FTX Restructuring, Falling from Grace

Welcome to the 120th edition of the Pari Passu newsletter.  

Following last week’s cryptocurrency post, today we will dive deeper into FTX’s bankruptcy that shocked the world in 2022 and is still discussed today. While not a standard bankruptcy in terms of its creditor base and valuation processes, FTX’s Chapter 11 shines a light on how the bankruptcy forum offers tools to navigate complex interests and corporate structures.

From the reasons behind financial distress, monetization strategies, and unexpected creditor recoveries, let’s dive deeper into what we can learn from one of the most influential bankruptcies of the digital era. 

But First, a Message from Hebbia

Work a grind? It doesn’t have to be. Hebbia lets you do in minutes what used to take hours. Need to summarize a credit agreement or build a CIM? With Hebbia, you can do it faster than your MD can say, “This is what you signed up for.” 

By automating repetitive tasks and synthesizing endless data into clear, actionable insights, Hebbia frees up PE / credit pros to focus on what they actually signed up for: making deals, growing companies, and maximizing returns (and avoiding Chapter 11).

Book a 20-minute demo to see why 1/3rd of the largest asset managers trust Hebbia. 

A Brief Overview of FTX [1] 

Co-founded in May 2019 by Sam Bankman-Fried (SBF) and Gary Wang, FTX was a global cryptocurrency exchange that offered a wide range of services like spot trading, future, options, and tokenized assets. It had more than one hundred subsidiaries of different types: FTX.com was the main exchange platform, FTX.US was the US-based exchange, Alameda Research was a trading firm and market maker, FTX ventures invested in crypto startups and blockchain projects, and foreign subsidiaries such as FTX Digital Markets operated in the Bahamas. 

FTX quickly gained traction, first in the VC world. It would receive early VC funding from Binance, a top crypto exchange at the time. After a very successful Series B round in July 2021 that raised nearly $1bn from investors such as Lightspeed Venture Partners and Sequoia Capital, FTX was already sitting at an impressive $18.0bn valuation. In the following Series B-1 and C funding rounds, FTX would raise about $800mn in total, garnering further investment from firms such as BlackRock, SoftBank, and Paradigm. 

In its peak of 2021, the company had more than one million users and was the third-largest cryptocurrency exchange by volume. By January 2022, FTX was valued at $32bn, a 77% increase from six months prior.

The Crypto Winter of 2022

As a quick recap of the last post, the crypto collapse started in May 2022 when the TerraUSD stablecoin collapsed. A stablecoin is a type of cryptocurrency designed to offer a stable value relative to another asset like the USD. One stablecoin is theoretically backed by one dollar, but this dollar soon decreased to $0.95. In reality, investors found out that there was no real collateral to support the price. As a result, the TerraUSD collapse wiped out $400bn of the $2 trillion crypto market cap. 

This wasn’t just a collapse of one stablecoin – it was the start of a panic sell-off that spread across the crypto market. Hedge funds like Three AC heavily invested in the TerraUSD faced low liquidity and filed for Chapter 15 bankruptcy as they could not pay back a total of $3.5bn loans to crypto exchanges and lenders. Between November 2021 to November 2022, the prices of many crypto tokens plunged, including Bitcoin’s drop by 77.5%. 

The Downfall of FTX

While FTX grew impressively within a few years, its downfall was quick and impactful as well. Months after the Terra-LUNA crash, crypto markets, although not fully recovered, have stabilized. Then, on November 2nd, 2022, SBF and FTX’s fates would change forever. For FTX, it was the November 2022 Coindesk article that raised concerns about Alameda’s insolvency. Behind the article was Coindesk’s senior reporter, Ian Allison, who was told by multiple sources to dig deeper into Alameda’s weak balance sheet even before the collapse of crypto platforms in early 2022[2]. 

Some key takeaways from the article were: 

1. FTX and Alameda were not independent of each other. Despite countless interview questions regarding the conflicts of interests between the two companies, SBF had time and time again insisted that he was no longer an employee of Alameda and that the operations of the two companies were completely separate. 

2. Alameda played a key role in the price regulation of FTT tokens, a type of token created by FTX, and offers discounts to FTT traders. In fact, Alameda held a huge portion of the available FTT pool and would regularly buy up huge amounts of the token to maintain FTT’s price when it dropped. 

3. Alameda’s assets were greatly inflated. Although FTT was listed as an asset of the company, and a major asset at that, it could not actually be liquidated. This was because since Alameda owned so much of the available pool, it could not sell large amounts of FTT without its price completely crashing to zero. Therefore, the FTT that was used as collateral to borrow customer funds was essentially worthless as an asset. 

