Drift recovery pool: 0.0104 USDT per DFX

The Drift recovery pool ledger classifies and signs every entry, and on October 7, 2026, across two readings separated by six hours, it contains not a single Tether contribution event. The protocol dashboard types each entry into five classes with literal labels —"seed-capital", "exchange-revenue", "tether-contribution", "other-contribution", and "redemption"— and under the third, there are zero records: the class is declared in the dashboard code and empty in the data. What does exist are 3,114,379.26 USDT of protocol seed deposited on September 30, 2026, two entries of 1 USDT, nine fee deposits totaling 44.849866 USDT, and 77 redemptions at 14:09 UTC. With that balance, each recovery token (DFX: the token that the Drift Foundation guide describes as issued one for every dollar of verified loss in the April incident) is exchanged for 0.010417 USDT, 1.04% of what that token claims, and the secondary market pays 0.0355 for it, a third less than four days prior. The 127.5 million USDT that Tether announced in April are written as a "paired deployment" that Drift's own documentation sizes against the previous quarter's revenue, and that revenue has left an average of 2.02 USDT per day in the pool across the eight full daily deposits from September 30 to October 7, 2026. This article measures the complete amortization mechanism: what goes in, through which channel, with what arithmetic ceiling, and what can be verified within a month.

This piece does not recount the April 1, 2026 attack nor discuss its attribution. That is covered in the incident report and in the account of the 577 million dollars that Lazarus took from DeFi in eighteen days. What follows is the economics of the recovery: a debt token, a USDT vault, and a revenue-indexed line of credit.

The exchange generating those revenues is called Velocity: it is Drift's own platform, rebranded in July 2026, a perpetuals market —futures contracts with no expiration date— settled in USDT on Solana. Its reopening has two dates from its own sources. On September 17, 2026, it published the OtterSec audit on its blog and opened the code, with 154 recorded issues and none classified as critical. On September 21 at 14:44 UTC, it announced on X that «the whitelist is off. Velocity Beta is now open to everyone», with wBTC, wETH, SOL, and USDT accepted as collateral in a single account and a 2 basis point fee for a limited time. The fees charged by Velocity are the only channel depositing into the pool every day, which is why they appear in all the following accounts. Its first entry in the pool's ledger is from September 29, 2026, the day the co-founder describes as the relaunch in the Solana Compass coverage.

How does the Drift recovery pool convert Velocity fees into USDT per token?

The Drift Foundation guide published on October 1, 2026, defines the mechanics in a literal line: "Redemption Amount = Recovery Pool Balance / Outstanding DFX Supply." The redemption amount is a quotient between two quantities that can be read directly on Solana: the USDT balance of the pool vault and the DFX supply remaining in circulation. There is no oracle, no auction, and no term discount.

The two readings from October 7, 2026, yield that ratio. In the first, at 14:09 UTC and in Solana slot 454,245,607, the vault held 3,111,695.765725 USDT and the circulating supply was 298,700,907.401891 DFX: the ratio is 0.010417429906 USDT per DFX. Four days earlier, on October 3, 2026, at 20:00 UTC, the vault held 3,112,017.956066 USDT and the supply was 298,732,555.716660 DFX, and the ratio was 0.010417404787.

The guide states that «a redemption leaves the Redemption Amount unchanged»: a redemption burns DFX and withdraws USDT in the same proportion, so it does not move the rate. This is verified within October 7, 2026 itself. In the second reading, at 20:11 UTC and in slot 454,326,028 —6 hours and 2 minutes after the first— the vault held 3,111,689.154319 USDT and the circulating supply was 298,700,272.753295 DFX. Between the two readings, there were two redemptions that withdrew 6.611406 USDT and burned 634.648596 DFX, and the ratio resulted in 0.010417429906 in both: identical up to the twelfth decimal place, with the first difference at the fifteenth, 4.23 × 10⁻¹⁵.

