A liquidator holding 250 tBTC should have a fairly straightforward job. Sell the Bitcoin-backed tokens, recover the debt, collect the liquidation bonus, and move on.
Except the numbers no longer work quite that neatly.
On November 3, a contributor to Aave’s governance forum tested what would happen if 250 tBTC were swapped for Wrapped Bitcoin through ParaSwap. The result was 228 WBTC. The missing 22 Bitcoin-equivalent units represented an 8.8% execution discount.
That was not a completed trade. It was a market quote, captured at Ethereum block 26,332,529. Still, it exposed something uncomfortable about a token designed to represent Bitcoin one-for-one.
Threshold Network says tBTC has enough Bitcoin to back every token in circulation. Its reported reserves actually exceed supply.
So why could selling just 250 tokens become so expensive?
The answer sits somewhere between a temporarily disabled Bitcoin bridge, a lending protocol’s pricing assumptions, and a market that looks much more liquid on a screen than it does when someone tries to leave.
The Bitcoin Is There. The Exit Is Another Story
On October 30, Threshold Network announced a precautionary suspension of new Bitcoin deposits, tBTC minting, and redemption requests.
Existing tBTC could still be moved through Ethereum and other supported networks. Users could hold it, transfer it, or trade it. What they temporarily lost was the normal route for converting those tokens back into native Bitcoin through Threshold.
That distinction matters.
When someone deposits BTC into Threshold, they receive tBTC and can use it across decentralized finance. The underlying Bitcoin remains secured through a distributed signing system, allowing the token to represent a claim on Bitcoin without relying on a conventional centralized custodian.
Ordinarily, users can reverse that process through redemption.
But Threshold’s response to a security concern interrupted that final step.
In its November 1 security update, Threshold explained that operators had patched their systems or taken affected nodes offline following the disclosure of a Linux kernel vulnerability, CVE-2026-98432.
The project said it had found no evidence that anyone exploited the vulnerability against tBTC nodes. A separate incident involving the theft of roughly 0.5 ETH from an operator’s transaction-fee account was disclosed, but Threshold reported no evidence that the operator’s Bitcoin signing material had been compromised.
There was no announcement of missing Bitcoin from the bridge.
Instead, Threshold presented a reassuring set of figures: roughly 4,320 BTC in bridge wallets backing about 4,185 tBTC in circulation.
That works out to a backing ratio of 103.23%, with roughly 135 BTC more in the reported reserves than the nominal token supply.
Those figures come from Threshold’s disclosures, rather than a new independent reserve audit by Ravel. They nevertheless point toward a different problem from the one investors typically fear when a bridge suspends operations.
The problem was not an identified shortage of Bitcoin.
It was the temporary loss of the mechanism that normally allows tBTC holders to claim that Bitcoin directly.
And because DeFi protocols elsewhere accept tBTC as collateral, the interruption quickly spread beyond Threshold.
On October 31, Aave’s risk team moved to prevent its exposure from growing. New tBTC deposits and new borrowing power backed by tBTC were disabled on the Ethereum Core market. Existing users could still repay debt, withdraw available assets, and undergo liquidations.
The intervention did not alter existing positions’ health factors or automatically liquidate borrowers.
But it left another question unanswered.
What happens if a position needs to be liquidated while the collateral’s normal redemption mechanism is unavailable?
A $136 Million Market Meets a Much Smaller Exit
The first clue appeared in the Aave governance discussion.
On November 2, a participant using the name SrAugust examined the Ethereum Core market and found roughly 1,749.5 tBTC supplied, representing about 42% of the token’s reported Ethereum supply.
The 50 largest suppliers accounted for 88% of that amount. Their accounts collectively carried around $90.8 million in debt, although not all of that debt necessarily depended exclusively on tBTC collateral.
At prevailing reference prices, the total tBTC supplied was worth over $130 million.
Yet the token’s market price appeared relatively calm.
SrAugust observed that tBTC had remained within 0.2% of Bitcoin since the bridge suspension.
