Tracing the silent logic where value meets code. The data suggests FTX has returned $10 billion to its creditors over five rounds. The most recent slice is $900 million. On the surface, this is a recovery headline. But trace the actual path of that value and you find no blockchain transaction, no smart contract execution. You find a court order, a centralized trust, and a manual reconciliation process that would make any DeFi protocol blush. The anomaly is not the amount. It is the method.
FTX collapsed in November 2022. At that point, the exchange was a black box. User balances were entries in a centralized database. The Recovery Trust, led by CEO John J. Ray III, has been unwinding that box ever since. The fifth round of distributions, announced in early 2025, brings the total returned to $10 billion. But these funds are not being released via an audited smart contract. They are being routed through traditional banking rails, subject to identity verification, legal delays, and the trust of a single administrator.
Context: The FTX bankruptcy estate has been selling assets, liquidating positions, and resolving legal disputes to fund these distributions. The process is entirely off-chain. Creditors must file proofs of claim, undergo KYC, and wait for their share. This is not how crypto was supposed to work. The industry promises trustless, automatic settlement. FTX’s payout is a reminder that when the code fails, the law takes over.
Now the core analysis. Let’s examine the numbers and the mechanics.
First, the recovery rate. According to court filings, total claims against FTX are estimated around $16 billion. The $10 billion distributed so far represents roughly 62.5% recovery for allowed claims. But this number is misleading. Many creditors sold their claims to arbitrage funds for 30-50% of face value immediately after the bankruptcy. Those funds are now receiving the full $10 billion payout. The original depositors, who held onto their claims, have waited over two years. If you factor in the opportunity cost of that time—say a risk-free rate of 4% per year—the effective recovery drops to about 58%. The ones who sold early did better in absolute terms because they avoided that delay. This is a tax on patience.
Second, the distribution bottleneck. Each round requires the trust to verify creditor identities, determine payment amounts, and execute transactions. This is done manually or through semi-automated systems. Contrast this with a DeFi liquidation mechanism. In 2020, I audited MakerDAO’s CDP system. When a position falls below the collateral ratio, a keeper automatically initiates a liquidation auction. The process takes minutes, not months. The FTX trust has to deal with legal complexities—disputed claims, jurisdictional issues, tax liabilities. The result is inefficiency. The $900 million in this round took months to prepare.
Third, the asset composition. The trust holds a mix of crypto assets and cash from asset sales. To pay creditors in fiat or stablecoins, it must sell large amounts of SOL, BTC, and other holdings. That selling pressure has been absorbed by the market over time, but it distorts prices. In my simulation work during the Terra collapse, I showed that forced selling of illiquid assets by a single entity can create 30-50% downward pressure. FTX’s gradual liquidation has been relatively orderly, but it still represents a non-market signal.
Fourth, the cost of legal overhead. The trust has paid over $300 million in professional fees since 2022. That includes lawyers, financial advisors, and consultants. That’s 3% of the distributed amount. In a smart contract-based distribution, the operational cost would be gas fees and maybe a few thousand dollars for a security audit. The legal overhead is a deadweight loss that reduces the total sum creditors actually receive.
Now, the contrarian angle. The mainstream narrative is that this distribution is good news—creditors are getting money back, the system is healing. But look closer. This process demonstrates the fundamental failure of the centralized exchange model. FTX was supposed to be a trusted custodian. It had audits, proof-of-reserves claims, and a reputable founder. None of that protected users. The recovery is happening because of the US legal system, not because of the crypto infrastructure. If FTX had been operating in a jurisdiction with weaker legal recourse, creditors would have seen far less.
Dissecting the corpse of a failed standard. FTX was a standard exponent of the exchange model: trust the brand, trust the founders. The corpse we are dissecting today shows that standard was built on sand. The $10 billion payout is not a validation of the exchange model; it is a crash test that the model barely survived.
The real winners here are the arbitrage funds and law firms. They understood the incentives. They bought claims at a discount and spent money on legal fees to maximize their share. The original users who deposited in good faith? They received less than they would have if they had held their assets in an audited smart contract.
Finally, the takeaway. The next generation of exchanges will need to embed automated liquidation and distribution mechanisms. Proof-of-reserves must be real-time and cryptographic, not quarterly PDFs. If an exchange fails, the code should automatically split remaining assets among verified creditors. The FTX trust shows that manual distribution is expensive, slow, and opaque. The industry should demand better.
I do not trust the doc; I trust the trace. The trace of this $10 billion is muddied by legal delays and centralized decision-making. The next bear market will test whether exchanges have learned from this failure. If they haven’t, we will see the same autopsy repeated. The logic is silent, but it is there: value should be returned by code, not by court order.

