the Euler story was the aftermath: the attacker, after initially appearing to launder the stolen funds, entered negotiations with the Euler team and ultimately returned approximately 90% of the stolen funds in return for not being prosecuted. The near-complete recovery (exceptional in DeFi exploit history) resulted from both on-chain communication between the attacker and Euler team and the attacker’s apparent inability to launder the funds anonymously enough to feel secure keeping them.
Cream Finance was exploited multiple times (most notably in October 2021 for about US$130 million and earlier in August 2021 for about US$19 million), making it a case study in persistent security failures at a protocol that had been audited and had responded to earlier exploits but continued to have exploitable vulnerabilities. The largest Cream exploit used a combination of flash loans and price oracle manipulation across multiple tokens and multiple protocols in a complex attack that required deep knowledge of Cream’s specific implementation. The repeated exploitation of Cream despite its audit status illustrates the limitation that audits cannot guarantee security against sophisticated adversarial analysis, particularly for complex protocols with many integration dependencies. For Australian DeFi investors, Cream’s history demonstrates that a protocol’s past exploit and survival does not guarantee future security: ongoing vigilance and DeFi risk management is required regardless of the protocol’s history.
The Compound governance vulnerability of 2021 (where a routine parameter update governance proposal accidentally allocated an excessive amount of COMP tokens to users, with estimates of COMP 280,000 sent incorrectly) was unusual in being a governance process failure rather than a smart contract code exploit. The Compound team discovered the error but were unable to pause or reverse the distributions because Compound’s governance has a 7-day proposal review period before implementation, and there was no emergency pause mechanism for token distributions. The loss was recovered partially through voluntary token returns by users who received excess COMP and chose to return it (demonstrating the community aspect of DeFi protocol recovery), but the incident highlighted that governance process errors can be as costly as smart contract bugs. The Compound incident led to industry-wide discussion of governance timelock designs and emergency response capabilities for DeFi protocols.
Flash loan attacks as an exploit category deserve specific attention in the context of DeFi hacks, as they represent an attack vector unique to DeFi with no traditional finance equivalent. A flash loan is an uncollateralised loan that must be borrowed and repaid within a single atomic transaction; if the repayment fails for any reason, the entire transaction is reverted and the lender’s capital is unaffected. Flash loans provide any user with access to enormous amounts of capital (hundreds of millions of USD worth) for the duration of a single transaction block, enabling attacks that would require prohibitive capital in traditional finance. The most impactful DeFi exploits of 2020-2023 used flash loans to: manipulate on-chain price oracles by temporarily imbalancing liquidity pools; create synthetic arbitrage conditions that exploit protocol assumptions about capital availability; and simultaneously occupy multiple positions across interconnected protocols to create situations the individual protocol designers did not anticipate. Defending against flash loan attacks requires designing protocols that assume adversarial actors have unlimited capital within a transaction block, which is a fundamentally different threat model than traditional finance security engineering.
The recovery track record across major DeFi hacks is relevant for Australian DeFi investors evaluating how protocols handle crises. Some recoveries: Euler Finance (2023) recovered ~90% through negotiations. Poly Network (2021) saw the attacker voluntarily return about US$611 million stolen in one of the largest crypto thefts ever, reportedly because the attacker found it impossible to launder funds of that scale and was moved by the community response. Wintermute (2022) negotiated a partial recovery. Most major DeFi hacks result in no recovery at all, and user compensation (when it occurs) comes from protocol treasuries, emergency fundraising, or investors. For protocols that have been exploited and subsequently compensated users (like the Euler situation), the subsequent trust recovery is often stronger because the community sees the protocol’s commitment to making users whole. For Australian DeFi investors tracking DeFi security developments, the combination of the DeFi exploit history and the post-exploit recovery track record for specific protocols provides important context for portfolio risk assessment.
Applying DeFi Hack Lessons to Your Investment Decisions
The recurring patterns across major DeFi hacks provide actionable guidance for Australian DeFi investors evaluating protocol security. Smart contract reentrancy (DAO-style attacks) is now largely addressed in modern code through standard guards, so its continued presence in an audit report finding would be a serious red flag for a 2024+ protocol. Bridge security failures consistently involve insufficient validator decentralisation, rushed code deployment, or governance failures rather than exotic cryptographic attacks: checking bridge validator counts, multisig thresholds, and audit currency before using any bridge is a practical risk reduction step. Flash loan attack vulnerability depends heavily on oracle design: protocols using spot price oracles for critical decisions (collateral valuation, liquidation thresholds) are substantially more vulnerable than those using time-weighted average prices or robust multi-source oracle systems.
Position sizing discipline in light of DeFi hack history means applying explicit allocation limits to any single DeFi protocol based on its security maturity. A rough framework based on DeFi hack pattern analysis: protocols with 3+ years of operation, multiple audits, no major exploits, and high TVL are relatively mature (though not safe); in one illustrative scheme, such protocols might hold up to 5% of a crypto portfolio each. Protocols with 1-3 years of operation, 2+ audits, and strong TVL trends are intermediate risk, with perhaps up to 2%. Newer protocols, those with recent minor exploits, or those with sparse audit coverage represent higher risk, with smaller amounts still. The exact limits are a personal decision. The position sizing discipline that applies these protocol risk categories to actual allocation limits directly applies DeFi hack history lessons to portfolio construction, reducing the impact of any single exploit on the overall portfolio while still allowing meaningful participation in the DeFi ecosystem.
