Evidence in the SBF Trial Sets Loss Amount Due to Fraudulent at Over $10 Billion

A. The Evidence at Trial Proved That the Loss Amounts Is at Least $10 Billion

The defendant takes issue with the loss amount as calculated in the PSR. His arguments, however, are entirely meritless and fail to address the facts of this case or the governing law. The evidence at trial proved that the losses to FTX customers are conservatively estimated to be at least $8 billion, the losses to FTX investors are at least $1.7 billion, and the losses to Alameda lenders are at least $1.3 billion. The total actual loss amount from the defendant’s interlocking fraudulent schemes, therefore, is at least in excess of $10 billion.

1. Applicable Law

“Loss” is defined in the Guidelines as “the greater of actual loss or intended loss.” U.S.S.G. § 2B1.1, cmt. 3(A). “Actual loss” is defined as “the reasonably foreseeable pecuniary harm that resulted from the offense.” Id., cmt. 3(A)(i). And “reasonably foreseeable pecuniary harm” is defined as “pecuniary harm”—that is, “harm that is monetary or that otherwise is readily measurable in money”—that “the defendant knew or, under the circumstances, reasonably should have known, was a potential result of the offense.” Id., cmt. 3(A)(iii), (iv)). “Intended loss” is defined as “the pecuniary harm that the defendant purposely sought to inflict,” and includes any “intended pecuniary harm that would have been impossible or unlikely to occur.” Id., cmt. 3(A)(ii). The “Guidelines do not require that the sentencing court calculate the amount of loss with certainty or precision.” United States v. Bryant, 128 F.3d 74, 75-76 (2d Cir. 1997). Instead, a court “need only make a reasonable estimate of the loss,” given the “available information.” U.S.S.G. § 2B1.1, cmt. 3(c)

2. The Trial Evidence Regarding Loss Amount

Here, there are three categories of losses: (1) losses to FTX customers; (2) losses to FTX’s equity investors; and (3) losses to Alameda’s lenders.

First, the defendant’s fraud on FTX customers resulted in a loss to those customers in excess of $8 billion. This number is calculated by conservatively estimating the total shortfall on the FTX exchange, that is, the amount by which customer account balances exceeded the amount of assets that FTX had available to return to those customers when FTX declared bankruptcy on November 11, 2022. The trial evidence fully supports this amount, which, if anything, is a conservative estimate. For instance, Professor Easton testified that in June 2022, as much as $12.6 billion in customer funds had been misappropriated. (Tr. 1773; GX-1005). Wang testified that in September 2022, it was around $14 billion. (Tr. 450). Ellison also maintained a spreadsheet, which she shared with the defendant in September 2022, showing that Alameda had taken around $13.7 billion from FTX customers. (Tr. 813).

By November 1, 2022, according to Professor Easton, the amount of misappropriated customer funds had been reduced to approximately $9.2 billion. (Tr. 1773; GX-1005). The shortfall decreased in the final days of FTX because the defendant returned some of the money he had misappropriated back to FTX to cover customer withdrawals. The Guidelines provide that the loss amount “shall be reduced” by the amount of “money returned … to the victim before the offense was detected,” U.S.S.G. § 2B1.1, cmt. 3I(i).[3] However, even if the defendant is given credit for these pre-bankruptcy returns of funds, the evidence at trial shows that there was still a total shortfall of around $8 billion in customer funds when FTX collapsed. For example, Wang testified that there was still a $8 billion shortfall (Tr. 456-58), and the defendant himself sent a text message on November 7, 2022, in which he acknowledged that there was a shortfall of $8 billion in customer assets that FTX was unable to repay, even if he could liquidate certain Alameda assets (Tr. 2789). Accordingly, the trial evidence fully supports the PSR’s conclusions that customer losses amount to $8 billion.

