Coinbase has switched on direct rupee bank transfers in India via IMPS, letting users move funds between local bank accounts and the exchange for spot and futures trading.

Sui experienced three mainnet outages on May 28 and May 29 tied to its v1.72 release. Two halts came from a gas-charging bug exposed by the new address balances feature. The third halt followed a latent randomness bug triggered during validator restarts. The Sui Foundation published its full review of the three mainnet outages that knocked the network offline across May 28 and May 29, 2026. The firm claimed that the disruptions were caused by two distinct bugs in the v1.72 software release. The post-mortem confirmed that the first two halts shared the same root cause, while a third halt was triggered by a latent randomness-state bug exposed during validator restarts. The first outage began at around 7 a.m. PT on Thursday, May 28, and ended at about 1:30 p.m. PT the same day. The second halt ran from roughly 5 a.m. PT to 8:30 a.m. PT on Friday, May 29. The third outage began at around 1:30 p.m. PT on Friday and ended at about 7:20 p.m. PT. The Foundation said no user funds were at risk across the three events and the network did not revert any committed transactions when it resumed. SUI , the network’s native token, trades at $0.8776, down 2.6 percent on the day, 15.7 percent over the past week, and 73.0 percent over the past year. Gas-Charging Bug Drove the First Two Sui Outages The v1.72 release added a feature called address balances, which gives users a new way to store funds and pay for gas without using coin objects. Sui transactions can now pay for gas using an address balance on its own, coin objects, or a mix of both, which the team calls hybrid gas. For transactions paying with coin objects or hybrid gas, the runtime performs gas smashing before charging the transaction. The process combines all input coins into a single coin that is then debited for gas. The step runs for transactions that execute successfully and for transactions that are cancelled. The root cause of the first two outages sat at an edge case inside the hybrid gas path. If a reservation attempted to overdraft an address balance during the budget check, the attempt was blocked and the transaction was cancelled with an InsufficientFundsForWithdraw error. The Foundation said the crash did not happen during gas smashing itself. Using an address balance in a transaction emits balance deltas that are reconciled by a system settlement transaction. The crash came from a negative delta produced by the cancelled-but-still-smashed gas being applied to a zero balance during settlement. The condition could only happen when two transactions hit the scheduler at the same time and competed to spend funds from an address balance that could not cover both. The scheduler cancels one of them with InsufficientFundsForWithdraw to prevent the overdraft, but the cancelled transaction still debited funds through gas smashing. Interim Fix Came With a Known Risk That Triggered the Second Halt The fix the core team proposed on Thursday afternoon stopped the system from smashing gas when a transaction was cancelled with InsufficientFundsForWithdraw. Enough validators adopted the patch to bring the network back at about 1:30 p.m. PT, with the team accepting a known risk attached to the interim approach. The Foundation said gas logic changes are delicate work. Address balances interact with coins in complicated ways. Changes must either preserve all previous behavior or apply version gating, since nodes can fork while replaying old transactions under new logic. Sui’s gas charging also includes conservation checks that prevent any transaction from creating or destroying SUI. Skipping the step that credits any charged funds to the appropriate place would cause a crash. Charging expensive transactions also serves as a key piece of denial-of-service protection. The interim fix had a shortcoming that the team flagged when it shipped. A transaction can carry multiple reasons for cancellation, and one reason can override the others. A transaction using address balances might be cancelled because too many higher-priority transactions are queuing to touch the same hot shared object, then also be cancelled for InsufficientFundsForWithdraw when another transaction spends from the same address balance. In that scenario, the InsufficientFundsForWithdraw error gets masked by the other error, bypassing the patch and triggering the same underflow. That exact scenario hit the network on Friday morning, leading to the second outage. The team was close to completing a more durable fix at the time and finished in time to propose the new patch to validators by about 8 a.m. PT. Enough validators adopted it to bring Sui back up by 9:40 a.m. PT. Randomness State Bug Caused the Third Sui Halt The network ran normally from 9:40 a.m. PT until about 1:30 p.m. PT on Friday, when the scheduled epoch change failed to complete and the network halted a third time. The Foundation said the third halt came from a latent bug whose conditions were set by the previous restart cycle. At the start of each epoch, Sui validators run a distributed key generation, or DKG, protocol that bootstraps the random beacon used by transactions that depend on on-chain randomness. The DKG requires a higher participation threshold than normal consensus. If participation falls short, randomness disables itself for the rest of the epoch as designed. When validators restarted to install the Friday morning fix, participation for the next epoch’s DKG was not high enough, and the protocol disabled itself. The latent bug meant the failure verdict was never written to disk. As further restarts followed, each validator came back up unaware that DKG had failed. Randomness-dependent transactions expect to either execute or be cancelled. With validators no longer holding the record that DKG had failed, neither could happen. The paused queue grew, and the end-of-epoch logic, which has to drain that queue before closing, was left waiting on a DKG that would never arrive. The fix carried two parts. The first piece corrected the bug and added logic to persist DKG status across restarts. The second piece added a mechanism that lets validators close a stuck epoch at a coordinated point. The team used the new mechanism once to close the affected epoch. The network then moved into the new epoch normally, and randomness was restored. What the Sui Team Took From the Three Outages The foundation set out four takeaways from the week. End-of-epoch resilience was the first, with the team noting that the existing safe-mode fallback for epoch transitions may be too narrow. The Foundation said the ecosystem needs to extend graceful-degradation patterns across the rest of the reconfiguration path and turn the force-close mechanism into a standing operational capability. The second takeaway covered the gas-charging logic itself. The crashes in parts one and two both stemmed from bugs in gas charging, a corner of execution that interacts with the address-balance settlement system, conservation checks, and the scheduler. The team said the logic is now complex enough that edge cases are hard to rule out by inspection alone. Coming out of the incident, the Foundation said gas charging deserves the same code-quality bar as the Move VM or the Mysticeti consensus engine. The third takeaway covered AI tooling. AI agents with access to production state, capable of querying validator logs interactively, inspecting cluster state, and assembling metrics on demand, materially accelerated diagnosis during the week’s incidents. The fourth takeaway covered failure containment. The crashes in the first two outages were each triggered by specific inputs the validators could not process safely. The Foundation said the system lacks a defense-in-depth layer that would bound the blast radius of such a crash. 

Were these outages indications of more serious architectural issues or just uncommon edge cases?

BNB is attracting renewed attention as stronger participation and improving market structure reshape its outlook.