Sui Foundation said Sui Mainnet suffered three outage incidents on May 28 and May 29, 2026, shortly after the network’s 1.72 release introduced address balances. The foundation said the halts were tied first to gas-charging edge cases and later to a separate randomness-state bug exposed during validator restarts. It also said no user funds were at risk and no committed transactions were reverted when the network resumed.
Sui Details Mainnet Outages After Upgrade
Sui Foundation said the first outage began at around 7:00 a.m. PT on Thursday, May 28, and ended around 1:30 p.m. PT. A second halt began around 5:00 a.m. PT on Friday and ended around 8:30 a.m. PT, followed by a third outage from around 1:30 p.m. PT to 7:20 p.m. PT. Validators have since addressed the known issues tied to both the gas-charging bug and the randomness-state bug, and network activity has resumed.
The foundation framed the incidents as separate but connected operational failures following the 1.72 upgrade. “The first two stemmed from crash bugs involving the interaction of gas charging logic and the recent 1.72 release, which introduced address balances,” Sui Foundation wrote. “The fix for the Thursday issue was an interim measure designed to restore functionality to the network while the Sui Core Team worked on a long-term solution.”
That interim response carried a known trade-off. “The interim fix had a known issue with a low probability of causing a halt,” the foundation said. “The team accepted the risk accompanying this proposal in order to bring the halted network back as quickly as possible while a robust fix was developed. On Friday morning, the network hit a variant of the known issue and halted.”
Gas and Randomness Bugs Behind Sui Halts
The initial technical failure centered on Sui’s new address-balance model, which allows users to store funds and pay gas without relying only on coin objects. Sui transactions can use address balances, coin objects, or a hybrid approach combining coin inputs with synthetic reservation objects. In the affected edge case, a hybrid gas payment could be cancelled for insufficient address-balance funds, but later gas-smashing logic still attempted to debit the same funds, leading to an underflow during settlement.
Sui Foundation described the issue as more subtle than a simple failed payment. “The crash does not actually happen during gas smashing,” it wrote. “To support concurrent withdraws from/deposits to the same address, using an address balance in a transaction emits balance deltas that are reconciled by a system settlement transaction that sums the deltas and determines the new values of each address balance. This is where the crash occurred, due to a negative delta from the gas smashing applied to a zero balance.” The foundation said the conceptual fix was to avoid gas smashing when a transaction is cancelled with InsufficientFundsForWithdraw.
The Friday afternoon outage came from a separate latent bug in how validators preserved distributed key generation status across restarts. Sui validators run DKG at the start of each epoch to bootstrap the random beacon used by randomness-dependent transactions, and DKG requires a higher participation threshold than normal consensus. “When validators restarted to adopt Part 2’s fix, participation threshold wasn’t high enough for the next epoch’s DKG, and it disabled itself as designed,” Sui Foundation said. “But due to a latent bug, that failure verdict was never written to disk — as further restarts followed, each validator came back up unaware DKG had failed.” The fix persisted DKG status across restarts and added a coordinated mechanism to close the stuck epoch, after which randomness was restored.
Sui Foundation said the incidents showed the need for stronger end-of-epoch resilience, cleaner gas-charging architecture, broader invariant testing, and better failure containment so validator crashes caused by specific inputs do not halt the full network. The foundation also said AI-assisted tooling helped engineers query validator logs, inspect cluster state, and assemble metrics during the response, making that area another target for continued investment.
AI Transparency Note: This article was prepared with the assistance of an AI system based on the sources listed and was reviewed, edited, and approved by a human editor before publication. All quotes, data points, and factual claims are intended to be grounded in the cited source material; however, errors cannot be ruled out entirely.