Second, the evidence at trial clearly supports the PSR’s conclusion that the losses to investors amounted to $1.72 billion, which was the total amount of equity investments in FTX.com that the defendant fraudulently obtained. (GX-26 (listing the total amounts invested in the Series B, B-1, and C round investments in FTX.com)). None of these investors received any return on their investment while FTX was operating, and the equity in FTX is now worthless. Accordingly, the actual loss to investors is the full amount of their investment.

Third, the evidence at trial showed that Alameda obtained approximately $1.7 billion in new loans from May to November 2022, the time period in which the defendant conspired to provide lenders with false financial information to induce them to continue lending. (GX-1014). Of that amount, approximately $1.3 billion were outstanding at the time of bankruptcy. (Id.). Thus, even if the defendant were given credit for pre-bankruptcy returns of funds, the actual loss for the lender fraud is $1.3 billion.

3. The Defendant’s Assertion That the Loss Amount for FTX Customers Is Zero Is Meritless

The defendant does not seriously dispute any of these numbers or the evidence supporting them. Rather, he argues that the actual loss should be estimated at zero based on developments that have happened long after his fraud was exposed and his companies declared bankruptcy. (Def. Mem. at 17-21). The sole basis for the defendant’s argument on loss amount is a statement by counsel for the debtors in the FTX and Alameda bankruptcy proceeding that FTX customers and creditors “will eventually be paid in full.” See Transcript of Hearing at 18:11-25 (“Bankruptcy Transcript”), In re FTX Trading Ltd., et al., 22 -11068 (JTD) (Bankr. D. Del. Jan. 31, 2024), Dkt. 6908. The defendant’s arguments relying on that representation, however, are fatally flawed both on the law and the facts.

First, as a legal matter, any potential recovery to victims in the bankruptcy proceeding should not be used to reduce the loss amount for purposes of the Guidelines. Pursuant to U.S.S.G. § 2B1.1 cmt. 3(E), there are only two circumstances where it is appropriate for a district court to make “credits against loss”: (1) where there are funds returned to victims “before the offense was detected,” and (2) in certain cases “involving collateral pledged or otherwise provided by the defendant.” Neither of those circumstances applies here. As to the first circumstance, the potential recovery to victims from the bankruptcy did not happen before detection of the fraud. On the contrary, as of this date, nearly a year and a half after FTX declared bankruptcy and months since the defendant’s conviction at trial, his victims have received no recovery and there is no timeline for when any such payments will be made. As to the second circumstance, it plainly does not apply here, as this is not a fraudulent loan case in which the defendant pledged collateral. Rather, this case involved the misappropriation of funds by the defendant, secretly taking money from the FTX exchange and moving it to Alameda, his own investment company, after assuring his customers he would do no such thing. Customers were never pledged any collateral for their FTX deposits. Thus, the plain text of the Guidelines does not provide for any reduction in loss amount based on the possibility that victims will receive compensation in the bankruptcy.

The defendant’s position is contradicted by case law as well. The Second Circuit has held that a defendant’s “arguments that the loss calculation should be offset … by his company’s bankruptcy filings are unpersuasive,” because the Guidelines only allow for reduction of the loss amount for money returned “before the offense was detected.” United States v. Ware, 399 F. App’x 659, 662 (2d Cir. 2010) (quoting U.S.S.G. § 2B1.1 cmt. 3(E)(i)); see also United States v. Bergstein, 788 F. App’x 742, 747 (2d Cir. 2019) (where defendant had misappropriated funds invested in company, affirming district court’s decision not to reduce the loss amount by the funds subsequently recovered by the company). Thus, the law is clear that the loss amount should not be reduced by “funds received in bankruptcy after the conspiracy was uncovered, or to the potential value of a future bankruptcy payout.” United States v. Bryson, 101 F. Supp. 3d 147, 157 (D. Conn. 2015).[4]

