At 15:28 UTC, one transaction recorded a sender-side inflow of 647,233.532222567 SUI while the minute was also unusually failure-heavy. The minute contained 2,691 eligible transactions, of which 528 failed—19.6%, compared with 8.36% across the frozen corpus. Of the 2,645 transactions with semantic interpretations, 511 failed, or 19.3%.
The activity otherwise looked like a busy orderbook and liquidity-maintenance window. Among those 2,645 interpreted transactions, 607 asserted order cancellation, 574 asserted limit-order placement, 254 asserted swap-route execution, and 222 asserted price updates. One address, 0xcde6dbe…c0e04, supplied 661 of the 2,691 transactions—24.6% of the minute—but an address is not evidence of a person or organization. Another 517 transactions were sponsored, meaning a different address paid the gas.
The outlier was a position close
The largest SUI holding movement in the minute was 6oHBxYdK, at checkpoint 300484463. Its pattern describes a concentrated-liquidity position close with staged reward handling. The transaction constructed an empty CETUS coin before one claim call, constructed an empty SUI coin before another, and then invoked the position-closing terminal.
At the sender boundary, the transaction changed by +647,233.532222567 SUI, +46,778.936028 USDC, and +1.77453194 CETUS. The detail does not reveal the individual reward amounts or attribute those boundary movements to a particular call. These are therefore observed balance changes, not proof of profit, a payout, or an exchange.
The same minute contains a smaller but suggestive sequence from 0x6b1f33…042c58. At 15:28:00.673, 81YbkTP1 closed a concentrated-liquidity position and recorded +399,994.346980035 SUI and +262,001.471347 USDC. At 15:28:28.100, 3yLWiPiA from the same address opened a position after preparing and funding several values, recording −399,994.346980035 SUI and −262,001.717696 USDC.
Together, those records resemble a position exit followed by a new funding step. They do not prove that the two transactions concerned the same position, nor do they establish an economic return.
Where the failures stopped
The failure pattern was not one universal error. The largest group—169 of 528 failures, or 32.0%—had no recorded abort site. Among named sites, 59 stopped at a quote-side flash-loan return call, 40 at the framework balance-splitting operation, 38 at a minimum-output check, and 31 at a base-side flash-loan return call.
C9ABmsuJ shows the route-shaped failure clearly. It opened a quote-side flash loan, passed the borrowed USDC through Bluefin and Cetus calls, and reached the matching return call. That return call aborted, so application effects were rolled back; the reached swaps describe attempted steps, not completed exchange outcomes. The sender’s retained movement was only −0.000790532 SUI for gas.
Another transaction, Hk6g7yDm, reached an order-placement path after submitting 130.815634814 SUI and a zero-DEEP quantity, then stopped at a minimum-output guard requiring 1,134,788,616 units. The guard value is not evidence of actual output, and the attempted application changes were rolled back.
The minute therefore reads as a mixture of rapid orderbook maintenance, liquidity-position cycling, and route attempts that often reached repayment or validation boundaries. The large SUI movement is real at the transaction boundary, but the frozen evidence does not establish who controlled the address, whether any swap settled economically, or whether the position close produced a gain.
Useful exits: 6oHBxYdK, the address 0x6b1f333b9c3f7a4ea29e0d4b9d7cf3d8bb9e4770161fd1e56dd9dde192042c58, and the route failure C9ABmsuJ.