Binance was an early investor in FTX, and it owned about 20% of the company. In July 2021, FTX bought back Binance’s  $2.1bn equity stake in FTX, financed part in cash and part in FTT tokens. On November 7th, 2022, four days after the article was released, the CEO of Binance, a Canadian-Chinese entrepreneur named CZ, would announce on X that his company was liquidating its remaining $550mn in FTT, which was sold to them when FTX bought out their shares. 

The announcement immediately sparked fear in the crypto space, and many FTX customers began withdrawing assets as they started doubting the reliability of FTX and crypto generally. The tweet alone generated an FTX customer bank run. To stay afloat during this emergency,  FTX would begin negotiations with Binance on November 7th to complete a strategic acquisition of FTX by Binance. However, when Binance analysts began their due diligence processes on FTX, they were unable to complete the deal. FTX’s poor accounting practices and abysmal balance sheet made it impossible for the analysts to gain any kind of understanding of FTX from a financial perspective. Within 72 hours, FTX faced a total of $6bn in customer withdrawals, which drove down the price of FTT by 77%  from $22 to $5 per token. FTX ended up halting the withdrawals, but the damage had been done: the decline in token prices directly eroded the asset value on the balance sheet. 

To assess the level of financial distress for a crypto exchange, we have to look beyond income statement metrics. In 2022, FTX’s operating income was projected to drop by 63% to $144mm while the revenue was projected to increase by 10% to $1.1bn as the exchange offered discounts to attract many users. These income statement figures did not fully capture FTX’s declining financial health. What really had been overlooked by the market was the problematic balance sheet. On November 8th, a Financial Times article leaked that FTX only had $900mm in assets compared to $9bn in liabilities. While adding less liquid and illiquid assets would lead to $9.59bn total assets, we will focus on the liquid $900mm assets as they truly represented what FTX was worth. For crypto exchanges, customer deposits are considered both an asset and a liability. They hold customer deposits but owe it to their customers. For an exchange to be solvent, liquid assets should exceed customer deposit liabilities to meet customer withdrawal demands [20].

The problem: only customer deposits in fiat currency (USD, EUR) were recorded as assets and liabilities. Since the customer deposits in crypto were not recorded, assets and liabilities were significantly understated. If the crypto deposits had been recorded as assets, customers would have easily noticed the $4.1bn of FTT transferred from FTX to Alameda, which will be elaborated in the next section [21].

Another problem: in total, FTX misused $10bn in customer deposits and corporate funds in the form of cash and stablecoins, leading to a severe liquidity shortage of $8bn (gap between $900mm assets and $9bn liabilities). However, these transactions from FTX to Alameda were never recorded on the balance sheet to avoid the scrutiny of regulators. Since only $9bn in liabilities were initially recorded, the balance sheet understated true liabilities by at least $4.1bn of hidden funds to Alameda [22]. 

On the assets side, customer deposits in crypto tokens were not recorded on the balance sheet. As a result, on the liabilities side, the balance sheet did not reflect loan payables to many other customer accounts and third parties (blue, red, grey boxes in Figure 2 below), resulting in an estimated range of true assets and liabilities between $10bn and $50bn.  

All this time, FTX was able to get away with this balance sheet because the value of tokens was artificially inflated (as explained below). Another reason was that the balance sheet didn’t need to be balanced. As a private company, FTX never had to audit or release its financial records publicly. Under SBF’s trustworthy public persona, investors, regulators, and customers never questioned FTX’s financial health. 

Most concerning was the percentage of assets consisting of highly illiquid tokens like FTT and Serum, which declined in value by 91% and 60% in one week, respectively. Similar to FTT, only 3% of Serum was publicly traded while 97% was held by FTX. Big problem: Serum was made up by FTX – it only traded mostly within the FTX exchange. So if FTX ever decided to sell Serum in the open market, it would flood the market with too much supply compared to the demand, dropping the prices significantly.

Although not all assets were created by FTX, only $900mm out of the $19.6bn in total assets were cash or liquid (easily convertible to cash) stock and cryptocurrencies. All this time, customers had faith that FTX could easily pay them back. But in the end, it turned out that assets were only one-tenth of reported liabilities [3]. More importantly, where did all this money go?  

Simply put, there are a few reasons for FTX’s failure that will recur throughout the post, including: 

• Absence of corporate governance: there were no third-party investors on the board and all decisions were in the hands of a group of inexperienced, unsophisticated individuals.

• Control failures: financial statements were either non-existent, limited, or completely unreliable. Assets and liabilities were shuffled among the subsidiaries and insiders without proper documentation, such as sending customer funds to Alameda without customer consent.

• Use of native tokens as collateral: FTT was used as collateral for Alameda’s trading activities, but this was highly risky since the value of FTT was dependent on the market’s perception of the FTX and the token’s future value. A simple analogy would be: the crypto token is the stock for FTX, the company. If FTX generates strong profits, the token’s value will rise. 

• High concentration of illiquid assets: a lot of FTX’s tokens had low trading activity and would result in low valuation under liquidation. 