In his submission, the sole case that the defendant relies on is United States v. Durham, 766 F.3d 672 (7th Cir. 2014), which affirmed a district court’s calculation of loss amount as the amount owed to investors less the amount recovered in bankruptcy. (Def. Mem. at 19-20). But the defendant’s reliance on Durham is flawed. In Durham, unlike here, the evidence at sentencing did not show exactly how much money the defendants had misappropriated, so the district court relied on the amounts invested by investors minus the amount recovered in bankruptcy as a reasonable proxy for that loss. Id. at 687. The Seventh Circuit merely affirmed the district court’s methodology as reasonable and did not suggest that as a general rule, loss amount should be reduced by the amount of money recovered in bankruptcy. Indeed, the issue was not even litigated in Durham, because the Guidelines range was not affected by whether or not bankruptcy recoveries were reduced from the loss amount. Id.[5]

The defendant’s argument is also contrary to common sense. In many criminal fraud cases, the combined efforts of the Government and the bankruptcy system are able to achieve substantial recoveries for victims. For instance, the Government recently announced that victims of the Madoff fraud scheme have been repaid at “91% of their fraud losses,” and that efforts to recover funds are ongoing. Press Release, Justice Department Announces Distribution of Over $158.9M to Nearly 25,000 Victims of Madoff Ponzi Scheme (Dec. 11, 2023), available at https://www.justice.gov/opa/pr/justice-department-announces-distribution-over-1589m-nearly25000-victims-madoff-ponzi. But that does not mean that the loss amount calculated at Madoff’s sentencing in 2009 was incorrect, or that the court in that case should have attempted to predict how much money victims would ultimately be able to recover.

Here, as in Madoff and many other cases, the defendant misappropriated funds from his victims and put large amounts of the money into assets such as real estate, stock in publicly traded companies, and other more illiquid assets for his own benefit. Many of those assets are recoverable, and can be liquidated to repay his victims, but that is due to the extensive efforts of the administrators of the bankruptcy proceeding and due to this prosecution, and these recoveries do not affect the appropriate measure of loss for sentencing purposes. As debtors’ counsel in the bankruptcy recently explained: “Had these Chapter 11 Cases not been filed in November 2022, or had different choices been made in the early months, customers of FTX.com and potentially other FTX creditors would have faced a near total loss.” Fourth Motion of Debtors for Entry of an Order Extending the Exclusive Periods During Which Only the Debtors May File a Chapter 11 Plan and Solicit Acceptances Thereof (“Debtors Motion”), In re FTX Trading, Ltd., Case No. 22-11068 (JTD) (Bankr. D. Del. Mar. 5, 2024), Dkt. 8621, at 2.

In addition to being legally baseless, the defendant’s argument that the loss to FTX customers should be reduced to zero is premised on a distorted depiction of the bankruptcy proceeding. The defendant seizes on one statement at a recent hearing that “customers and general unsecured creditors with allowed claims, will eventually be paid in full.” Transcript of Hearing at 18:11-25, In re FTX Trading Ltd., et al., 22-11068 (JTD) (Bankr. D. Del. Jan. 31, 2024), Dkt. 6908. But he overlooks critical context for that statement. First, the meaning of “paid in full” here is that customers are expected to receive, at some point in the future, “100% of their allowed claims over time, measured at petition time value and subject to agreement by certain governmental creditors to voluntarily subordinate their own very large and impaired claims.” (Debtors Motion at 2).

There is a significant difference between the defendant’s claim that customers will “get back all of their money” (Def. Mem. at 17), and the Debtors’ more measured statement that it hopes to return “100% of their allowed claims over time.” The example of the two customer witnesses who testified at trial demonstrates this difference. Marc-Antoine Julliard, for instance, transferred the equivalent of about $140,000 in fiat currency to his FTX account in April and May 2022. (Tr. 78). Julliard used his deposits to purchase Bitcoin, and intended to hold Bitcoin for five to ten years. (Tr. 73-74). He purchased the Bitcoin based on the defendant’s false representations that when an FTX customer purchased Bitcoin in the spot market, FTX was holding Bitcoin that the customer could withdraw on demand. (Tr. 80). Julliard used almost all of his deposited funds to purchase three Bitcoin, which were trading around $40,000 at the time of his purchase. (GX425). As of approximately November 11, 2022, FTX was still representing to Julliard that his account contained three Bitcoin, which were then valued at a total of approximately $52,000, or around $17,000 per Bitcoin. (GX-425). This was the lowest price Bitcoin had closed at since November 2020.[6] In fact, however, FTX was not holding any Bitcoin for Julliard, nor was it holding his initial fiat currency deposits of $140,000. Instead, the defendant had used that money, along with the deposits of countless other FTX customers, to fund investments through his private investment company, Alameda.

