STARKNET OUTAGE DUE TO EXECUTION LAYER STATE DISCREPANCY

• Starknet's mainnet experienced a temporary outage due to a discrepancy between the execution layer and the proving layer, preventing faulty transactions from being finalized.
• The proving layer functioned as intended, detecting the error and halting the faulty execution before it reached Ethereum's finality.
• Approximately 18 minutes of network activity was reverted, but
  
  no funds were lost, and the network has since returned to normal operations.

Starknet, an Ethereum layer-2 scaling network, temporarily halted its mainnet on January 11, 2026, due to a software bug in the blockifier execution layer. The issue stemmed from a specific combination of cross-function calls and variable writes that led to an incorrect state retention after a function reverted. This discrepancy triggered a block reorganization, rolling back approximately 18 minutes of on-chain activity.

The proving layer, a core design feature of Starknet, identified the inconsistency and prevented the faulty transactions from being finalized on Ethereum. This safeguard ensured that the incorrect execution was never settled, reinforcing the network's commitment to security and accuracy. The network resumed normal operations after the block reorganization, and the total value locked (TVL) remained at $840 million throughout the incident.

What Caused the Starknet Outage?

The outage originated from a specific bug in the blockifier execution layer, which incorrectly remembered a state-write that should have been discarded after a function reverted. This occurred under a combination of cross-function calls and reverts, leading to an incorrect transaction execution. The blockifier operates differently from the proving layer, using a simplified simulation to maintain high throughput. The divergence from expected behavior during these function calls led to the mismatch detected by the proving layer.

The proving layer, responsible for verifying the correctness of transaction execution, flagged the inconsistency and halted the faulty execution before it could be finalized. This mechanism ensured that the error never reached Ethereum's finality, preventing any permanent damage. The team acknowledged the need to improve stability as layer-2 technology continues to evolve.

What Are the Implications for Users and Investors?

The market reaction to the outage was muted, with the STRK token trading around $0.089 at the time of writing. Despite the relatively short duration of the outage, STRK's price remains well below its early-2024 peak of $4, reflecting broader market conditions and concerns about supply and network reliability. The incident highlights the challenges of maintaining stability as Starknet scales and introduces new features.

Users affected by the outage had to resubmit transactions that were rolled back, which could be problematic for time-sensitive operations. However, the proving layer's intervention ensured that no funds were lost, and the network resumed normal operations quickly. Traders and applications that depend on fast and predictable transaction execution were impacted by the disruption.

What Steps Are Being Taken to Prevent Future Disruptions?

Starknet has committed to improving testing and code audits to prevent similar incidents in the future. The team is implementing new fuzz-testing suites to compare blockifier execution results directly with the proving system, helping to identify edge-case interactions that could lead to incorrect state handling. Additionally, the team plans to shorten the time between transaction execution and prover-compatible execution.

This will allow mismatches to be detected sooner, limiting how much network activity would need to be rolled back in the future. The team acknowledged the need to improve overall stability as layer-2 technology continues to mature. These steps aim to reduce the frequency of disruptions and enhance the network's reliability as it transitions toward a more decentralized architecture.

The recent outage follows a similar incident in September 2025, where a protocol upgrade caused a more severe disruption lasting over five hours. The team has since committed to architectural changes and expanded monitoring to reduce the likelihood of such issues in the future.