That difference in the fifteenth digit is the rounding, and in this interval it appears in its minimum unit. The 634.648596 DFX burned were worth 6.611407 USDT at the 14:09 rate and the ledger paid 6.611406: the guide rounds down each payment to 0.000001 USDT and what is not paid remains in the vault. With zero deposits in between, the identity of the mechanism remains in that residue: (20:11 rate − 14:09 rate) × 20:11 circulating supply = 0.0000013 USDT, one micro-USDT at the six decimals with which the contract pays, distributed between two redemptions. The same result was measured on October 3, 2026, with a single redemption in between —40.024999 USDT and 3,842.127745 DFX between the 13:54 and 20:00 readings, separated by 6 hours and 6 minutes, with the balance dropping from 3,112,057.981065 to 3,112,017.956066 USDT— and the rate also came out identical to twelve decimals.

The other side of the same identity is seen over four days, and there the rate does increase: 2.5119 × 10⁻⁸ between October 3 and 7, 2026, with 17 redemptions and four fee deposits in between. The equality is the same as before with the deposits term included, and it also closes to the twelfth decimal place: (rate at 14:09 on the 7th − rate on the 3rd) × circulating supply at 14:09 = deposits + (DFX burned × rate on the 3rd − USDT paid), and both sides yield 7.502964786772 USDT. The four deposits from October 4 to 7 total 7.467653 USDT, so the remaining term is 0.035312 USDT: the same downward rounding, accumulated over 17 payments instead of two. The 17 redemptions burned 31,648.314769 DFX, which were worth 329.693306 USDT at the October 3 rate and 329.694101 at the October 7 rate, compared to the 329.657994 paid by the ledger, and the distance from the first of those two valuations is that 0.035312 USDT term. The pair of readings with only redemptions leaves the rate steady, and the four-day pair raises it by exactly what entered through the channel, minus what is retained by rounding.

The practical consequence is that the redemption rate only increases when new money enters. And there is only one channel entering every day: fees from Velocity, Drift's perpetual exchange on Solana. The guide describes a three-part distribution —15% to the Insurance Fund (the protocol's guarantee fund, which covers losses from liquidations that run out of sufficient margin), 15% to vAMM capital (the exchange's own virtual automated market maker, which quotes against a curve instead of a third-party order book), and the remaining 70% as "Net Protocol Revenue"— and sets what fraction of that 70% goes to the pool based on marginal tiers of the net revenue from the last 24 hours:

• First 30,000 USDT of daily net income: 60% to the pool.
• Between 30,000 and 100,000 USDT: 70% to the pool.
• Above 100,000 USDT: 90% to the pool.
• Each tier applies only to income within its range, and the deposit is executed once daily at 00:00 UTC as an on-chain transaction.

The dashboard API returns exactly that first tranche. The daily cutoff read on October 7, 2026, corresponds to the closed session of October 6 and was deposited into the vault at 00:02:00 UTC on the 7th, two minutes late compared to the seven previous cutoffs, which all entered within the first minute of the day: the "totalFees" field marks 1.238352 USDT. From there, "netProtocolRevenue" retains 0.866847 and "poolShare" —the portion entering the vault— receives 0.520108; "treasuryShare" collects 0.346739 and the "insuranceFund" and "vammCapital" fields receive 0.185752 USDT each. The "revenueSplit.recoveryPool" field is at 0.599999: the 60% tranche. Out of 1.238352 USDT in total fees for a single day, 0.520108 reached the pool.

This distribution determines how much of the revenue generated by the exchange can reach those affected. For every USDT of net trading commissions, 0.30 remains in the Insurance Fund and the vAMM capital, excluded from the recovery; of the remaining 0.70, the current bracket sends 60% to the pool. The pool therefore receives 0.42 USDT for every USDT of net commissions as long as daily income does not exceed 30,000 USDT, reaching up to 0.63 USDT in the high band, above 100,000 USDT daily. The proportion measured in the October 7, 2026 snapshot is exactly that: 0.520108 out of 1.238352 is 42.0%, the same percentage yielded by the October 3 snapshot with commissions eleven times higher. The Velocity blog describes the same distribution from the exchange side on September 17, 2026 —the commission for each perpetual is divided between protocol liquidity, the Insurance Fund, and the protocol itself— and refers to the same schedule: «a majority share of net protocol revenue goes to the recovery pool on the tiered schedule outlined in the DFX update».

