Ethereum layer-2 network Taiko halted block production on June 22, 2026, after an attacker exploited a leaked SGX signing key to forge bridge withdrawal proofs and drain approximately $1.7 million from its L1 Bridge and ERC20Vault contracts. The TAIKO token fell over 20% to $0.07, near its all-ti...
"An RSA private key committed to a public GitHub repo just became a $1.7M exploit." — Quill Audits, Blockchain Security Firm
Ethereum layer-2 network Taiko halted block production on June 22, 2026, after an attacker exploited a leaked SGX signing key to forge bridge withdrawal proofs and drain approximately $1.7 million from its L1 Bridge and ERC20Vault contracts. The TAIKO token fell over 20% to $0.07, near its all-time low, on a market capitalization of $14.5 million. The team contained the exploit within hours, paused affected contracts via its Security Council, and urged all users to withdraw funds from every bridge deployed on the network.
The incident is the 15th major cross-chain bridge exploit of 2026. Cumulative bridge losses now exceed $342 million across the year, according to PeckShield data. Bridges account for approximately 42% of all DeFi exploit losses in 2026, despite representing a fraction of total protocol count. The root cause — a single cryptographic key stored in a public GitHub repository — underscores the persistent gap between bridge security architecture and the value these systems are asked to secure.
At approximately 2:00 a.m. ET on June 22, 2026, blockchain security firm Blockaid detected anomalous withdrawal activity on Taiko's Ethereum L1 bridge contract (0xd60247c6848B7Ca29eDdF63AA924E53dB6Ddd8EC). An attacker had submitted forged cross-chain proofs that convinced the bridge to release funds against withdrawal requests with no corresponding deposits on the Taiko L2 chain.
The attacker drained an estimated $1.7 million from the L1 Bridge and ERC20Vault (0x996282cA11E5DEb6B5D122CC3B9A1FcAAD4415Ab) before the Taiko team froze the contracts. Approximately 2 million TAIKO tokens, worth roughly $170,000, were moved to the MEXC exchange prior to the freeze.
Taiko's bridge TVL stood at approximately $20.7 million at the time of the incident, according to DeFiLlama data, meaning the attacker captured roughly 8.2% of total bridge value before containment.
The exploit originated from an operational security failure, not a smart contract logic bug.
The leaked key. Taiko's proof-generation system, Raiko, uses Intel SGX (Software Guard Extensions) enclaves to produce cryptographic attestations that verify L2 state transitions. These proofs anchor Taiko's "based rollup" architecture, where the Ethereum L1 serves as the sequencer. The system depends on an RSA-3072 private key (enclave-key.pem) remaining sealed inside SGX hardware so that only authorized provers can generate valid proofs.
BlockSec's Phalcon analysis identified that this RSA-3072 private key had been publicly committed to Taiko's open-source GitHub repository (taikoxyz/raiko). The key was accessible to anyone who cloned the repository.
The attack chain. With the exposed key, the attacker executed a two-phase operation:
Prover registration. The attacker used the leaked key to register attacker-controlled SGX instances via Taiko's SgxVerifier.registerInstance function, presenting themselves as legitimate provers within Taiko's verification framework.
Proof forgery and fund extraction. The attacker generated fraudulent L2 state attestations that passed on-chain verification checks. These forged proofs triggered processMessage() calls on the L1 bridge, setting withdrawal statuses to RETRIABLE. The attacker then called retryMessage(), which executed with minimal additional validation, releasing ETH and ERC-20 tokens from the bridge and token vault on Ethereum mainnet.
The fundamental issue: the bridge verified that a proof came from a registered SGX instance, but the registration mechanism itself was compromised. BlockSec Phalcon noted that "because the enclave signing key was publicly accessible, the SGX prover trust model may have been broken."
Taiko's Security Council acted within hours. The team issued a statement confirming "a compromise of Taiko's chain state verification mechanism," adding that "the security assumptions of all bridges deployed on Taiko can no longer be relied upon."
Response actions included:
The team stated it is coordinating with "ecosystem partners to contain the incident, pause affected systems where possible, and take all necessary technical and legal actions." Taiko said remaining bridge funds are secure and a full incident report would follow.
The TAIKO token dropped over 20% following disclosure, trading near $0.07 on CoinGecko — close to its all-time low — on a total market capitalization of $14.5 million.
