When one missed deadline reaches the rest of the system

Positive marked equity can coexist with an unfunded exit

The additional 7 October comparison shows both token debts increasing while accepted dollar debt fell by $733.93437306. Between the 7 October, 16:21:23 cutoff and 15 October, 06:13:35 UTC, the token obligations again increased while accepted prices partly offset their dollar effect. Gross marked equity improved by $807.63901772. Neither comparison establishes a realized return or payment. Dated accounting reconciliation.

The 5% accepted-collateral decline is an independent financial sensitivity with debt units and debt prices fixed, not an observed spot shock or security test. It shows eligibility changing before gross marked equity is exhausted; it does not establish a liquidator’s funding or eventual proceeds. A common valuation factor affecting collateral and dominant USDT debt can partly offset rather than compound. That countercase depends on the applicable input relationship, which is not an exceptional-control guarantee. Favorable and adverse conditions.

At the new endpoint, substantial USDT and USDC reserve cash is favorable for their suppliers, but is not reserved for this borrower or every concurrent claimant. A truly independent repayment can replenish the reserve receiving it; borrowing or withdrawing from that same resource must also be debited. An original lender can be paid while a replacement lender remains exposed to the collateral-sale interval. Cash, claims and the independent-payment example.

Four conditional paths—and where each can stop

Seriousness and counterevidence · Observed control boundary · Actual valuation paths.

A stabilizing repayment can transfer the remaining credit exposure

7 October event: 05:44:35 UTC. The original Aave USDT reserve receives 10,006.329226 USDT. Morpho’s temporary 10,000-USDC principal returns in the same matched transaction. Yet Aave’s USDC reserve has supplied replacement credit, and the borrower retains its sUSDe collateral. The own-cash contribution of 4.688632 USDT is repayment, not a fee. Actual funding sequence.

The later state at 14:04:11 UTC records 5,866,650.101604 USDT and 10,000.460197 USDC still owed. Original and replacement reserves cannot be collapsed into one undifferentiated dollar pool. The USDT payment is a genuine stabilizing counterflow, but the new USDC receivable remains exposed. A returned temporary advance is not an equity injection or repayment of all funding.

With an eventually recoverable discount, an admitted refinance can protect the borrower from a forced collateral sale. Under genuine impairment, retaining collateral and rotating the debt token cannot manufacture missing value. These are conditional consequences, not a finding that backing is impaired. No scalable full-position financing follows from the observed small draw.

The selected shareholder’s new standard-withdrawal window is a different type of evidence: a successful query with zero matches over Base blocks 52,294,247–52,296,246. It does not show an attempted failure or selected-holder costs. October 2 ordinary admission remains positive at its original date. The separate historical fee-denominator request remained unavailable; missing data cannot become zero expense. Selected-holder claim and expense boundary.

A payment shortfall and a fee claim reach different people

5 October documentary analysis; the natural event remains dated 4 October. A refinance can pay the original supplier while moving the remaining conversion and waiting risk to another creditor. A shareholder can receive a direct lending claim rather than cash. A fee recipient can receive new shares rather than a USDC payment. None of those transitions supplies the selected holder’s missing all-in expense record. Connected consequences.

Valuing the 3.309050057993409206 fee shares from Base block 52,154,675 on 4 October, 07:24:57 UTC at that transaction’s withdrawal ratio gives about 3.349853 USDC-equivalent, approximately one quarter of the event’s 13.399412-USDC accounting interest increase. It is not a new 25% parameter read, principal charge, gas expense or exact selected-holder dilution. The other holder’s 24,663.478510-USDC payment precedes the later snapshot and is not deducted again.

Coinbase’s described 1% spread applies to optional MORPHO reward conversion, not all withdrawn principal. The served lending help page also links a different vault address from the selected one. Product terms establish neither a corporate advance to these sampled accounts nor their actual fee schedule. The ordinary October 2 admission findings remain positive; missing cost data are not a failed withdrawal. Applicable fee bases and product identity.

A new independent capital contribution or admitted refinance can genuinely reduce urgency. Cash circulated out of and back into the same resource set cannot be credited as both new funding and unchanged sale inventory. The bounded cash ceilings in the new chapter retain their named corridors, conversion assumptions and October 4 clocks; they are not a scenario in which every route necessarily fails.

One holder’s cash receipt is not another holder’s fee result

4 October 2026 · Base block 52,154,675 · 07:24:57 UTC. A natural Withdraw event and burn of 24,363.064586372408152658 shares match a 24,663.478510-USDC transfer to another holder’s receiver. This is positive scoped payment evidence, not an observed withdrawal by the selected approximately three-million-USDC shareholder. What the event match establishes.