• Poor Cybersecurity: Ignoring security controls, FTX put all their assets in “hot wallets,” which are less secure online cryptocurrency wallets used for conveniently managing digital assets. On November 11th, 2022, A group of hackers stole $477mm worth of cryptocurrency stored in FTX using a service called ‘mixers’ that hid their traces across the blockchain network. The hacked money represented 15% of the total $3.1bn owed to the largest creditors [4]. 

Where did my money go? FTX’s Commingling of Funds [5]

What makes FTX’s bankruptcy particularly interesting is that it wasn’t simply about splitting up the pie of fixed assets – the main objective among lenders of a distressed company. The debtors not only had to locate lost assets but also understand how much could be recovered. 

Without clear intercompany loans between subsidiaries, assets and liabilities traveled through different entities without proper records. The graphic below may look complex, but there are a few main avenues where the $10bn of customer deposits and corporate funds were ‘commingled’ by SBF and other directors, meaning they were used for unrelated purposes. 

As a customer of FTX, you would want your funds to be stored safely or only used for trading. However, these were transferred to bank accounts (the black boxes) of Alameda Research, a subsidiary called North Dimension, and FTX Digital Markets, a non-debtor subsidiary that was solvent. Here, FTX clearly breached its fiduciary duties. They used their customers’ money for Alameda’s general corporate purposes, speculative trading, venture investments, purchase of luxury properties, and political donations to bolster FTX’s status. Then, the customer funds flowed into various debtors’ wallets (the blue boxes) across various jurisdictions. In some cases, they would end up in the gray boxes in the hands of third parties with the value to be determined. 

A transfer with a significant impact on customer recoveries was transferring FTT, FTX’s native cryptocurrency, to Alameda Research. By June 2022, the Coindesk report revealed that 163mm FTT tokens with a market value of $4bn were secretly transferred to Alameda. As a quick recap from the crypto post, FTX had a feature called spot margin trading that allowed users to borrow funds from other users. Just like a traditional loan, users would pay principal and interest. For example, if the leverage was 10x, users could put down $100 of their money to make a $1000 trade. Before regulators stepped in to bring down maximum leverage to 20x, users were allowed to trade at leverage as high as 101x. To fund their trades, Alameda took out loans from different crypto lenders like Genesis and Voyager Digital, with customer funds in cash or FTT as the collateral backing these loans.  

Making trades with FTT gave the false impression that the token was in high demand, so Alameda could continuously borrow more funds. In reality, there was little value to the tokens. When collateral is a tangible asset like real estate, lenders have more certainty about the real value of the assets. The problem was that FTT was not a reliable source of collateral – it did not adequately capture the risk behind the FTX exchange or any liabilities on FTX’s balance sheet. Key point: the value of FTT was directly correlated to the market’s perception of FTT’s future demand. But if FTX collapses, how would FTT exist? When the market lost confidence in the stability of FTX, FTT prices declined by 75% in one day and automatically triggered a liquidation of customer loans, meaning FTT was sold off in a fire sale. Within five days, FTT’s market value was wiped out by 90.5%, from $2.94 trillion to $281bn. As panic rippled through the market, customers holding the top six crypto tokens withdrew a total of $20.7bn across the crypto exchanges between November 2nd and November 13th, 2022 [19]. Specifically, users on FTX withdrew at least $5bn worth of crypto tokens. Firstly, this meant that true liabilities on the balance sheet would have been at least $14bn (from the existing $9bn, which did not include crypto deposits as explained above). Secondly, customers could still withdraw less liquid crypto tokens ($5.5bn as recorded above) since this was just moving their tokens to external wallets. However, beyond this, FTX could not meet withdrawal demands and froze customer accounts.

The possibility of monetizing FTT looked dim. Having a separate balance sheet from FTX, Alameda held around 25% of their $14.6bn assets as unlocked FTT tokens. $292mm out of $7.4bn in liabilities were in locked FTT tokens. Intuitively, listing tokens as an asset would make the most sense since Alameda can freely sell the tokens in the market for profit. However, locked tokens were treated as a liability as they could not be monetized to pay back customers. Locked tokens are contractually prevented from being traded freely in the market for a period of time, so they have little ability to boost FTX’s liquidity. Any future unlocking would suddenly increase the supply and drive down FTT’s market price. Similarly, unlocked tokens could contribute very little, given FTT’s market value already declined heavily after FTX’s bankruptcy [6]. 

This left customers unable to withdraw their funds as FTX and Alameda did not have enough liquidity. Furthermore, they have never agreed to be exposed to Alameda’s risk. Naturally, customers demanded that they should be first in line to receive any returned funds. Where the tension rose in court, however, was customer creditors demanding a senior position in the capital structure. More discussion between the two competing creditor groups will come below. 