Two channel conditions changed between the May plan and the October 2026 guidance, and the shift lies in the two primary sources. The May 5, 2026 plan stated that "redemption opens once the Recovery Fund exceeds $5M" and described the contribution as quarterly and in qualitative terms: "each quarter, a substantial portion of the exchange's net revenue flows directly into the recovery pool," without percentages. The October 1, 2026 guidance opened redemptions with the pool at approximately 3.11 million USDT, below that announced threshold, and replaced the quarterly contribution with a daily deposit at 00:00 UTC with the 60, 70, and 90% tiers explicitly written. The channel cap remains the same in both documents: deposits stop when the pool has accumulated the total verified losses. The May 5, 2026 plan writes that amount in dollars, "$295,426,725.97," and the October 1 guidance counts it in the pool's currency—"one DFX per verified USDT of loss"—so in this piece it is listed as 295,426,725.97 USDT.

The dashboard publishes two different time windows across two different endpoints —an endpoint being each of the API addresses that returns a specific data type— and each one measures something different. The fees endpoint with the "range=24h" parameter always refers to the daily cutoff at 00:00 UTC, a period already closed and deposited: on October 7, it returned the data for the 6th. The markets endpoint moves with the clock: the perpetuals volume for the last 24 hours stood at 126,681.47 USDT at 14:09 UTC on October 7, distributed across five perpetual markets, with 4,233,626 USDT deposited in the exchange; on October 3, in two readings on the same day, that same field went from 30,065.89 to 10,410.41 USDT. A fee-to-volume ratio constructed using figures from both endpoints would be measuring two different days.

Three different magnitudes describe the same recovery. The redemption rate is measured in USDT per token: 0.010417 USDT per DFX. The percentage recovered by the holder is what that rate represents relative to the dollar that the token claims: 1.0417%. And the pool coverage of the verified loss is the vault balance against the 295,426,725.97 USDT that Drift published in May 2026: 1.0533%. Both percentages stem from the same balance with denominators that differ by 4,074,085.03 units, which explains why 1.0417% and 1.0533% coexist in this piece.

What does the recovery pool ledger record regarding the 127.5 million Tether?

The recovery panel exposes its ledger via API, and each entry includes the type, amount, resulting pool balance in the "recoveryPoolValue" field, and the signature of the supporting transaction. The complete inventory of entries as of October 7, 2026, consists of twelve records:

• «other-contribution», 1.000000 USDT on September 29, 2026, the first entry in the pool's history.
• «seed-capital», 1.000000 USDT on the same day and in the same transaction.
• «exchange-revenue», 28.700732 USDT on September 29, leaving the pool at 30.700732 USDT.
• «seed-capital», 3,114,379.255730 USDT on September 30, 2026: the actual seed. The October 1 guide only provides the balance («the Recovery Pool holds about 3.1M USDT today»); the source of these funds is detailed in the May 5, 2026 plan, which describes the pool as seeded with «the protocol's remaining assets … approximately $3.8M», about 0.69 million more than what was ultimately deposited.
• Eight more daily «exchange-revenue» deposits, seven of them within the first minute of the UTC day: 0.505894 USDT on September 30; 1.554790 on October 1; 0.659123 on October 2; 5.961674 on October 3; 0.678060 on October 4; 2.302913 on October 5; 3.966572 on October 6 and 0.520108 on October 7, 2026.
• Zero entries of the «tether-contribution» class.

The nine fee deposits total 44.849866 USDT, which is exactly the value of the accumulated "poolShare" field returned by the API with the "range=all" parameter. Eight of the nine are full daily cuts and yield an average of 2.018642 USDT per day, with a maximum of 5.961674 USDT on October 3, 2026, and a minimum of 0.505894 on September 30. The ninth is not included in that average: the 28.700732 USDT entry on September 29 was recorded at 08:25 UTC, outside the daily cut schedule, and carries fees accumulated before the pool had a ledger. Including it raises the average of the nine entries to 4.983318 USDT per day, a figure that mixes the channel's startup with its subsequent pace. The dashboard series begins at the end of September —with the "range=all" parameter returning the same as "7d" and "30d"— so the Velocity private beta remains outside these figures. The pool ledger measures how much the vault has received, which is the amount that divides the redemption rate.