Protocol selection based on security track record is a positive-screening approach that systematically favours protocols that have demonstrated resilience through market cycles and security challenges. Protocols like Uniswap (no major exploit in 5+ years of operation), Aave (minor incident responses that demonstrated good governance without major losses), and Compound (despite the governance incident, no fund loss to external hackers) have track records that newer protocols cannot claim. For Australian DeFi investors who want to participate in DeFi yield with lower tail risk, concentrating the majority of DeFi capital in battle-tested protocols with proven security track records (accepting lower yields in exchange for lower exploit risk) and limiting allocation to newer or less-tested protocols reduces the expected loss from exploits over a multi-year investment horizon.
Monitoring active exploit reports and DeFi security alerts is an ongoing requirement for active DeFi participation. DeFi exploit events are announced on Twitter/X within minutes of occurrence, with blockchain analytics firms (PeckShield, BlockSec, Certik) typically among the first to identify and publicise attacks in progress. Following these accounts (and the official channels of any protocol you are invested in) allows rapid response when an exploit occurs: the ability to exit a position within minutes of an exploit alert can be the difference between losing a small amount and losing everything in a protocol that is actively being drained. Setting up protocol-specific monitoring and maintaining the discipline to monitor alerts during periods of DeFi activity is a practical risk management practice that
The ultimate lesson from the history of DeFi hacks is that the DeFi ecosystem is still in an early and imperfect stage, where technical and economic design failures are common and costly, but where each failure has driven meaningful improvement in security standards, tooling, and protocol design. The trajectory from The DAO’s unguarded reentrancy (2016) to Euler’s complex multi-function flash loan attack (2023) shows both that attack sophistication has increased and that the low-hanging vulnerabilities of early DeFi have been systematically addressed. For Australian DeFi investors with a multi-year investment horizon, understanding this trajectory of improvement alongside the remaining open security challenges is essential context for informed participation. The DeFi ecosystem rewards participants who understand its risks and manage them actively, and
Frequently Asked Questions
What are the most significant DeFi hacks?
The history of DeFi hacks is simultaneously a chronicle of the most expensive software bugs in human history and the most valuable security education library available for DeFi investors. The total losses from DeFi exploits, bridge hacks, and smart contract vulnerabilities since 2016 exceed AUD 10 billion, with individual incidents ranging from relatively small protocol exploits to single events causing losses of AUD 1-2 billion. Each major exploit has taught the DeFi security community something important: about the specific vulnerability class exploited, about the economic design failures that made the attack profitable, about the response mechanisms that protocols use to manage crises, and about the governance structures that determine whether affected users are compensated.
What happened in The DAO hack of 2016?
The DAO was an Ethereum-based investment fund that raised about US$150 million worth of ETH. In June 2016 an attacker used a reentrancy bug to keep withdrawing before balances updated, draining about 3.6 million ETH, then worth roughly US$50-60 million. Ethereum's community chose a hard fork in July 2016 to return the funds, and those who rejected it continued as Ethereum Classic.
How did The DAO hack shape DeFi governance?
The DAO hack's structural context is important for understanding subsequent DeFi governance design. The DAO had no upgrade mechanism, no pause function, and no emergency response capability: once the attack began, there was nothing the development team could do within the protocol to stop it. Subsequent DeFi protocol designs incorporated emergency pause mechanisms, upgrade timelocks, and governance structures with faster response paths for security emergencies, directly in response to the DAO vulnerability response crisis.
What happened in the Ronin, Wormhole and Nomad bridge hacks?
In 2022 three bridges were drained. Ronin lost about US$625 million in March after attackers took control of five of its nine validator keys. Wormhole lost about US$320 million in February when a signature-checking flaw let an attacker mint 120,000 wrapped ETH on Solana. Nomad lost about US$190 million in August after a bug let hundreds of copycats drain it.
What do bridge hacks teach DeFi users?
Bridges hold large pools of assets, which makes them prime targets. Native rollup bridges, secured by the networks they connect, are generally lower risk than third-party bridges run by a separate set of signers. Limiting how much you move or leave in any one bridge, checking how many independent parties control it, and watching for incidents all reduce exposure.
What happened at Euler Finance, Cream Finance and Compound?
Euler Finance lost about US$197 million in a March 2023 flash loan exploit, but the attacker later returned almost all of it. Cream Finance was exploited several times in 2021, including about US$130 million in October. Compound's 2021 incident was a faulty upgrade that wrongly paid out around 280,000 COMP, which governance delays prevented the team from quickly fixing.
How should hack history shape your investment decisions?
The recurring patterns across major DeFi hacks provide actionable guidance for Australian DeFi investors evaluating protocol security. Smart contract reentrancy (DAO-style attacks) is now largely addressed in modern code through standard guards, so its continued presence in an audit report finding would be a serious red flag for a 2024+ protocol. Bridge security failures consistently involve insufficient validator decentralisation, rushed code deployment, or governance failures rather than exotic cryptographic attacks: checking bridge validator counts, multisig thresholds, and audit currency before using any bridge is a practical risk reduction step.
What are the risks associated with Famous DeFi Hacks?
The recurring pattern across major DeFi exploits is that audits, size and operating history have all failed to prevent losses, with bridges representing the largest single category. Recovery is rare and usually depends on the attacker choosing to return funds. For Australian investors the practical conclusion is to size positions on the assumption that any individual protocol can go to zero, prefer native rollup bridges over third-party ones, and avoid leaving assets in a protocol longer than the strategy requires.