A simultaneous 3.309050057993409206-share performance-fee mint creates a separate asset claim. Its approximate 3.349853-USDC equivalent at the event’s payment-per-burned-share ratio is not another USDC outflow, gas or a charge wholly attributable to one holder. No independent receipt, fee/sponsor record or exact selected-holder dilution denominator was obtained. Matched fields and transaction.

The payment is 16 blocks and 32 seconds before the new Base snapshot. It must not be subtracted again from a later stock. Of 35 outflow logs, 29 led to the adapter; those were not labelled shareholder withdrawals. Nor is the matched payment attributed to a particular underlying market without evidence. Cash classification and ordering.

The selected holder now previews 3,000,466.674703 USDC on unchanged shares. Its positive October 2 normal admission stays dated and valid as evidence, but the new seven-market count is not a refreshed five- or eight-market cash screen. Actual selected-holder net expense and timely competing access remain unestablished. Selected holder and the remaining conditions.

An in-kind recipient can inherit the unpaid market exposure

3 October documentary analysis; October 2 claim and penalty inputs. A cash withdrawal, a force withdrawal from liquid allocations and an in-kind exit do not deliver the same asset. Published in-kind mechanics can return idle USDC plus direct market supply positions. Leaving the vault’s future curation for the transferred portion does not remove borrower credit or liquidity risk. Claim-transfer explanation and source limits.

The measured 3,000,011.478695-USDC shareholder preview still fits the 39,707,049.090363-USDC ordinary cash screen under the October 2 no-prior-competition comparison. The separate hypothetical zero-idle in-kind budget instead returns 2,999,981.478880 accounting units of direct claims, approximately 29.999815 of retained share-penalty value, and zero USDC before external costs. It is not an observation of new illiquidity or proof of alternative-route admission.

A creditor demanding cash cannot count those claims as a USDC receipt unless it separately agrees to accept or finance them. The penalty does not create an external cash buffer; an optional referral-fee exclusion does not establish zero network or later sale costs. Competing use may change the amount and form available, but no new October 3 cash movement or cost record was observed. The positive normal admission below is preserved.

A claim can fit ordinary cash—and become costly or delayed after other users act

The 2 October shareholder result is more specific than an aggregate cash estimate. At Base block 52,071,788, 09:22:03 UTC, exact shares previewed 3,000,011.478695 USDC, while the designated market’s claim-capped ordinary cash screen was 39,707,049.090363 USDC. Earlier normal views returned true for the holder sending shares and both the holder and published bundle receiving assets. These observations narrow the old personal-access gap. Measured claim and normal checks.

With the sampled conditions maintained and no earlier competing cash use, the claim fits ordinary cash before external costs; no forced leg or force-deallocation penalty is economically necessary in that case. The bundle allowance and holder native-ETH balance nevertheless returned zero. Neither means there is no ordinary cash, and neither supplies execution funding or authorization. The native zero maxWithdraw follows a different, conservative maximum convention.

The middle case is primarily a share-cost effect: mobilized cash can serve the remaining gross budget, but part of share value bears the penalty. The last case is primarily a cash constraint. Burning share value does not produce the unavailable USDC. These are financial comparisons, not executable limits or promises that an under-covered request automatically pays partially. Ordering, cash caps and cost boundaries.

The actual Aave partial repayment and these vault resources are not an observed common funding arrangement. A repayment financed by cash first withdrawn from the same set changes that set by minus withdrawal plus repayment, not by repayment alone. No shared beneficial owner or rescue agreement is inferred. Track the remaining creditor and count each cash unit once.

A gross exit budget can exceed the cash reaching the holder

1 October analysis; September 28 inputs remain frozen. That shareholder model distinguished gross share-value debit from net USDC, using the pinned Morpho SDK design’s share-penalty and receiver conditions. It did not identify an actual investor or establish receiver admission. The 2 October result above now supplies exact shares and positive normal sender/receiver views; authorization, costs and subsequent cash competition remain separate. The older illustration is retained with its own amounts and date.

The illustrative 50-million cash demand still has the earlier 1,304,682.927618-USDC gap against the frozen 48,695,317.072382-USDC cash screen. A penalty-inclusive 50-million gross debit budget answers a different question; it cannot be substituted to make that cash gap appear smaller. An under-covered request need not partially fill.

The illustrative competition sensitivity assumes a prior independent 5-million-USDC withdrawal with no replenishment. It reduces the later cash screen to 43,695,317.072382 USDC and increases the gap against the same assumed cash demand to 6,304,682.927618 USDC. Neither demand nor withdrawal is observed.

Repayment helps only when its source is counted correctly. In the separate accounting illustration, lending 1 million from this resource set and then receiving 1 million back adds zero new cash; an independent outside repayment adds cash before later uses. No authority or facility for that recycling is asserted. Gross/net conventions · Competition and funding provenance.