Similarly, Tareq Morad funded his FTX account with approximately $500,000 in fiat currency. (Tr. 1288). Like Julliard, Morad bought cryptocurrency on the spot market, and simply held those funds. When FTX declared bankruptcy, the company was still falsely representing to Morad that his account held more than two Bitcoin and 86 Ethereum as well as certain other cryptocurrencies, with a total value at the time of $257,000. (GX-539). But due to the defendant’s misappropriation of funds, the company was holding neither the cryptocurrency that Morad had been told he had purchased, nor his initial fiat deposits.

Since FTX’s bankruptcy, cryptocurrency prices have increased. Bitcoin is currently trading at more than $60,000, and Ethereum is trading at nearly $4,000. Thus, if not for the defendant’s fraud, Julliard would currently be holding three Bitcoin that he would be able to withdraw for approximately $180,000, and Morad would be holding cryptocurrency worth well in excess of $400,000. But instead, Julliard and Morad are holding a claim in bankruptcy for the dollar value of their FTX accounts as of the bankruptcy date, which are $52,000 and $257,000, respectively, and have no clear timeline for when those claims will actually be paid. Thus, even if the Debtors are successful in paying “100% of their allowed claims over time,” Julliard and Morad will never get back either the amount of actual fiat money they deposited in FTX nor the cryptocurrency they were falsely told their deposits had been used to purchase. The defendant’s fraud took away these funds from these victims: he left them stuck not with “all of their money,” but with a claim for future receipt of the dollar value of their cryptocurrency as of the time of the bankruptcy petition. Several victim impact statements make this point: the bankruptcy recovery, as currently conceived, will not in their view make them whole.[7] (See, e.g., Victim Impact Statements Nos. 32, 33, 38).

The defendant’s claims about customer recovery additionally lack context because they fail to take into account the substantial efforts undertaken by the bankruptcy estate to recover value for FTX customers. This has included, for example, lawsuits to claw back money given to the defendant’s parents and money invested in K5 for the defendant’s benefit. See Debtor’s Motion at 10-11 (describing efforts undertaken to recover assets, including “significant litigation claims”). As proven at trial, these funds were customer money that the defendant invested for his own benefit, and the fact that these funds may be recoverable through the bankruptcy proceeding does nothing to reduce the defendant’s culpability. (Tr. 1320-22 (Singh testimony about defendant’s investment of customer funds in K5)).

Thus, the defendant is engaged in nothing more than fanciful speculation when he suggests, without citing any evidence, that FTX could have simply “resumed withdrawals on November 15th after selling sufficient assets to cover the $8 billion.” (Def. Mem. at 18). In the real world, even in March 2024, and after sustained effort by the bankruptcy estate, the company has not yet paid out any customer claims, has not yet been able to recover all the assets misappropriated by the defendant, and has not yet been able to fully liquidate the assets that it has recovered. Similarly, the defendant’s suggestion that FTX was “solvent as of the date of the bankruptcy petition” is contradicted by overwhelming evidence. As noted, the defendant himself circulated internal messages in the days before declaring bankruptcy in which he acknowledged that there was an $8 billion hole in customer deposits that could not be repaid on any short-term time horizon. (Tr. 2789). Even now, the defendant acknowledges that he would have had to suspend FTX withdrawals while he attempted to liquidate assets that he was holding in a separate company, Alameda, in order to keep the business operating. That is the textbook definition of insolvency. Pereira v. Farace, 413 F.3d 330, 343 (2d Cir. 2005) (under Delaware law, “a company is insolvent if it is unable to pay its debts as they fall due in the usual course of business”).