On the outflows side, 77 redemptions totaling 2,730.339871 USDT, an average of 35.46 USDT per redemption and a range from 0.000108 to 1,115.600809 USDT. The figure is verified without the API: the sum of all recorded inflows is 3,114,426.105596 USDT and the vault held 3,111,695.765725 USDT in the October 7 reading. The difference —2,730.339871 USDT— matches the total redemptions from the ledger to the sixth decimal place. Two independent sources, the dashboard ledger and the vault balance on Solana, reconcile the same account. And there is a second reconciliation that does not involve the balance: the sum of the nine "exchange-revenue" class entries is 44.849866 USDT, the same number as the accumulated "poolShare" from the fees endpoint, two different paths from the dashboard to the same figure.

The complete list of vault signatures since its creation contains 90 entries as of 14:09 UTC, and all of them are accounted for. Eleven correspond to the twelve entry records in the ledger, because two of them share a transaction, and 77 correspond to redemptions. The remaining two predate any deposit and have been decoded: the one from August 21, 2026, is the «InitializeConfig» instruction of the redemption program, which creates the pool's USDT account with a zero balance, and the one from August 25 is an SPL transfer of 1.000000 USDT that leaves the vault with 1 USDT. That dollar is accounted for in the ledger, and the account only balances in one way: of the two 1 USDT entries that the panel dates September 29 under the signature of the first «Contribute» instruction, one records money entering that day and the other records the dollar that was already in the vault since August 25, five weeks earlier. With that allocation, the sum of the twelve entries minus the 77 redemptions yields the exact vault balance. Among those 90 signatures, there is no USDT movement outside of that inventory, and the last one is from October 7, 2026, at 07:19:05 UTC, 6 hours and 50 minutes before that reading. At 20:11 UTC the list has 92: the two new ones are the two redemptions from the interval, the last of which is from 16:19:32 UTC, and the ledger entries remain at twelve, without any «tether-contribution» class record.

Absence has the same support as any present entry. If Tether had disbursed anything to the pool, the entry would be typed with its own class, including the amount, resulting balance, and signature, just like the 0.520108 USDT in fees from the October 7, 2026, cutoff. The class exists in the panel code and is empty.

What is and what is not included in the Tether package with Drift?

The commitment is published by both parties and dated. What is stated in the primary source:

• Tether, April 16, 2026: "up to nearly $150 million" for the recovery plan, of which "up to $127.5 million from Tether," and a pacing condition: "Capital support will be introduced progressively and aligned with performance."
• Drift, April 16, 2026: the package structure is "a $100 million revenue-linked credit facility, an ecosystem grant, and loans to market makers," and "other partners are proposed to contribute $20 million."
• Drift, May 5, 2026: the sizing rule for the Tether tranche: the deployment "will be based on the exchange's prior-quarter revenue." The same plan rewrites the partner contribution as "Strategic partners have committed up to $20M to support user recovery."
• Drift, June 3, 2026: the update reporting the relaunch mentions the package only qualitatively —"the strategic support package from Tether and other partners"— and does not include any figures: neither the 5 million threshold, nor the 20 million tranche, nor the 127.5 ceiling.
• Drift Foundation, October 1, 2026: in the claims guide, the commitment appears as "Tether Matched Deployment: Tether has committed up to 127.5M USDT to support relaunch and user recovery," and the partners' commitment as "Strategic partners have committed up to 20M USDT to support user recovery."

The partner tranche underwent wording changes in the documents that do mention it, and the shift always moves in the same direction. On April 16, 2026, they were "other partners" who "are proposed to contribute $20 million": a proposal, in dollars and without the "up to" phrasing. On May 5, they became "strategic partners" who "have committed up to $20M": a formal commitment, with the amount converted into a ceiling. On June 3, the tranche disappears from the text. On October 1, the guide reinstates the commitment and changes the unit to USDT. These are three versions of the same commitment and one intermediate document that does not repeat it, and the rewriting is of the same nature as that of the 5 million threshold: a condition stated in one document and absent from the next.

And what is not included in any of those documents or on the dashboard:

• The breakdown of the 127.5 million between the credit line, ecosystem grant, and loans to market makers.
• The deployment schedule, the interest rate for the credit portion, and the seniority of repayment relative to DFX holders.
• Which exact metric is paired and with what multiplier: "matched deployment" does not specify whether it matches gross revenue, net protocol revenue, or the deposit to the pool, nor in what proportion.
• Whether any tranche has already been disbursed outside the pool, for example as a loan to a market maker. The pool ledger only tracks what enters the vault.