Without the funds to pay back its customers and with no one to bail it out, FTX filed for Chapter 11 bankruptcy in the District Court of Delaware on November 11th, 2022. Total liabilities were $7.8bn based on the unaudited balance sheets prepared by SBF. But based on the petition filing, both assets and liabilities were about to look much higher at an estimated $10bn-$50bn as the new debtors stepped in to uncover the missing figures. 

The Critical Role of New CEO John J. Ray III [7]

Before SBF was arrested and extradited to the U.S. on December 12, 2022,  John Ray stepped in as the new CEO on November 17th, 2022, to navigate FTX through bankruptcy. As a restructuring veteran, Ray tried to use Chapter 11 to achieve five goals: implement controls, protect and recover assets, conduct transparent investigations, efficiently coordinate with any non-US proceedings, and maximize value. The top priority for John Ray and the debtor’s advisers was reorganizing the corporate structure and building an accurate balance sheet. 

“I have over 40 years of legal and restructuring experience. I have been the Chief Restructuring Officer in several of the largest corporate failures in history…Never in my career have I seen such a complete failure of corporate controls and such a complete absence of trustworthy financial information as occurred here.” 

– John J. Ray III

Cleaning Up the Corporate Structure with ‘Silos’ [8]

Many companies operating in different segments often have multiple or even dozens of subsidiaries. In the debtor’s case, they were operating without an appropriate organizational structure. Instead of a parent company making a centralized decision, the debtor had separate subsidiaries with different owners and interests. SBF himself owned a majority interest in the subsidiaries, while the two other co-founders, Gary Wang and Nishad Singh, owned minority interests. However, once John Ray stepped in during Chapter 11, he reorganized the debtor’s capital structure into four clear categories called ‘silos’ with new independent directors to address issues of overlapping ownership, interconnected financial accounts, and unclear legal entities. 

Was the debtor a single subsidiary or the parent company? Where did the assets and liabilities exactly sit? Resolving these questions was key to determining creditor recoveries. One caveat is that the balance sheets below were produced by SBF and the previous management team updated them as of September 2022, two months before the filing. With inflated assets and missing liabilities, the materials portrayed the debtors as a financially healthy company. For example, some silos did not reflect loans payable and intercompany loans on the liabilities side. Given how unreliable these metrics were, one of the top priorities for the debtors throughout Chapter 11 was reconstructing these balance sheets to conduct an accurate valuation. Let’s evaluate how much real value the assets recorded on the balance sheets had at the onset of the bankruptcy (all figures below are in thousands). 

1) West Real Shires Inc. (WRS) Silo 

The WRS Silo mainly included FTX US, the exchange market for digital assets and tokens with 1mm users (out of 7mm total FTX users) as of August 2022 and other subsidiaries. The most notable asset line item was the $250mm loan receivable from BlockFi, which the debtor lent to bail out BlockFi. However, BlockFi filed for Chapter 11, making the prospect of repaying the loan slim as they would prioritize their creditors, and asset value was tied to the crypto market. 

2) Alameda Silo 

This silo had around a third of assets coming from loans receivables, but these were difficult to be monetized as actual cash. As the provider of liquidity to other subsidiaries, Alameda made $4.1bn loans to related parties (these are called intercompany loans): $1bn to SBF, $543mm to Nishad Singh, the Director of Engineering, and $2.3bn to SBF’s insolvent subsidiaries. They were also loans to the debtor’s executives and SBF for personal investments, but proper documentation was hard to find. Furthermore, many of these loans were backed by illiquid collateral, such as FTT tokens, so they were nearly unsecured (more to be discussed below). The $4.1bn of crypto assets in fair value would have been a significantly inflated value. When investors lost confidence in the FTX exchange, withdrawals plunged the FTT prices from $22 to $2 in a week after the debtors filed for bankruptcy.

3) Ventures Silo 

Under this silo, the debtors made a total of 200 private investments in tokens, loans, and venture capital projects, which represented nearly all of the asset value. As the venture arm of FTX, FTX Ventures made investments but also spread its funds across a company called Clifton Bay Investments. Due to venture investment valuation being dependent on future value, the fair value was ambiguous as some investments were still early stages in a market with declining investor confidence. 

4) Dotcom Silo

This silo included FTX.com and licensed subsidiaries in non-US jurisdictions such as the EU and Japan. Nearly half of the assets came from U.S. dollar-denominated stablecoin, which is a type of cryptocurrency intended to have a value of $1 for every $1 of US dollars. Many customers held their funds in stablecoins on the platform. Theoretically, stablecoins can be easily converted into cash. However, panic selling caused the stablecoin prices on the debtor’s exchange to drop from $1 to $0.97. This was still a minor drop compared to drops in cryptocurrency value, so the value erosion wasn’t too extreme.