The defendant suggests in the alternative that the loss to FTX customers should be calculated based on their “cost of collection in bankruptcy.” (Def. Mem. at 22). That argument is flawed for the reasons already given, because it would only come into play if the defendant could claim credit for bankruptcy recoveries to reduce the loss amount, which he cannot. And the defendant’s citation to United States v. Abiodun, 536 F.3d 162 (2d Cir. 2008), provides no support for his position. In the section of Abiodun relied on by the defendant, the Second Circuit was considering how to quantify the “number of victims affected by defendants’ conduct,” for purposes of the Guidelines enhancement for number of victims. Id. at 167-68. In that context, the court held that victims of an identity theft scheme who had been reimbursed by their credit card companies could still be counted as victims, if the district court included the monetary value of the time those victims spent seeking reimbursement in the total loss amount for the offense. Id. at 169. Thus, the holding of Abiodun is that the costs of collection should be added to the total loss amount, not that such costs are the only permissible loss amount. See id. at 169 (remanding for district court to “recalculate the loss amount … to include the time lost by these potential victims”).

4. Even Setting Aside FTX Customers, the Loss Amount to FTX Investors Is Greater Than $1.7 Billion

As discussed, the defendant’s arguments attempting to minimize FTX customer losses are legally and factually meritless. However, even if all of the defendant’s arguments about customer losses were credited, that would have no effect on his Guidelines range because the losses to FTX investors alone are far greater than the highest loss amount enhancement in the Guidelines. The defendant was convicted of securities fraud for his scheme to fraudulently obtain more than $1.7 billion in equity investments into FTX.com, and even the defendant acknowledges that equity investors are not expected to recover from the bankruptcy plan on which he relies for his arguments about customers. (Def. Mem. at 20). Instead, he simply tries to wave off the losses to investors by saying that they “received an equity stake, not a ‘cash-back guarantee.’” (Def. Mem. at 20).

That is not the law. In a securities fraud case, the loss amount is calculated as the decline in value of the securities purchased by investors, insofar as that decline is a “result of the fraud.” United States v. Rutkoske, 506 F.3d 170, 179 (2d Cir. 2007). In cases where the securities decline in value over time for multiple reasons in addition to a defendant’s fraud, the “portion of a price decline caused by other factors must be excluded from the loss calculation.” Id. at 179. In this case, however, the defendant raised funds from investors at a valuation of $32 billion for FTX.com in early 2022, and continued to seek fundraising at that valuation into the fall of 2022. Then, once his fraud was exposed, the company declared bankruptcy within a matter of days and the investors immediately wrote the value of their investments down to zero. This is therefore an example of a case “where share price drops so quickly and so extensively immediately upon disclosure of a fraud that the difference between pre- and post-disclosure share prices is a reasonable estimate of loss caused by the fraud.” Id. Accordingly, the full amount of the funds invested is the appropriate measure of loss.

5. In the Alternative, the Intended Loss Is Even Greater Than $10 Billion

U.S.S.G. § 2B1.1, defines “loss” as “the greater of actual loss or intended loss.” U.S.S.G. § 2B1.1 cmt. n.3(A). The commentary further defines “intended loss” as “the pecuniary harm that the defendant purposely sought to inflict,” even if that loss did not occur or was impossible, id. cmt. n.3(A)(ii). Due to the more than $8 billion dollars of actual losses suffered by the countless victims of the defendant’s fraudulent scheme, there is minimal need to turn to “intended loss” to assess “the seriousness of the offense and the defendant’s relative culpability.” See U.S.S.G. § 2B1.1 cmt. (background); see id. App. C Supp., at 104-05 (Amend. 793) (noting “the Commission’s belief that intended loss is an important factor” because it focuses “specifically on the defendant’s culpability”). If, however, intended loss was applied in lieu of actual loss, the result would be no different for Guideline purposes. The trial record makes clear that the defendant intended to cause more than $14 billion in combined losses to FTX customers, investors, and Alameda lenders.