The sizing rule is what converts the headline into a function of another figure. If a quarter's deployment is calculated using the previous quarter's revenue, what the dashboard API reveals from the prior period is 74.749781 USDT of accumulated "netProtocolRevenue" and 106.785401 USDT of "totalFees." A pairing calculated on that basis would have three digits. And the guide itself limits the destination: the 127.5 million are "to support relaunch and user recovery," two distinct things, of which only the second ends up in the vault that divides the redemption rate.

How far does the arithmetic ceiling of Drift's recovery reach?

The guide lists four growth channels for the pool: net protocol revenue, paired Tether deployment, up to 20 million USDT from strategic partners, and stolen funds recovered through freezing, bounties, or legal action. Combined, these represent the maximum that can exist. It is a ceiling, not a forecast: it assumes that every promise is fulfilled in its entirety and that everything ends up in the vault.

The ceiling with the two capital promises is 50.98 % of the verified loss and 50.29 % of the issued token supply; adding the 9.2 million dollars that the Drift Foundation itself considers frozen on September 30, 2026, 54.10 % and 53.36 %. The denominator of the actual distribution is set by the guide and is a third party: «Redemption Amount = Recovery Pool Balance / Outstanding DFX Supply», the circulating supply, which on October 7, 2026, was 298,700,907.401891 DFX and can only decrease. The percentages of the issued supply are therefore a floor: with the circulating supply of that day, that same ceiling yields 50.42 % instead of 50.29 %, and the gap between the two figures will grow with each burn. The percentages of verified loss and tokens are not subtracted from each other either: one divides by dollars recognized in May 2026 and the other by tokens, with a base 4,074,085.03 units larger.

Between the 1.05% of the verified loss covered by the pool on October 7, 2026, and the 50.98% of that same loss provided by the cap, lies the channel that actually operates daily, and its scale is best measured by two calculations. First: from October 7, 2026, until the closing of the claims window, 451 calendar days remain; at the observed average of 2.018642 USDT per day, the fee channel would contribute 910.41 USDT in total, or 0.0003% of the verified loss. Covering only the partners' portion—the 20 million—through this method would require 44,345.90 USDT per day for 451 days, which is 21,968 times the average of the eight data points. The second calculation looks at the rate: increasing it by a single thousandth of a USDT per token—from 0.010417 to 0.011417 USDT—requires an additional 298,700.91 USDT in the vault; at the best observed day of 5.961674 USDT on October 3, this would take 50,104 days, or approximately 137 years.

What is the Drift loss and why are six different figures circulating?

Any recovery percentage depends on the denominator, and for the Drift incident, six different figures are circulating. All six measure different things: three are April 2026 estimates of the stolen value, one is the verified loss against the position snapshot —the photo of all accounts frozen at a specific time—, one is circulating without a cited source, and the sixth is the issued token supply presented as the loss.

Two of those rows were written in the same week. The 299.5 million travel with two labels that do not mean the same thing —“verified losses” on October 1, “token supply” on the 3rd— and the second one is correct. The 311 million come from a phrase that repeats the same three digits in the two amounts it provides, with a factor of one hundred between them. Searching for the literal text in the four Drift documents and in the Tether note, the only “311” is the 3,318,311.45 dollars of syrupUSDC in the stolen assets table from April 16, 2026, which measures a specific item of the loot.

The 285 million used in this site's April pieces represents the valuation circulating at that time, four weeks before the protocol published the count against the snapshot on May 5, 2026. Anyone comparing recovery percentages across different coverages is dividing by denominators that differ by up to 31 million dollars. The figures in this article include their basis: 295,426,725.97 USDT when the denominator is the verified loss, 299,500,810.998 DFX when it is the issued supply, and 298,700,907.401891 DFX when it is the circulating supply as of October 7, 2026.

Until when can those affected by the Drift incident claim their DFX?

The deadline is set in the primary source with an exact time. The Drift Foundation guide states it as follows: «Claim window: closes at 00:00 UTC on 1 January 2028». October 1, 2026, is the claim opening date: the day that same guide was published. What expires on January 1, 2028, is the right to claim: any DFX not claimed when the window closes will be permanently burned.