Monetizing the Debtor’s Estate [9] [10]

As the bankruptcy proceeded, more litigation claims were filed against the debtors, and more assets were dug up to compensate various stakeholders in real time. Unlike traditional businesses, crypto companies do not have significant physical inventory, real estate, or other tangible assets. Hence, an asset’s liquidity would impact the timeline of the bankruptcy and, ultimately, creditor recoveries. As a reminder, the leaked balance sheet in November 2022 recorded only $900mm in assets, but this value gradually increased as the debtors monetized new and existing assets. Let’s break down the $6.7bn assets, including $2.4bn of cash, that the debtors identified as of September 2023, but as a caveat, the projections were subject to significant changes (note, Liquid defined as “L” and Illiquid defined as “IL”)

1. Cash: mainly driven by venture investment monetization and stablecoin conversion. The debtors started off with $1.7bn of cash on the petition date, which rose gradually to $4.4bn by January 2024, as the debtors monetized assets into cash during Chapter 11. 

2. Crypto Assets (L&IL): a crypto token is considered illiquid if its trading volume and market cap are low. It could also be where FTX holds a significant portion of the token, so liquidating it would drive down the market value due to oversupply. 20 types of tokens made up 98% of the illiquid value. At the start of the bankruptcy, the debtors only found $3.5bn of crypto assets and continued searching for more from various sources, especially on third-party exchanges. 

3. Venture Investments (IL): a total of 438 investments worth $4.5bn of cash and crypto in equity, funds, loans, and tokens. Since it was hard to value these intangible assets, valuation relied on a combination of bids for the portfolio, company-provided information, funded values, recent financing rounds, and market conditions. However, recovery values were expected to be lower than the acquisition price because interested buyers would most likely demand a discounted price, and the private nature of investments would make valuation difficult.

4. Brokerage Assets (L): originally, the debtor had stakes across different Bitcoin trusts, such as Grayscale Bitcoin Trust (GBTC) created in 2013, which purchases Bitcoin on behalf of investors. In January 2024, the SEC approved Bitcoin ETFs since they were a safer way to get crypto exposure. After GBTC converted into an ETF, the debtor sold off $1bn worth of shares in GBTC [12]. 

5. Real Estate in the Bahamas (L): FTX DM and FTX Bahamas PropCo were two subsidiaries in the Bahama area with 35 real estate properties worth $222mm in book value.

6. Avoidance Actions (L): simply put, avoidance actions are legal actions from debtors to recover money or property from pre-petition transactions. The goal is to improve creditor recoveries with more assets. Here, the debtors estimated $16.6bn could be recovered and redistributed to creditors. To clarify, these were not directed at the current debtors but at parties that pushed FTX into bankruptcy. For example, creditors could demand $2.2bn of cash and crypto value received by fifty insiders, including SBF, $86.6mm in political donations to third parties, $3.2bn received by third parties not part of the debtor’s business, $5.0bn in loans made to crypto lenders like BlockFi, and $5.3bn worth of investments the debtors had stakes in, such as venture capital projects, which were initially not included in the bankruptcy but later accounted for under the avoidance action. As of September 2023, $588mm was successfully clawed back, including $474mm of venture investments in hedge fund Modulo Capital [23]. While not all were clawed back throughout the bankruptcy due to litigation settlements, the debtors successfully sold their $500mm stake in AI startup Anthropic [24]. 

7. Sale of Subsidiaries (L): despite operating in different jurisdictions, the court approved the sale of foreign subsidiaries such as FTX Europe. Located in Switzerland, FTX Europe did not have enough assets to cover its $125mm in liabilities. However, they successfully attracted buyers and raised $32.7mm, which contributed to the overall debtor’s estate.

8. Contribution from Solvent Subsidiaries (L): FTX Japan and Ledger Prime had enough assets to pay out creditors, leaving $202mm of residual equity value to flow into the consolidated debtors.

Given the scale and multiple lines of operating businesses, the debtors had a wide range of options to raise additional liquidity. In fact, most of them were classified as liquid. By June 2024, the debtors arrived at $9.9bn of cash and an expected extra $2.7bn from selling digital assets, which they believed could pay off all the non-governmental customers and creditors in full. The strategy was not to conduct a fire sale of assets to exit bankruptcy quickly. Instead, the debtors opportunistically waited for the right market timing that would give the assets a maximum valuation. 

Going forward, the debtors received the court’s approval that would transfer all assets from different silos into a ‘Consolidated Wind Down Trust.’ This was allowed under the concept of substantive consolidation, where all assets and liabilities of multiple debtors are combined into a single estate. The intercompany loans are eliminated as well. While not common due to the court’s strict requirements, it is used when the debtor's assets and operations are extremely interconnected. The Wind Down trust would be free from any existing or future legal claims and liens, meaning they could get full control over assets without having to pay out existing liabilities. The only purpose would be to liquidate assets and distribute cash to creditors. Even though creditors have to share the same assets, the trust was meant to maximize the proceeds of asset sales. 