As a threshold matter, the defendant’s attempt to have this Court disregard the Guidelines commentary regarding § 2B1.1’s definition of “loss” based on the Supreme Court’s reasoning in Kisor v. Wilkie, 588 U.S. ---, 139 S. Ct. 2400 (2019), should be rejected. (Def. Mem. at 16). The Second Circuit and district courts within this Circuit have, even after Kisor, consistently applied the commentary on “intended loss” when calculating the Guidelines range. Specifically, the law is clear that, “[f]or the purposes of calculating the Guidelines range, loss is defined as ‘the greater of actual loss or intended loss.’” United States v. Powell, 831 F. App’x 24, 25 (2d Cir. 2020) (quoting United States v. Certified Envtl. Servs., Inc., 753 F.3d 72, 103 (2d Cir. 2014)). This makes sense because “the larger intended amount is a better measure for the defendant’s culpability.” United States v. Lacey, 699 F.3d 710, 720 (2d Cir. 2012). Other courts have also rejected attempts to rely on Kisor to disregard the Guidelines use of intended loss. See United States v. You, 74 F.4th 378, 396–98 (6th Cir. 2023) (finding—after canvassing several dictionaries—that the term “loss” in § 2B1.1. “has no one definition,” “can mean different things in different contexts,” and deferring to the Sentencing Commission’s interpretation)

The defendant relies on the Third Circuit’s decision in United States v. Banks, 55 F.4th 246 (3d Cir. 2022), to argue that “loss” “unambiguously” means “actual loss.” (Def. Mem. at 16). But Banks is neither binding on this Court, nor persuasive. As noted in You, “without consulting the ‘traditional tools’ of the commentary’s ‘structure, history, and purpose’ to reach that conclusion, Kisor, 139 S. Ct. at 2415, Banks’s attempt to impose a one-size-fits-all definition is not persuasive.” 74 F.4th at 397. Where the Guidelines note that the purpose of estimating “loss,” including intended loss, is designed to assess the “seriousness of the offense and the defendant’s relative culpability,” § 2B1.1 cmt. (background), the defendant’s interpretation would lead to vastly different sentences for similarly culpable defendants where one successfully stole money before causing actual losses, but the other did not. Moreover, the Guidelines commentary with respect to § 2B1.1 “loss” is the very type of “fair and considered [agency] judgment” described in Kisor that has been applied by courts and litigants across the country for years to uniformly calculate loss where fraud schemes, thefts, and embezzlements often vary significantly. See Mistretta v. United States, 488 U.S. 361, 379 (1989) (rejecting a constitutional challenge to the Sentencing Commission and finding that “[d]eveloping proportionate penalties for hundreds of different crimes by a virtually limitless array of offenders is precisely the sort of intricate, laborintensive task for which delegation to an expert body is especially appropriate”). For these reasons, the Sentencing Commission’s § 2B1.1 “loss” guidance is perfectly appropriate to consider when evaluating the defendant’s Guidelines range.

Turning to the facts established at trial, it is clear that the defendant intended to deprive FTX customers and investors, along with Alameda lenders, of even more than the $8 billion suffered in actual losses. See United States v. Jacobs, 117 F.3d 82, 95 (2d Cir. 1997) (“[T]he concept underlying the distinction between actual and intended loss is that the defendant may have the goal of depriving the victim or victims of more than the constraints of the situation ‘actually’ permit. The significance is that the defendant’s acts should be measured by intentions.”).