Redemption, on the other hand, has no expiration date. It is deactivated by a threshold, not a calendar: «Deposits stop once the pool has received the full amount of verified losses in total». The pool stops receiving deposits once it accumulates the total amount of verified losses, and the redemption rate, by design, never decreases.

What the guide requires for claiming is also locked and verifiable before attempting: the wallet that controlled the Drift account on April 1, 2026 —any other address shows nothing—, a small amount of SOL for the network gas fee, and the verification of the allocation against the loss snapshot via a Merkle proof (a cryptographic check that a specific entry belongs to a closed list, without needing to publish the entire list), which returns the exact number of tokens. The Insurance Fund claim is a separate process with its own terms and is not DFX. Once the tokens are claimed, the guide describes three possible destinations —redeem, sell, or hold—, and details that the DFX burn and the USDT payment occur in a single transaction, that amounts are rounded down to 0.000001 USDT, and that redemptions are final.

Regarding a specific claim, the arithmetic is straightforward, and it is what allows the affected party to put the three figures in this piece on the same scale. Anyone with 10,000 USDT of verified loss recognized holds 10,000 DFX. At the rate on October 7, 2026, the pool pays them 104.17 USDT for the total. At the price of the secondary pair with the most depth that same day, 355.00 dollars; four days earlier it was 531.20. And if the 127.5 million Tether and the 20 million from partners arrived in full, 50.29 % of the issued supply, 5,029 USDT. The three amounts measure different things: the first is the money that was in the vault on October 7, 2026, the second is what a third party was paying for that right on October 7, 2026, and the third is the sum of promises, two of which have no accounting entry.

How many DFX are there in excess of the verified loss and what is known about the 537,804 tokens burned without payment?

Two facts about the token have no officially published explanation, and both are resolved on-chain down to the token without any document stating why they occurred.

The first is a base mismatch. The issued supply is 299,500,810.998352 DFX, minted in a single transaction on August 4, 2026, with the issuance authority revoked 21 seconds later; the verified loss that Drift published in May 2026 is 295,426,725.97 USDT. The guide describes the supply as "fixed at one DFX per verified USDT of loss," but there are 4,074,085.03 more tokens than dollars of verified loss, a 1.38% excess. No Drift document explains the difference. Its effect on the recovery ceiling is limited: it moves it between 50.29% and 50.98%, depending on whether it is divided by issued tokens or by dollars of verified loss.

The second consists of unpaid burns. Out of the total 799,903.596461 DFX destroyed since minting —0.2671% of the supply— 262,100.007515 correspond to redemptions paid by the pool. The other 537,803.588946 DFX were burned without a single USDT leaving the vault: 59.678821 on September 19, 2026, and 537,743.910125 on September 29, two days before claims opened, across three transactions that same morning, the largest of which was 514,109.379254 DFX from a single token account. The first redemption in the ledger occurred later, on September 30, 2026, for 1 token and 0.010417 USDT. Neither the excess supply nor the burns are explained in the four Drift documents, in the Drift Foundation thread from September 30, 2026, nor in the coverage by The Block on October 2 and Solana Compass on October 1, and there is no documentary basis to assign them a cause.

The arithmetic, on the other hand, adds up: 299,500,810.998352 minted, minus 537,803.588946 burned without payment, minus 262,100.007515 burned via redemption, results in the 298,700,907.401891 DFX recorded on October 7, 2026. Verification from the other side reaches the same conclusion with a known residual: the 2,730.339871 USDT paid divided by the daily rate yields 262,093.423774 DFX, which is 6.58 tokens less than the subtraction —0.0025%—, consistent with the downward rounding of each redemption and the slightly lower rate during the initial days.

The effect of these burns on the remaining holders is indeed documented, as it is the core mechanism. The guide illustrates this with an example: if 10% of the supply is redeemed, each remaining DFX receives approximately 11% more from each subsequent deposit. The 537,803.59 tokens burned without payment represent 0.18% of the supply, thereby increasing the share of each future deposit allocated to the surviving tokens by that same proportion. Based on daily deposits of 2.018642 USDT, this additional 0.18% amounts to 0.0036 USDT.

