At 13:09 UTC, Sui processed 1,394 eligible transactions. The striking event was not a single dominant operation but a late-minute split: one successful position-close workflow recorded a boundary inflow of 515,247.556531884 SUI and 89,356.091704 USDC, while 24 transactions stopped at the same flash-loan repayment call.
The large movement belongs to digest Ay1ujwtmGeAU83Zidcskk9z9HBmtjyVYG6EQiuYr3mjL. Its two-step workflow ran collect_reward and then close_position. The sender boundary shows the SUI and USDC inflows, but that evidence does not establish profit, price, or an exchange outcome. It is simply the largest proven sender movement in this minute.
A busy but familiar surface
Of the 1,394 eligible transactions, 1,387 had enough execution evidence for semantic labels; seven did not. The leading asserted operations were cancel-order at 296/1,387, place-limit-orders at 253/1,387, update-price at 230/1,387, and record-activity at 149/1,387. Together, those four accounted for 928/1,387 semantically labelled transactions.
The largest sender, 0xcde6dbe01902be1f200ff03dbbd149e586847be8cee15235f82750d9b06c0e04, contributed 342/1,394 eligible transactions. A second address contributed 176/1,394, so the two largest senders accounted for 518/1,394. These are on-chain identities only; the data does not establish whether either address represents a person, service, or automated system. Sponsorship was also substantial: 356/1,394 transactions had a different gas owner from the sender.
The minute moved almost continuously. The median gap among adjacent transactions was 0 ms, the mean was 42.869 ms, and the longest idle period was 473 ms. The final five-second bucket contained 170/1,394 transactions; 54/170 were labelled execute-swap-route. That late burst explains why the minute feels more like a repayment stress test than a simple order-book census.
What failed
There were 85 failed transactions out of 1,394, or 6.1%, below the frozen atlas baseline of roughly 83,649/1,000,000. Failures were concentrated in route execution: 48/85 failed rows carried the execute-swap-route label. Another 29/85 had no asserted behavior, so the evidence does not justify assigning them a more specific purpose.
The leading abort site was flash_loan::return_flashloan_quote, with 24/85 failures. The next largest recorded sites were an unrecorded location at 12/85 and framework balance::split at 10/85. This is a repeated stopping point, not proof that every route had the same economic cause.
Digest GZJykSQW9hxWa3DgKcUGbU2w7PDtk1cwSiypETCdd1u2 shows the sequence clearly. It opened a quote-side flash obligation, routed through DeepBook and Bluefin, and reached repayment with 3.057632 USDC as the submitted amount. It then aborted at return_flashloan_quote with code 14680251. The application effects were rolled back; the sender movement that remains proven is a 0.000389568-SUI gas charge. Sui documents this all-or-nothing behavior for aborted Move transactions, while gas is still charged. [Sui systems paper](https://docs.sui.io/doc/sui.pdf)
A related row, FpGpGVA8ugAVRn4UoKJQoSKuxUv1Uo41hyoW5YUdMH1Y, reached arb_flash::flash_begin, pool::flash_swap, and pool::repay_flash_swap, then stopped at arb_flash::flash_repay_plus. Those package calls had no authored semantic explanation, so the record supports a sequence of attempted calls—not a conclusion about what the route intended to accomplish.
Money without a price tag
The minute’s SUI holding movements totaled 524,467.800070664 SUI in across 610/1,394 transactions and 22,546.726392060 SUI out across the same governed movement set, for a boundary net of 501,921.073678604 SUI. USDC showed a holding net of 86,636.334270 across 67/1,394 transactions; WAL showed 16,978.697636911 across 52/1,394. Gas was tracked separately: the SUI gas lane recorded a net outflow of 1.393090594 SUI across 1,036/1,394 transactions. These lanes are not additive, and none of them supplies prices, profit, or cross-asset value.
For contrast, digest BU1H3NYeZb7FpWQoHn6kMCTbm7cYhtsesFMRVVR8y2Zs ran a three-step WAL/USDC route and did not fail. Its sender boundary changed by -200 WAL, +0.007504744 SUI, and +5.91439 USDC. The route was labelled execute-swap-route, but the stored evidence explicitly does not establish that an exchange occurred.
The store also places 15 exact revision minima in this minute. They are first stored uses after about seven minutes of observation, not proof that packages were newly deployed or introduced on-chain. Their timing does not explain the repayment failures. The strongest concrete explanation remains narrower: ordinary order and price maintenance filled most of the minute, then a concentrated family of flash-route attempts repeatedly reached repayment and stopped there.
Useful exits are Ay1ujwtmGeAU83Zidcskk9z9HBmtjyVYG6EQiuYr3mjL, GZJykSQW9hxWa3DgKcUGbU2w7PDtk1cwSiypETCdd1u2, and the route pattern sui/execute-swap-route/coverage93-r004-committed-or-repayment-abort/v1.