Given the different levels of liquidity and jurisdiction of the assets, they were pooled into three categories: US, Dotcom, and General. By separating the assets into different groups, the debtors wanted to distribute proceeds from monetizing assets to specific creditors – an attempt to organize the assets and liabilities shuffled between subsidiaries.

• US pool: all fiat currency (government-denominated currency), digital assets, litigation claims, and asset sales related to the FTX.US exchange 

• Dotcom pool: all fiat currency (government-denominated currency), digital assets, litigation claims, and asset sales related to the international exchange FTX.com 

• General pool: all assets not part of the US Pool, Dotcom Pool, or to a Separate Subsidiary (Ledger Holdings and FTX Japan), all excess distributable value of non-Debtor subsidiary and Separate Subsidiaries, all proceeds from Avoidance Actions and litigation claims, the Alameda US Customer claim, all proceeds from the sale of the debtor’s property (besides the Bahamian subsidiaries) 

While the picture below captures the recovery at one point of the bankruptcy, it shows that General Pool had the most options to monetize digital assets and brokerage investments. Each pool had a separate waterfall recovery schedule to determine the recoveries of different creditor groups. 

Settlement with Non-Customer Creditors

Back to the earlier discussion on customer vs. non-customer creditors, it’s important to make a distinction that not all creditors were customers. By June 2023, the debtors were faced with 2,300 non-customer claims for $379bn. After remaining duplicate claims, the remaining $65bn claims were represented by 67% from the Internal Revenue Service (IRS), 14% from FTX Digital Markets, 9% from crypto lenders Genesis and Celsius, and 10% from fraud claims, contract claims, and loans payables [10]. While these tax claims were later reduced from $44bn to $24bn, the IRS claimed that the debtors still owed significant income tax on stolen or misused funds. Without settling these non-customer claims first, the debtors found it difficult to project and distribute cash to the customer claims and receive approval for the plan. 

In June 2024, the debtors settled a $24bn claim from the IRS, including tax obligations accumulated between 2022 and 2024. If the debtors didn’t settle this claim, they would have nothing to return to creditors. Ultimately, the claim was settled for a mere $200mm in cash as the debtor’s legal team argued FTX never earned enough revenue for such a liability, successfully freeing up cash to return to its creditors. However, as a tip to the IRS, the debtors promised an extra $685mm as a Junior Subordinated claim that would be junior to the customer claims. In addition, the debtors also had to settle $5.0bn of claims from crypto lending platforms like Genesis, BlockFi, and Voyager Digital. These were the loans used for margin trading that Alameda Research defaulted on as FTT’s price declined, pushing the crypto lenders to file for bankruptcy. However, some crypto lenders initially pushed back a settlement until the Judge ruled against them. For example, Celsius argued that the debtors had to compensate $2.0bn for destroying the lender’s reputation, which led to huge customer withdrawals that pushed them into bankruptcy.  

From Owner to Creditor – the Creation of ‘Customer Creditors’ 

On the customer creditors' side, the debtors were faced with claims for a total of $16bn as of August 2023, which was gradually reduced to $9.2bn by March 2024 as more claims were settled. Since the beginning of the bankruptcy, some customers have demanded better treatment. In December 2022, a group of FTX US and non-US customers (the Ad Hoc Committee) filed a complaint to the debtors and the previous management team, arguing they should not be treated as unsecured creditors. They believed that the cash and assets never belonged to the debtors and should not be part of the bankruptcy estate as it would be returned to the general pool of creditors. In other words, even if there was a crypto asset that currently does not exist in the customer’s account or is owned by the debtor, it should be returned to that customer. This would prevent lower customer recoveries from sharing cash distribution with non-customer creditors like government agencies [11]. If the court approves this interpretation, customer claims should be higher in the seniority of repayment over non-customer creditor claims.

The debtor responded that when the debtor and customers signed the business contract pre-petition, they did not agree on specific legal ownership of assets, leaving room for open interpretation. For any customer of a crypto exchange, this statement would raise serious concerns in future crypto bankruptcies. When customers signed their user agreements, they believed they were owners of their digital assets. However, the law didn’t legally protect their assets by mixing them with the debtor’s money. Hence, there was no guarantee that they would get their money back. Basically, customers were tricked into being creditors and could not automatically demand ownership over their money in the bankruptcy. Ever since similar issues were found in other crypto bankruptcies, some crypto exchanges have updated their user agreements and started using “Proof-of Reserves (PoR)” that verifies customer assets are fully backed and separate from the exchange.  

Customer-Friendly Plan of Reorganization [14]

The most unique aspect of FTX’s bankruptcy is that the debtors were insolvent, but eight of ten creditor classes were projected to recover more than 100% of their claims. Specifically, 98% of retail investors would get at least 118% of their allowed claims within 60 days after the plan becomes effective (from January 3rd to March 4th, 2025). Intuitively, this does not make sense. While customers found these recovery rates to be misleading, the debtors were able to achieve these unusually high figures due to the following reasons.