For an observer checking only the quoted price, that might look like evidence that the market was functioning normally.
But a quoted market price does not reveal how much someone can sell before exhausting the available liquidity.
Another forum participant raised exactly that concern, pointing to the limited depth available across decentralized exchanges and centralized order books.
SrAugust subsequently tested larger swaps through ParaSwap.
The results became increasingly uncomfortable.
A swap involving 150 tBTC could reportedly stay within a 1.6% discount. At 200 tBTC, the quoted output dropped to roughly 191 WBTC, or about 4.6% below nominal value. At 250 tBTC, the output fell to 228 WBTC.
The additional 50 tBTC between those last two quotes produced just 37 more WBTC.
In other words, increasing the input by 25% increased the output by only about 19.4%.
That’s a substantial deterioration for an asset whose purpose is to maintain a one-to-one economic relationship with Bitcoin.
And this is where the liquidation mechanics become particularly interesting.
According to the forum analysis, tBTC’s standard liquidation bonus on Aave was 7.5%, equivalent to around 6.75% after the protocol’s liquidation fee.
Consider a simplified liquidation in which a liquidator receives 250 tBTC after the net bonus.
At a 6.75% incentive, the debt repaid would have a BTC-equivalent value of:
250 ÷ 1.0675 = 234.19 BTC
Now apply the reported ParaSwap quote.
Selling those 250 tBTC would return only 228 WBTC.
That leaves a shortfall of 6.19 BTC-equivalent, before considering gas costs, execution changes, or additional fees.
At the smaller 200-tBTC size, the same calculation produces a different result. Repaying debt worth 187.35 BTC-equivalent and selling the received collateral for 191 WBTC would leave a theoretical surplus of 3.65 BTC-equivalent.
The liquidation can look profitable at one size and unprofitable at another.
These are Ravel’s calculations using the forum’s reported inputs, not evidence of actual liquidation losses. They assume the stated net bonus and exchange outputs apply to the illustrative transactions. Real liquidations depend on position structure, execution routes, liquidity, and changing market conditions.
Still, the reversal is striking.
The cost of selling troubled collateral can wipe out the bonus that Aave offers to encourage liquidators to take on the risk.
SrAugust estimated that the break-even point was around 220 tBTC under the tested market conditions.
With roughly 1,745 tBTC sitting in Aave’s reserve at the time, that transaction size represented just 12.6% of the total supplied balance.
That does not mean the remaining 87.4% was impossible to liquidate. Liquidators could execute smaller trades, find alternative venues, retain the collateral, or wait for redemption to resume.
But waiting introduces another risk: a liquidator whose business model depends on rapidly recycling capital may have little appetite for holding Bitcoin-backed tokens with no immediately available redemption route.
And the problem becomes more complicated when we examine how Aave values that collateral.
The Price Oracle Hasn’t Forgotten Bitcoin
Aave does not determine collateral values simply by asking how much a large trade would receive on a decentralized exchange.
Its risk calculations rely on price oracles.
In a November 2024 assessment, LlamaRisk documented an important distinction between tBTC and WBTC pricing on Aave.
tBTC used Chainlink’s BTC/USD market price feed, while WBTC incorporated a WBTC/BTC exchange-rate feed alongside BTC/USD.
The arrangement reflected the underlying assets’ expected relationship with Bitcoin while avoiding reliance on potentially manipulable decentralized exchange prices.
The November 2026 governance discussion also described the same tBTC pricing arrangement, although Ravel has yet to independently verify the latest live oracle configuration against Aave’s contracts.
The consequence is worth examining.
Imagine Bitcoin trades at $77,000, while a sufficiently large tBTC sale can only obtain $69,300 per token after slippage.
That is a 10% execution discount.
Under the documented BTC/USD oracle arrangement, Aave could still value one tBTC at $77,000 for collateral accounting, even though selling that token immediately at the hypothetical execution price would recover considerably less.
This is an illustration, not a claim that tBTC currently trades at a persistent 10% discount.