Other resources can reduce the gap without proving a withdrawal

The 28 September supplier comparison keeps the existing assumed 50-million-USDC demand but replaces the unexamined other-assets residual with an eight-market resource screen. It is not an observed withdrawal queue or a specific shareholder’s entitlement. Source clocks, permissions and competing uses still matter. The exact screen and accounting boundary.

These rows are alternative resource cases, not four additional cash balances. The later borrower/quote batch enters as a conditional future repayment, not as same-block cash. A repayment arriving too late does not fund the earlier demand; a share penalty is not an outward USDC fee. Deadline and non-overlap qualifications.

The creditor transfer remains explicit in the September 28 comparison. A prior USDC funder can pay down the Base loan while still relying on later proceeds; its million-unit enquiry did not confirm full consumption. That date’s Aave examples owed USDe, leaving their lenders with dollar-value risk even under a common collateral/debt dollar-mark change. The 2 October USDT account is a different cross-token case, not a change of denomination in those older positions. Earlier creditor and funding comparison.

The same return can be sufficient—and still arrive too late

The 20 September investigation adds a threshold based on observed Mint USDC inventory, with demand explicitly assumed. Hold 30,934,109.172 USDC at Ethereum block 26,019,040 fixed. Assume four separately accepted orders, each for ten million USDe and each paying one USDC per USDe at successive permissible deadlines. Assume no other uses, pre-consumed limits or conversions. These are not observed orders, a queue, a price promise or a four-block throughput forecast. Inputs and conditional model.

Here, a funding bridge means temporary financing, not a blockchain bridge. No Maple or JAAA receipt date is assumed. The amount is the largest cumulative deficit before any deadline, not a sum of repeated deficits or a claim that a lender will advance it. An initially funded bridge brings a matching debt or assigns an existing recovery claim. It cannot repair a permanent principal loss without genuinely loss-absorbing capital. The cash ledger and separate impairment test.

For a claim with principal P and net recoverable principal R, a positive P − R is a value shortfall even with unlimited waiting. A full recovery after an earlier deadline instead leaves a timing problem. Both can affect a forced seller, but they must not be counted as independent losses without tracing the same resource. Recovery costs and impairment · Unreconciled issuer-wide buffers.

The same invented balance sheet can answer different questions

The following normalized balance sheet belongs to the 18 September foundation. It is a separate hypothetical example, not a portfolio calibrated to the newly observed Maple, JAAA or Mint amounts. Do not add its scenarios to the funding example above.

The supplied balance sheet uses liabilities of 1,000 and assets of 1,015: liquid stablecoins 300, DeFi loans 350, Treasury claims 150, CLO claims 100, net hedged-crypto backing 100 and a reserve 15. Categories are nonoverlapping by assumption; the reserve is already included in assets. These are hypothetical normalised value units, not scaled estimates of Ethena’s portfolio. Complete supplied sensitivities.

Recoverable assets and accessible cash are different dimensions. A freeze need not imply final loss; a default can. A margin payment is a cash use in the timing example, not an additional assumed economic loss. Reusing the same reserve for several obligations would overstate protection.

Repaying the protocol can move the remaining risk to another creditor

The 27 September funding example assumes an existing Aave position and 90,000 Ethereum USDC provided by a distinct term lender. The funds arrive first, repay Aave and allow collateral to be released subject to account and reserve conditions. The new creditor then waits for the staking claim and later sale. This is a financial construction with explicit funding—not an observed loan, actual Aave account or executable route.

At the frozen 100,000-USDe sale indication, 99,923.537191 USDC arrives later. With assumed 10% simple annual term credit, principal plus interest is 90,024.657534 USDC after one idealized day; the residual is 9,898.879657 USDC before gas or other costs. At three days the residual falls to 9,849.564588 USDC. A positive eventual residual cannot finance the earlier 90,000-USDC repayment. A day-one debt with day-three sale proceeds remains late without an extension or further funded arrival. Assumptions, chronology and conserving cash ledger.

The one-million-USDe quote cannot be priced at ten times the smaller output. Under the separate 900,000-USDC debt example, the same pool’s large-size indication leaves 657,162.08 USDC of extra contribution required. A negative residual is a funding shortfall, not a completed payment. Independent Ethereum quotes and their size limit.

There is a genuine supplier stabilizer on Base: an independently funded, timely full repayment by the sampled borrower reduces a hypothetical 50-million-USDC selected-market withdrawal gap from 11,227,446.31 to 8,522,886.48 USDC. The same payment arriving later does not meet the earlier deadline. Neither demand nor that repayment is observed; no other vault asset is assumed worthless. Same-batch inputs and conditional counterflow.