Here, witness testimony and internal spreadsheets established that between June and November 2022 the defendant knew Alameda owed FTX billions of dollars, up to around $14 billion by September 2022, which could not be legitimately repaid. With the assistance of Government Exhibit 50, a spreadsheet titled “Alameda Balances by FTX sub-info,” dated June 14, 2022, Wang testified that the defendant reviewed the spreadsheet and knew at that time that Alameda owed FTX $11 billion. (Tr. 436-437; GX-50). When excluding Alameda accounts, such as Cottongrove, which contained large amounts of illiquid FTT, Alameda’s debt expanded to $16 billion. (Tr. 619-20). Yet, even after discussing this disastrous financial situation, Bankman-Fried ordered that Alameda’s lenders be repaid with FTX customer money, which enabled the fraud to continue undetected. (Tr. 440, 620). Wang, Ellison, and Singh all testified that by September 2022, conversations circulated amongst the group, including the defendant, that Alameda could not be shut down because it owed FTX—and therefore, FTX customers—approximately $14 billion. (Tr. 447-449, 817-20, 1403, GX-11). When confronted by Singh on the balcony of the Orchid 6 penthouse, the defendant admitted that “in 24 hours, if pressed” FTX could only “deliver around $5 billion,” which left a $8 billion shortfall. (Tr. 1407). Even after this confrontation with Singh, the defendant continued to take FTX customer money and put it into his own investments, which belies his claims that he intended no losses to customers. (Tr. 1417-18).

In his submission, the defendant relies on United States v. Confredo, 528 F.3d 143 (2d Cir. 2008), to support his assertion that the Court should credit that the intended loss is zero because he purportedly intended to repay his victims. (Def. Mem. at 23). But in Confredo, the Second Circuit was interpreting a special Guideline amendment to U.S.S.G. § 2F1.1, which applied only to fraudulent loan and contract procurement cases. Id. at 150-51. The court’s reasoning, therefore, does not apply to this case. And even if Confredo applied here, it would mean only that the Court could consider whether there was evidence that the defendant had an intent to repay. Id. at 152-53 (remanding to allow defendant to present evidence of his intent). Such evidence would need to take the form of a concrete expectation that the losses would be lower than the potential losses implicated in the scheme, such as the anticipated loan repayments at issue in Confredo. Nothing in Confredo suggests that intended loss can be reduced based on the defendant’s general hope that things would “work out,” or his expectation that he could reap personal profits from his investments and then return victim funds before his fraud was discovered. To reduce the intended loss amount on that vague basis would be inconsistent with the Second Circuit’s admonition that “where some immediate loss to the victim is contemplated by a defendant, the fact that the defendant believes (rightly or wrongly) that he will ‘ultimately’ be able to work things out so that the victim suffers no loss is no excuse for the real and immediate loss contemplated to result from defendant’s fraudulent conduct.” United States v. Rossomando, 144 F.3d 197, 201 (2d Cir. 1998).

Here, the evidence undermines the defendant’s claim that he did not intend any loss. Aside from his own self-serving claims, he cites just two things. The first is a letter from an FTX employee who did not begin directly working with the defendant until around November 7, 2022, and who claims that the defendant was seeking to recover funds for customers after that point in time. (Def. Ex. A-9). But even if the defendant was working to recover funds for customers at that time, those efforts to repay victims took place after his fraud had been exposed, when he had an incentive to reduce his potential liability. It says nothing about his intent when he was engaged in the actual fraudulent conduct in the years prior. Moreover, there is no real evidence that the defendant actually undertook any efforts that facilitated customer recovery during this time. Notably, the Debtors’ counsel in the bankruptcy has never suggested that the defendant provided any significant assistance. And one of the few concrete actions the defendant took during this time period was to file an affidavit attempting to claim personal ownership of the $650 million worth of stock that he had purchased with FTX customer funds, in an apparent effort to prevent the bankruptcy estate from recovering those assets. (Tr. 2710-11; GX-933).