The distinction becomes dangerous only under certain conditions. Liquidations still depend on borrowers’ health factors, debt positions, and collateral thresholds. A temporary discount in a large swap does not automatically create bad debt.
But it can create a gap between the value a lending protocol assigns to collateral and the amount a liquidator can actually recover from selling it.
SrAugust modeled a more severe scenario. Under a hypothetical 20% repricing of tBTC, three accounts would fall below a health factor of one. At a 30% discount, one account’s collateral would be worth less than its debt.
Those were community stress-test results, not realized losses or independently reproduced calculations.
Aave had not reported tBTC-related bad debt in the materials reviewed for this investigation.
Nevertheless, the mismatch between oracle valuation and executable liquidity is hardly a new concern.
Risk researchers had flagged the warning signs years earlier.
The Risk Was Written Down Years Ago
In November 2023, LlamaRisk published a detailed collateral risk assessment of tBTC.
The report examined security, market liquidity, governance, and the mechanisms responsible for maintaining tBTC’s relationship with Bitcoin.
It acknowledged that liquidation demand could exceed available decentralized exchange liquidity, pushing tBTC below its intended peg.
But it also identified the mechanism expected to correct that imbalance.
Redemption.
When tBTC became sufficiently cheap relative to native Bitcoin, arbitrageurs could purchase the discounted token, redeem it for BTC, and capture the difference.
The report estimated that redemptions normally took three to five hours.
That delay was manageable enough to support the broader argument that tBTC could maintain a credible peg under ordinary conditions.
Then, in November 2024, Aave risk advisers considered another configuration involving tBTC and WBTC.
Chaos Labs specifically warned that discounts in tBTC’s market price could consume liquidation bonuses, citing an earlier episode in which tBTC traded at a discount close to 3% before recovering.
The proposed configuration was ultimately not implemented after its sponsor withdrew support. Its unusually low proposed liquidation bonus should therefore not be confused with the standard parameter discussed during the current suspension.
But the historical warning remains relevant.
The risk was not that the Bitcoin would necessarily disappear.
It was that selling a Bitcoin-backed claim could become expensive precisely when that claim was supposed to function as dependable collateral.
The October 2026 pause removed the normal redemption mechanism from that equation, at least temporarily.
And there is a final irony.
Just weeks before the suspension, Threshold itself published a due-diligence framework for institutional Bitcoin.
Its central argument was that ownership and operational control must be evaluated separately. Institutions needed to understand not only where Bitcoin was held but also who could move it and how recovery worked when technology failed.
The framework was discussing a broader custody architecture, including Threshold’s planned institutional products, rather than diagnosing the present tBTC incident.
Still, its underlying principle applies remarkably well.
A claim backed by Bitcoin is not the same thing as immediate access to Bitcoin.
Threshold has emphasized that the bridge suspension is precautionary, that its reserves remain sufficient, and that normal operations will resume after security remediation. As of this November 4 review, Ravel could not independently confirm a subsequent announcement restoring normal redemption operations.
That leaves a legitimate counterargument to the more alarming interpretations of the incident.
If redemption resumes promptly, large-scale arbitrage may return, secondary-market liquidity could improve, and much of the current liquidation concern may prove temporary.
Aave’s decision to freeze additional exposure also shows that the protocol has mechanisms for responding before a liquidity problem becomes a realized loss.
Nothing in the evidence establishes that tBTC is insolvent, that its Bitcoin reserves have been stolen, or that Aave has already suffered bad debt from the suspension.
But the episode has exposed something that reserve figures alone cannot answer.
Threshold can report 4,320 Bitcoin sitting behind 4,185 tBTC, and both figures can be accurate without guaranteeing that a holder can sell hundreds of tokens at anything close to face value.
The reserves answer whether the Bitcoin exists.
The redemption mechanism answers whether users can claim it.
And the liquidation market answers what that claim is worth when someone urgently needs to turn it into money.
For tBTC, those three answers were supposed to stay closely connected.
The pause shows what happens when they don’t.
The Bitcoin may be fully backed. The exit still has a price.