What is the secondary market paying for a DFX after falling by a third in four days?

DFX is a standard SPL token —SPL is the Solana token standard, the equivalent of Ethereum's ERC-20— and is traded on the network's secondary markets. The following figures come from DexScreener, an aggregator that publishes the price and reported liquidity of each pair, as read on October 7, 2026, at 14:09 UTC; these are not direct readings of the reserves of each pool. The deepest pair, DFX/USDT on Meteora, with $225,337.02 in reported liquidity and $17,322.22 in 24-hour volume, was trading at 0.03550 USDT. On October 3, 2026, at 20:00 UTC, that same pair stood at 0.05312 with $272,516 in liquidity: a 33.17% price drop in four days.

The order of the pairs also changed: on October 7, 2026, the aggregator lists four pairs, compared to six on October 3. The other DFX/USDT pair, on Raydium, went from 170,604 to 380.64 dollars in reported liquidity —a 99.8% decrease— with 9.35 dollars in daily volume and a price of 0.03711 USDT, ceasing to be a usable benchmark. With that drop, the second place by depth moves to DFX/SOL on Meteora, the only pair that rose: from 25,771 to 33,802.85 dollars, trading at 0.03553. The third is DFX/SOL on Raydium, with 1,041.73 dollars in liquidity and a price of 0.03641, and the fourth is that DFX/USDT on Raydium which remained at 380.64. No public source documents what moved that liquidity.

Those 3.55 cents are 3.4 times the pool redemption rate; on October 3, it was 5.1. Regarding the whole: the circulating supply at the price of the deepest pair is worth 10.60 million dollars, 3.59% of the verified loss, compared to an arithmetic ceiling of 50.98% of that same loss with the two capital promises included. That 50.98% ceiling is 14.20 times the 3.59% capitalized by the market: at 3.55 cents per token, the market prices in 7.0% of the promised package, compared to 10.5% on October 3. The pool rate rose during those four days and the market price fell by a third, so the premium the market pays over the redemption has narrowed from both sides at once.

There is a measured precedent for how long an amortization of this type takes when the only real channel is protocol revenue. In the remediation of the cbETH incident at Moonwell, the protocol had distributed 12.29 % of the obligation by June 4, 2026, nearly four months after the incident, and the affected parties themselves estimated about five years at the protocol's pace. Drift is at 1.05 % of its verified loss six days after opening claims, with a daily channel of 2.018642 USDT and announced external commitments totaling 47.4 times the vault balance that have no entry in the ledger.

The other useful precedent is that of pursued stolen funds, and here the source is primary. The Drift Foundation published a thread on September 30, 2026, at 14:21 UTC titled "Drift Recovery Update" —the only internal link that the October 1 guide includes in its "Recovered Funds" section— and this is where the Foundation quantifies the pursuit. It estimates the amount taken on April 1 at "~$295.4M". The money was bridged to Ethereum as "~130,259 ETH across four wallets", and of those four wallets, three have not moved and hold 107,165 ETH, or 82.27% of the bridged amount; on July 23, one of them sent approximately 23,094 ETH to Tornado Cash, the remaining 17.73%. Regarding the frozen assets, it provides a figure and a cause: "About $9.2M of stolen funds has been frozen so far, after the attacker moved them through Tornado Cash in August". Those 9.2 million dollars represent 3.11% of the 295.4 million that the same thread acknowledges as taken. Mandiant, zeroShadow, and SEAL 911 are involved in the investigation, and the public 10% bounty program for recovered funds, which the May 5, 2026 plan already linked to, is running, according to the thread, in collaboration with Bybit. As of October 7, 2026, this fourth growth path for the pool, "Recovered Funds", has no entries in the ledger.

Of the two cases with measured figures on this site, what is recovered depends on which issuer controls the asset. In the Bitget hack, issuers had frozen 339,231.54 dollars of the 387.5 million transferred to attacker addresses, or 0.09%, as of September 29, 2026. In the KelpDAO bridge hack, the 116,500 unbacked rsETH minted were sent to Aave as collateral, the loan obtained against them went through Tornado Cash, and Kelp did not publish a loss allocation plan. Here, the 107,165 ETH is native ether, which no issuer can freeze, and the pool pays out in USDT, the stablecoin from the same company promising the 127.5 million: Tether has frozen 3.29 billion dollars across 7,268 wallets and that capability applies to a recovery pool vault just like any other account. The other path that has worked at scale is for other protocols to provide the funds, such as the 300 million dollars that seven protocols raised for Aave in April 2026; here, the external capital has had a name, figure, and date since April 16, 2026, and still no entry in the vault.