The second item on which the defendant relies is a footnote in the March 2023 FTX Debtors’ report, which states that there were a number of deposits into FTX from Alameda in the days immediately prior to the bankruptcy. (Def. Mem. at 23 (citing Def. Ex. G at 18 n.1)).[8] But those returns of funds at the very end could just as well be evidence of the defendant’s continuing fraudulent intent, in that the defendant was trying to cover customer withdrawals and reinforce his false public statements that FTX had sufficient funds on hand. Nothing cited by the defendant comes close to countering the overwhelming trial evidence that during the fraud scheme, he intended to deprive customers, investors, and lenders of billions of dollars for his personal benefit. The defendant misappropriated those funds to invest in a variety of speculative assets. Even if he hoped that those investments would prove profitable, enabling him to escape detection for his misuse of victim funds, that is not sufficient to reduce the intended loss calculation.

Based on the trial record, there is ample evidence to establish that at the height of FTX’s decline, Bankman-Fried intended to cause in excess of $14 billion of losses to FTX/Alameda victims, which he had no intention of ever legitimately repaying.

[3] Arguably, the defendant should not benefit from this provision, because he returned these funds in an effort to conceal his fraud and prevent more customers from making withdrawal requests. To the extent the Court were to find that the defendant’s return of misappropriated funds between September and November 2022 was intended to keep his fraudulent schemes operating, the loss amount to customers would be properly calculated at $14 billion, rather than $8 billion. See, e.g., United States v. Vilar, 729 F.3d 62, 96 n.34 (2d Cir. 2013) (rejecting the argument that loss amount should be reduced by amount returned to investors because “defendants should not benefit from attempting to ensure the continuation of their scheme”); United States v. Carrozzella, 105 F.3d 796, 805 (2d Cir. 1997), abrogated in part on other grounds, United States v. Kennedy, 233 F.3d 157, 160-61 (2d Cir. 2000) (“We have held that loss in fraud cases includes the amount of property taken, even if all or part has been returned.... One reason for this rule is that … the return of money as interest or other income is often necessary for the scheme to continue.”); United States v. Mucciante, 21 F.3d 1228, 1237-38 (2d Cir. 1994) (“Although [defendant] returned some of [victim-1]’s money, and repaid [victim-2] and [victim-3], he did so as part of a meretricious effort to maintain their confidences. He is therefore not entitled to credit for sums returned, or for sums spent for [victim-1]’s benefit.”).

[4] Additionally, the defendant was convicted of money laundering, and the Section 2S1.1 Guideline groups with Section 2B1.1. There is no “credit against loss” in the money laundering context: under Section 2S1.1, defendants who launder money are accountable for the “offense level for the underlying offense from which the laundered funds were derived.” U.S.S.G. § 2S1.1(a)(1); United States v. Eckstein, No. CR 12-3182 JB, 2016 WL 546663, at *7 (D.N.M. Feb. 3, 2016) (no “credit against loss” in money laundering case because that credit “would not reflect a money laundering scheme’s full impact.”).

[5] Even if Durham did support the defendant’s position here, it is contrary to the Second Circuit’s decisions in Ware and Bergstein. And the district court’s sentencing in Durham is hardly persuasive authority for this Court. As the district court in Durham itself said: “I don’t know about what goes on in the Southern District of New York. I visit there only rarely. This is the Heartland. This is where we work hard. … We drive Chevies and Buicks and Fords, not Bugattis.” Durham, 766 F.3d at 685.

[6] Though not necessary for the Court’s analysis, the Government has every reason to believe that at least a significant causal factor in the drop in Bitcoin’s value in the weeks leading up to FTX’s collapse was market instability due to revelations about the defendant’s misconduct.

[7] Some FTX customers are receiving even lower value in bankruptcy. Holders of FTX’s token FTT are near the bottom of the priority list—sitting in thirteenth place—and the Government understands that it is unlikely that holders of that token will receive compensation from the estate. See In re FTX Trading Ltd., et al., 22-11068 (JTD) (Bankr. D. Del. Jan. 31, 2024), Dkt. 4861 at 27-28.

[8] The defendant’s brief erroneously cites Ex. F rather than Ex. G.