The Taiko incident extends what has become the most damaging year for cross-chain bridge security since the 2022 cluster of Ronin ($624M), Wormhole ($320M), and Nomad ($190M) exploits.
| Date | Protocol | Amount Lost | Root Cause | |------|----------|-------------|------------| | Jan 2026 | IoTeX ioTube | $4.4M | Private key compromise | | Feb 2026 | Gravity Bridge | $5.4M | Validator key breach | | Mar 2026 | Purrlend Bridge | $1.5M | Unauthorized multisig transaction | | Apr 2026 | Hyperbridge | $2.5M | Mint function vulnerability | | Apr 2026 | Axelar | $4.7M | Cross-chain messaging flaw | | Apr 2026 | Kelp DAO (LayerZero) | $292M | RPC quorum set to 1-of-1 | | May 2026 | Verus-Ethereum | $11.6M | Source-side balance verification bypass | | May 2026 | Aztec Connect | $2.1M | Proof verification failure | | Jun 2026 | Taiko Bridge | $1.7M | Leaked SGX signing key |
Select incidents. PeckShield tracked 14 major bridge exploits through May 2026 totaling $328.6M. With June incidents, cumulative losses exceed $342M.
Kelp DAO accounts for 85% of the year's bridge losses by dollar value. The attacker exploited a default RPC quorum configuration set to 1-of-1 in Kelp's LayerZero-based bridge, meaning a single node could authorize fraudulent messages. The attacker drained 116,500 rsETH tokens (approximately $292 million) backed across more than 20 chains.
PeckShield data shows 60 security incidents in May 2026 alone — the highest monthly count of the year — with a 13.7% fund recovery rate, meaning approximately 86% of stolen funds remain unrecovered.
Cross-chain bridges have accumulated more than $2.8 billion in cumulative losses since 2022, representing roughly 40% of all value hacked in Web3, according to security research compiled by Yellow Network. The pattern is structural, not incidental.
The trust problem. A bridge must answer one question: did something actually happen on another chain? Destination chains cannot natively verify source chain state. This forces bridges to rely on external trust mechanisms — validator sets, oracle networks, SGX enclaves, or optimistic verification windows — each of which introduces a potential single point of failure.
Vulnerability concentration. Security firm Halborn estimates that insufficient validation of cross-chain message provenance accounts for approximately 40% of all bridge exploits since 2022. The Taiko exploit falls into this category: the bridge verified proof format but not proof origin integrity.
Key management failures. Private key compromises accounted for 88% of stolen funds in bridge exploits during Q1 2025, per Chainalysis data, and the pattern has continued into 2026. The Humanity Protocol breach in June 2026 ($36M) traced to three of six Ethereum keys and three of five BNB Chain keys stored on a single employee laptop. In Taiko's case, the key was stored in a public repository.
Economic incentive mismatch. Bridge validator fee revenue is typically small relative to the total value locked. This creates an asymmetry: the cost of attacking a bridge can fall below the value of assets held in it. IC3 researchers have noted that "validator-set bridges become economically insecure whenever the cost of corrupting validators falls below the value of assets."
The $342 million in 2026 bridge losses represents a direct destruction of economic value that flows downstream through the ecosystem.
Token holders absorb immediate market impact. TAIKO's 20% decline wiped approximately $3.6 million in market capitalization — more than double the $1.7 million directly stolen. Kelp DAO's rsETH experienced broader contagion across 20+ chains where the wrapped asset was deployed.
Bridge users face opportunity costs from frozen funds. Taiko's advisory to withdraw from all bridges creates friction and gas costs for users, with remaining bridge TVL at risk of further decline as trust erodes.
Protocol treasuries bear recovery costs. The Humanity Protocol launched a token swap program following its $36 million exploit, effectively socializing losses across the ecosystem. Recovery rates across 2026 bridge exploits average 13.7%, according to PeckShield.
Infrastructure providers face reputational damage. LayerZero, which provided Kelp DAO's bridge messaging layer, and Intel SGX, whose enclave technology underpins Taiko's proof system, face questions about the security assumptions their products are marketed on.
Total DeFi exploit losses reached $840 million in the first five months of 2026, per Decrypt data. Bridges account for a disproportionate share relative to their protocol count, suggesting the category carries systemic risk for the broader ecosystem.
The Taiko exploit is small by 2026 standards — $1.7 million against a backdrop of $342 million in cumulative bridge losses this year. Its significance is diagnostic rather than systemic. An RSA-3072 private key committed to a public GitHub repository is not a sophisticated attack vector. It is an operational failure that the project's security architecture should have rendered harmless but did not.
The bridge security problem has not improved materially since the $1.1 billion lost across Ronin, Wormhole, and Nomad in 2022. The attack surfaces have shifted — from validator compromise to SGX key leaks to single-node RPC quorums — but the underlying dynamic persists: bridges hold high-value assets behind trust assumptions that are cheaper to break than to maintain. Until cross-chain verification can be performed natively rather than through external attestation mechanisms, bridges will remain DeFi's most concentrated point of failure.