What should be measured on November 3, 2026, to determine if the pool has changed regimes?

The question has a numerical answer and can be verified with two public queries. With only commissions entering at the observed rate —an average of 2.018642 USDT per day, with a maximum of 5.961674— the redemption rate for November 3, 2026, should remain between 0.01041 and 0.01043 USDT per DFX: the twenty-seven deposits remaining until that date, all at the best observed day, would leave it at 0.010418. On October 7, 2026, the rate stands at 0.010417429906 USDT per DFX, within that range. Above that, there are two levels, and they do not mean the same thing. A reading between 0.01043 and 0.0110 is already outside the observed pace: reaching 0.01043 requires an additional 3,754.70 USDT in the vault, which represents 630 days at the best recorded cut and 1,860 at the average; therefore, in twenty-seven days, it can only be reached by funds unrelated to the commissions of those days. A reading above 0.0110 requires an additional 174,014.22 USDT, or 86,204 days at the average rate: that level can only come from an external deposit. In both cases, the entry will be categorized in the ledger with its specific class and signature.

The second threshold measures whether the revenue channel can replace external capital, and it requires more than ten times the average daily fees that Drift recorded in the first quarter of 2026 to do so. Bringing the vault from 3.11 to 20 million USDT in the 85 days remaining until December 31, 2026, requires 198,686 USDT daily to the pool. Applying the published tiers —60% of the first 30,000, 70% of the next band, and 90% above 100,000— that amounts to 246,318 USDT in daily net protocol revenue and 351,882 USDT in daily total fees. This is 10.6 times Drift's daily fee average in the first quarter of 2026 according to DefiLlama (33,073 dollars per day, with the last day of the series on April 1, 2026) and 24,790 times the largest daily cut recorded by the dashboard, the 14.194462 USDT from October 2nd deposited on the 3rd. In practice, the 20 million threshold can only be crossed by external capital.

One thing remains that, as of October 7, 2026, should not be treated as verified: the token price, which comes from an aggregator rather than a direct reading of each pool's reserves. The reopening schedule is indeed found in Velocity's own sources and includes specific dates—September 17, 2026, for the audit and open-sourcing of the code, and September 21, 2026, at 14:44 UTC for the open beta—while the data from September 29 reflects the first entry in the pool's ledger; "relaunch" is the word the co-founder uses to describe that day in Solana Compass coverage, though no Drift or Velocity document employs it. The numerical core can be reproduced by anyone with two public calls: the vault balance (3,111,695.765725 USDT), the circulating supply (298,700,907.401891 DFX), the ratio between the two (0.010417429906), and the five types of ledger entries, with zero records under "tether-contribution" in the reading from October 7, 2026, at 14:09 UTC and the same zero in the 20:11 reading.

Sources and links: Drift Foundation — DFX claims and redemptions dashboard (source for typed ledger and pool balance) · Drift Foundation — DFX Claims and Redemptions are now open (Oct-1-2026) · Drift Foundation — Drift Recovery Update, X thread from Sep-30-2026 14:21 UTC (bridged funds, frozen funds, and bounty program) · Drift Foundation — DFX dashboard recovery pool page · Drift — Drift Recovery Update, June 3, 2026 · Drift — Recovery Plan for Affected Users (May-5-2026) · Drift — Incident Recovery Update (Apr-16-2026) · Tether — Tether leads support to the 150M Drift recovery plan (Apr-16-2026) · The Block — Drift opens exploit recovery claims (Oct-2-2026) · Solana Compass — DFX claims and redemptions near one cent per dollar · Velocity Exchange — Velocity publishes its OtterSec audit and opens its source code (Sep-17-2026) · Velocity Exchange — open beta without whitelist, X announcement from Sep-21-2026 14:44 UTC · DefiLlama — Drift daily fees · DexScreener — DFX pairs on Solana