Sui's mainnet is processing transactions again after an outage that stopped block production for several hours, the network's validator team confirmed in a post on its official account dated May 30, 2026. The team said the chain hit an end-of-epoch halt, that more than two-thirds of stake had upgraded to a patched build before the restart, and that user funds were never at risk.
The stall ranks among the longest in Sui's history. Several outlets, including Cointelegraph and CoinDesk, put the downtime at roughly six to seven hours and described it as the chain's longest recorded stall. SUI fell during the event and traded near $0.92 on May 29, 2026, down about 16% over the prior week, according to market data cited in that coverage.
A bug in the gas-charging path froze block production
Independent reporting traces the halt to a crash bug in the gas-charging logic introduced in version 1.72 of the node software. The defect caused validators to crash and stopped the network from producing new checkpoints, the mechanism Sui uses to finalize batches of transactions. Sui's own message framed the incident around an end-of-epoch halt rather than naming the exact code path. The two descriptions line up if the faulty logic triggered as the network crossed an epoch boundary. A full post-mortem has been promised in the coming days.
For users, the practical effect of a checkpoint stall is simple. Transactions stop confirming. Wallets show pending transfers that do not settle, decentralized exchanges on the chain cannot match orders, and anything that depends on on-chain finality, including stablecoin payments routed through Sui, waits until block production resumes.
Validators patched and re-quorumed before the chain resumed
Recovery on a proof-of-stake network requires coordination. The fix had to be distributed, and enough validators, weighted by stake, had to run the corrected build before the chain could safely advance. Sui put that threshold at more than two-thirds of stake, the standard quorum needed to keep the network's safety guarantees intact. Once that share had upgraded, block production restarted and the backlog cleared.
The detail that funds were never at risk matters here. A liveness failure, where the chain stops advancing, is different from a safety failure, where the ledger records conflicting or invalid state. Sui experienced the former. Balances stayed correct throughout. They were simply frozen in place until the network came back.
Two long halts in five months sharpen the reliability question
This is the second multi-hour outage Sui has logged in 2026. A similar stall in January knocked the network offline for a comparable stretch, and the repeat has reopened a debate about how often high-throughput chains trade resilience for speed. Sui's design pushes for parallel execution and very low latency, and critics argue that aggressive performance targets leave less margin when a single release carries a defect.
The counterpoint from the team is that both incidents were liveness events resolved without loss, and that fast coordinated patching is itself evidence the validator set functions. Both things can be true. A chain that recovers cleanly beats one that does not, and a chain that halts twice in five months still owes its users a clearer reliability record than one that never stalls at all.
For anyone holding or moving assets on Sui, the takeaway is operational rather than existential. Network uptime is infrastructure risk that sits on top of the usual custody questions. Holding your own keys protects against an exchange freezing withdrawals, but it does not keep a chain online. When the base layer stalls, self-custodied and custodial users wait the same length of time. The post-mortem, once published, will show whether the v1.72 release path had test coverage that should have caught the bug before it shipped.
Overview
Sui restored mainnet after a multi-hour end-of-epoch halt, with more than two-thirds of stake upgrading to a patched build before block production resumed and no loss of funds. Reporting attributes the stall to a gas-charging bug in version 1.72. It is the network's second long outage of 2026, and a full incident review is pending.