Collateral release can require money that arrives before the exit

Take the supplied hypothetical borrower with an initial health factor of 1.05. Assume fixed short-interval debt and a 5% fall that the actual collateral oracle recognises. The resulting health factor is 0.9975, below the liquidation threshold. Liquidation or repayment funding can be needed within blocks or minutes, before released sUSDe can enter a queue and ultimately reach a separate sale or issuer redemption. Ordered health-factor example.

This is not a statement that a 5% external USDe spot discount automatically enters Aave’s ratio / USDT input or Base’s selected par primary. If backing is sound and the discount recovers, keeping the accepted value stable can protect borrowers from forced loss. Under genuine impairment or unavailable local exit, delayed recognition can instead expose suppliers to poorer recovery. The Base switch / fallback semantics are not certified. Current-design countercase and limits.

For a venue account, a haircut can obstruct a transfer before liquidation. The separate supplied illustration assumes collateral factors of 95%, 90% and 79% against initial margin 90 and maintenance margin 80. At the assumed 90% factor, the transfer threshold binds while the maintenance ratio is only 88.8889%. Those factors are not measured USDe UTA tiers or a complete venue margin engine. Published gate distinction.

A protective decision can change who must wait

Customer-admission and last-look checks can protect common backing while postponing a legitimate holder’s payment. A fund board or custodian can preserve an orderly recovery process while the issuer’s customer faces an earlier deadline. A contract delay, a committee’s deliberation and the holder’s own clock are not synchronized. These are conditional pathways, not reported incidents. Four distinct clocks; external release authority.

Revenue policy supplies another possible transmission path. The 27 August 2026 fee-switch proposal and the Foundation’s 8 September reply reporting passage describe conditional allocation toward ENA buybacks. Implementation and actual transfers were not independently verified. A lower staking distribution can make leveraged carry less attractive; a subsequent unwind may require outside USDC before collateral is free. Income change is not itself principal loss, and no resulting outflow is measured here. The documented decision and its limits.

Calling a backing asset “liquid” or shortening a cooldown does not speed up a borrower, fund or bank. The March cooldown-policy discussion distinguishes liquidity routes, including dependence on USDe itself. A reserve commitment to a local facility is not an additional free buffer, and a proposed token seed does not supply the external payment asset. The earlier hypothetical scenarios remain separate from these documentary findings. Liquidity classification; conditional local-conversion proposal.

A stopped route can outlast the notice window

The 26 September source-based continuation sharpens the service branch without measuring a new outage. Selected immediate revocation can be protective, but re-granting or repairing a route may use another authority path. Cancelling a pending instruction does not perform the payment or restoration it was meant to support. A ready ordinary operation has no general maximum execution time in the examined source. Stop/restore and source limits.

A holder’s relevant comparison is notice + collateral release + remaining cooldown + claim/conversion, not the controller delay alone. In the dated 18 September configuration, controller minimum and cooldown were each 86,400 seconds; those two labels leave no positive margin for additional steps. This does not declare every exit too slow: free USDe, an already-running queue, independent repayment cash or a longer scheduled wait can change the outcome. Necessary-condition analysis and countercases.

Restoration also needs the resource. The later PSM deployment claim does not prove usable local payout stock; a local atomic swap does not prove bridge-independent restocking. An independently funded arrival can close a timing gap, but it must be received by the required entity before the deadline and counted once. The earlier numerical scenarios above remain unchanged and separate; no new portfolio weights, recovery dates or loss probabilities are introduced. Local capacity conditions.

Repayment is a stabiliser only when it can actually arrive

Borrowers repaying USDC can restore lender liquidity and reduce collateral exposure. But repayment may need outside funds before the pledged collateral is released. Issuer redemptions, lending withdrawals and venue margin can compete for an asset in the same location at the same time. Counting eventual proceeds before that order has been respected turns a possible counterflow into invented immediate cash. Lending and ownership evidence.

The narrow dust result in one named Ethena-oriented Base adapter weakens one proposed reciprocal-funding path. It does not supply a portfolio-wide legal-owner or shared-funds map. Therefore no current circular-funding amount or system-wide death spiral is assigned to these scenarios.

The 10 October 2025 episode supplies a bounded historical comparison, not an exact replay. The issuer’s 17 November retrospective reported a Binance USDe/USDT low of 0.6567 and more than $2bn of primary redemptions in 24 hours. Those remain attributed issuer claims, not independently reconstructed prices, summed receipts or proof every holder could exit. The Aave discussion began 24 October, with a contrasting technical response on 2 November. Historical accounts and their limits.

The useful distinction is local exit access versus primary backing—not a guarantee that any proposed support facility is funded today. No full event-block replay, current reserve reconciliation or measured cross-route stress liquidation capacity was obtained.