One, the assets and liabilities were frozen since the petition date. This meant that customers were paid back in cash instead of crypto tokens based on the value of crypto holdings in November 2022, not at current market prices. During fraud investigations (and while not limited to bankruptcies), government agencies can freeze assets so that they cannot be accessed or transferred to customers. In the debtor’s case, the legally disputed assets were frozen, including some crypto tokens, so there were simply not enough tokens to return to creditors. The debtors were able to generate much more money from the other liquid tokens by holding them during the bankruptcy rather than immediately meeting customer withdrawals when the debtors filed. 

A simple way to think about this is liabilities were fixed but assets went up. For example, a customer had one Bitcoin deposited in the debtor’s exchange. This claim would be worth $16,871 (price as of petition date in November 2022) instead of $66,096 (price as of POR filing in September 2024). Clearly, there was a huge opportunity cost of missing out on the crypto market recovery. By the time the debtors emerged on January 3rd, 2025, Bitcoin prices skyrocketed to $97,942, increasing by 481%. While the customer’s Bitcoin on the petition date was $16,871, if their recovery was in Bitcoin rather than cash, they would have been able to recover $97,942 on the day of emergence. Compared to this, a 118% recovery would only return them $16,871*1.18=$19,907 since the claim was frozen at the lowest price right after the collapse. The customer is only able to recover $19,907/$97,942= 24.6% and will not be able to enjoy the upside if the price of Bitcoin continues to increase over time. 

At first glance, at least 118% in recovery sounded generous, but some customers believed the cash recovery based on the petition date values was unfair. This explains why acceptance rates below never reached 100%. 

Two, the recovery rates above 100% (going as high as 143%) were numerically achievable because of the post-petition interest payments of 9%. These payments were a ‘tip’ from the debtors to compensate BlockFi (Class 3), customers (Class 5 & 7),  and unsecured creditors (Class 6) for the money they could have gained if paid out in crypto tokens. During the bankruptcy, there were still claims that were under review or disputes over the exact amount of claims. When these disputed claims turn into allowed claims, the debtors would pay post-petition interest payments at 9% as if the claims had been paid from the beginning (as an exception, the senior loan claims were paid the market interest rate). 

As a simple example, a claim worth $1bn will be paid back on February 18th, 2025, which is 831 days after November 11th, 2022 (the petition date). Over 831/365=2.3 years, accrued interest at 9% would be 9%*2.3= 20.5%. The post-petition interest would be $1bn*0.205=$205mm, resulting in a $1.205bn/$1.0bn = 120.5% recovery.

According to the bankruptcy code, the debtors usually cannot pay post-petition interest payments to customers and unsecured creditors unless the debtor is solvent or the creditor has more collateral than its claim (see more analysis in Hertz). However, in this case, the creditors were forced to lend to the debtors during the bankruptcy, so the Board approved boosting their returns using post-petition interest.

Three, new assets were discovered in real time and monetized into cash. In May 2024, the projected gross distributable value for the Wind Down Trust was $19.9bn from $12.6bn in cash and cash equivalent and a potential $7.2bn from asset sales. However, the volatility of venture investment and digital asset values resulted in a range from $14.9bn to $16.7bn of distributable value, a figure combining the $12.6bn of cash and $2.6-4.6bn asset sale proceeds after subtracting $448mm of costs from the Wind Down trust. Even so, the fulcrum security (the first class to be impaired) was not a customer creditor class but BlockFi’s $250mm Senior Loan claim. The main question was how FTX decided to design the seniority of the capital structure, given they had to juggle the interests of multiple types of creditors. 

Four, the debtor favored the customers over non-customer creditors. To minimize litigation, customers of the international and US exchange were given a partial priority claim against the General Pool over the general unsecured creditors. In fact, the government agencies agreed to subordinate their claims and allowed the debtors to pay post-petition interest to customers first. Note below how the Class 10 non-IRS government claims were worth $8.9bn, but their recoveries were only 2-17%, while the $24bn IRS claims were settled for $200mm in cash and $685mm junior IRS claims worth 0%. If the government agencies hadn’t been so cooperative, it would have been difficult to boost customer recoveries.

In total, there were $11.2bn of claims, with the following claims most noteworthy in terms of recovery values. In Chapter 11, even though recoveries are above 100%, claims are considered impaired if 1) creditors are paid less than what they were originally owed 2) paid over a longer period 3) a change in contractual rights. As discussed above, the cash recovery with the petition date value fell short of the current value of crypto tokens, impairing claims. The Senior Loan claim from BlockFi was also settled at a much lower claim than the original. On points 2 and 3, before the bankruptcy, the contracts allowed customers to withdraw funds immediately. Even though their claims would be recovered more than 100%, now, they had to wait in line based on their waterfall recovery at a later distribution date, which could be after the plan becomes effective – this would be a change in contractual rights. Among a total of eighteen classes, fourteen were impaired, and seven were the key voting classes. For simplicity, we will focus on the treatment main impaired classes and their estimated allowed claims and recoveries.

FTX 2.0: Possibility of a Restart

Would it make sense for FTX to restart, given its structural problems? Even if 98% of the creditors would be fully paid out in cash or stablecoins under the plan, creditors were dissatisfied that they missed out on the increasingly valuable crypto value. All of their claims were based on crypto values back in November 2022. Since then, crypto prices soared: by January 2025 (when FTX emerged), Bitcoin prices increased by 481% from $16,543 to an all-time high of $97,942 [15]. Similarly, the value of other popular tokens on the exchange rose. It might be counterintuitive to think that the market was optimistic about crypto during a wave of crypto bankruptcies. But in this period, investors grew excited about the approval of Bitcoin ETFs as a sign of wider adoption, and riskier assets gained traction in the expectation that the Fed will start cutting rates in early 2024 [16]. 

Yesterday’s FTX was not the same as today’s FTX. Restarting the exchange would redeem some of the customers’ lost crypto value from their frozen accounts at the start of the bankruptcy. The unsecured creditor committee consisting of crypto firms especially pushed for a restart. The Customer Ad Hoc Committee (CAHC) also filed a complaint against the POR, demanding customers should be given crypto tokens over cash. However, they were not able to establish a sufficient blocking position of one-third in the dollar amount of the voting class to reject the plan. 

While the debtor’s lawyers and advisors explored this option during the bankruptcy, they decided it wasn’t worth it. In September 2023, among seventy-five potential bidders (including a crypto exchange, fintech startup, and investment firm), the debtors found none attractive enough to cover the costs and delays of a restart. On top of the uncertainty of receiving regulatory approval, John Ray questioned the overall purpose of cryptocurrencies and believed creditors should have the freedom to use their returned cash for their own purposes rather than signing up for a new exchange [16]. As a result, the debtors no longer operate as a business. Since its exit from bankruptcy, they’ve only focused on returning cash to the creditors.

The Real Winners

When someone loses, someone always wins. Even though the crypto market collapsed in 2022, some were still bullish on a crypto rebound – while risky, it turned out to be one of the all-time best distressed trades. In the early days of FTX’s bankruptcy, hedge funds such as Attestor Limited, Farallon Capital, Diameter Capital, and Baupost Group bought up steeply discounted creditor claims at 20 cents. Gradually, it rose to 120 cents of face value. By March 2024, their claims were worth more than $1bn combined. Even some individuals like bankruptcy claim broker Thomas Breziel bought an $8mm claim for $240k (3% of face value) as early as November 2022 [17]. The early birds do win, after all. 

State of the Industry and Key Takeaways 

In October 2024, the plan was approved by Bankruptcy Judge John Dorsey followed by an exit in January 2025. In early 2025, the post-reorg debtors will repay customers starting from the convenience claims. Without Chapter 11, FTX would have faced more asset seizures, hacks, bank runs, and conflicts across dozens of creditors across the world. Under Chapter 11, debtors had more time and resources to investigate lost assets, litigate claims, and monetize illiquid investments, resulting in creditor recoveries of over 100%. However, there are still ongoing discussions about how ideal the outcome was. While the debtors prided themselves on the numerical recovery value, perhaps creditors were not fully compensated for the surge in crypto value ever since the petition date. 

Meanwhile, the crypto markets continue to flourish as if the crypto collapse never happened. At the end of 2024, the stablecoin market value reached an all-time high of $190bn as it became more accepted in global commerce with investments from large companies. Crypto has outpaced the growth of the US stock market, with Bitcoin increasing by 122% since November 2023 under expectations of eased regulations [18]. 

No matter the precedence, as long as there is market demand, there will be perceived value, which will drive more demand. In fact, the very creditors who lost and recovered their FTX investments could be fueling this growth. FTX’s bankruptcy was not simply a one-time failure that generated losses for financial institutions and individuals. Its effect also carried an intangible value, impacting millions in making financial decisions and reflecting on investment risk. 

While ironic that SBF himself declared the “protection of investors and the public as a top priority” for the FTX exchanges, FTX’s regulation statements in 2021 are worth revisiting to guide the crypto industry: “ensure customer and investor protection, promote market integrity, prevent financial crimes, and ensure overall system safety and soundness.” 

Sources: [1], [2], [3], [4], [5], [6], [7], [8], [9], [10], [11], [12], [13], [14], [15], [16], [17], [18], [19], [20], [21], [22], [23], [24]

Interested in partnering with Pari Passu? Complete the form here.

Fundamental Edge is launching its next Academy in one week! Get $450 off!