Free to read, search, and study on this site. Cite with attribution; no redistribution or commercial reuse (CC BY-NC-ND 4.0) — License & Terms.

← All cases
F-009Failure series

FTX / Sam Bankman-Fried — collapse of a crypto exchange

2019–2023 · Scandal/Fraud · scored under OTA methodology v4

Scoring

Attribution weights under OTA methodology v4. Percentages express how much of the episode’s outcome each phase and modality accounts for — not a performance grade.

Phase attribution

Observe
0%
Think
100%
Act
0%

Observe Easy-Correct · Think Easy-Wrong · Act Easy-Almost-correct

Modality weights

Direction
35%
Structure
30%
Culture
35%

Modalities scored at zero weight are omitted; the case narrative records why an evidenced modality carries no independent weight.

Primary modality
Direction
Reliability band
High
Fraud-related
Yes

1. Episode summary

FTX Trading Ltd. was a cryptocurrency derivatives exchange founded in 2019 by Sam Bankman-Fried, operating alongside the affiliated proprietary trading firm Alameda Research. By 2021–2022, FTX had raised roughly $1.8 billion from equity investors — including approximately $1.1 billion from about 90 U.S.-based investors per the SEC complaint — at peak private valuations near $32 billion, while marketing itself as a safe, institutionally-run venue. On 2 November 2022, CoinDesk published a leaked Alameda balance sheet showing Alameda's assets were concentrated in FTX's own exchange token FTT rather than independent collateral. Binance's CEO announced on 6 November that it would liquidate its FTT holdings; FTT's price collapsed, customer withdrawal requests at FTX surged past the platform's ability to meet them, and a proposed Binance rescue was abandoned on 9 November. On 11 November 2022, FTX, Alameda, and over 100 affiliated entities filed for Chapter 11 bankruptcy in Delaware. The court-appointed CEO John J. Ray III stated in his first-day declaration that he had "never seen such a complete failure of corporate controls." The U.S. Department of Justice indicted Bankman-Fried on 13 December 2022; he was convicted on seven counts of fraud, conspiracy, and money laundering in November 2023 and sentenced to 25 years and $11 billion in forfeiture in March 2024. The strategic question the episode turned on: whether the governance and accounting discipline required to run a regulated-style customer-asset platform would be built and enforced, or deliberately suppressed in favour of using customer deposits as working capital for an affiliated trading book.

2. Sources

Primary:

  1. U.S. Securities and Exchange Commission, "SEC Charges Samuel Bankman-Fried with Defrauding Investors in Crypto Asset Trading Platform FTX," Press Release 2022-219 and accompanying civil complaint, 13 December 2022 (sec.gov).
  2. U.S. Department of Justice (SDNY), "United States Attorney Announces Charges Against FTX Founder Samuel Bankman-Fried," indictment and press release, 13 December 2022; and DOJ, "Samuel Bankman-Fried Sentenced to 25 Years for His Orchestration of Multiple Fraudulent Schemes," 28 March 2024 (justice.gov).
  3. Declaration of John J. Ray III in Support of Chapter 11 Petitions and First Day Pleadings, In re FTX Trading Ltd., et al., U.S. Bankruptcy Court, District of Delaware, Case No. 22-11068, 17 November 2022.
  4. Commodity Futures Trading Commission, "CFTC Charges Sam Bankman-Fried, FTX Trading and Alameda with Fraud and Material Misrepresentations," Press Release 8638-22, 13 December 2022 (cftc.gov).
  5. Trial testimony of Caroline Ellison (former CEO, Alameda Research), United States v. Samuel Bankman-Fried, SDNY Case 1:22-cr-00673, 10–12 October 2023 (as reported contemporaneously by court reporters and wire services).
  6. Trial testimony of Gary Wang (co-founder and CTO, FTX), United States v. Samuel Bankman-Fried, SDNY Case 1:22-cr-00673, 5–6 October 2023 (as reported contemporaneously by court reporters and wire services; key evidence: order to write "allow_negative" code on 31 July 2019, unlimited credit line for Alameda rising to $65 billion).
  7. Trial testimony of Nishad Singh (co-lead engineer, FTX), United States v. Samuel Bankman-Fried, SDNY Case 1:22-cr-00673, 16–17 October 2023 (as reported contemporaneously by court reporters and wire services; key evidence: September 2022 penthouse-balcony confrontation with Bankman-Fried after learning $8 billion in customer funds was missing; SBF dismissed concern as manageable).
  8. U.S. Securities and Exchange Commission, "SEC Charges Audit Firm Prager Metis for Negligence in FTX Audits and for Violating Auditor Independence Requirements," Press Release 2024-133 and settlement order, 2 October 2024 (sec.gov); confirms Prager Metis issued audit opinions on FTX Trading Ltd. and that the firm lacked the competence and independence to perform those audits.

Secondary:

  1. Ian Allison, "Divisions in Sam Bankman-Fried's Crypto Empire Blur on His Trading Titan Alameda's Balance Sheet," CoinDesk, 2 November 2022 — investigative reporting that triggered the run; synthesises leaked balance-sheet documents.
  2. Michael Lewis, Going Infinite: The Rise and Fall of a New Tycoon, W. W. Norton, October 2023 — long-form narrative based on hundreds of hours of access to Bankman-Fried and FTX executives; used here for chronology and internal atmosphere only, noting that reviewers have critiqued the book as overly sympathetic.
  3. Zeke Faux, Number Go Up: Inside Crypto's Wild Rise and Staggering Fall, Crown Currency, September 2023 — investigative journalism on FTX, Tether, and the broader crypto industry context.
  4. Chloe Cornish and Joshua Oliver, FTX coverage archive, Financial Times, November 2022 – March 2024 — aggregated contemporaneous investigative and court reporting.

Tertiary (flagged):

  1. Wikipedia, "FTX," "Bankruptcy of FTX," and "Trial of Sam Bankman-Fried," used for cross-checking dates and structural facts only; not load-bearing.

3. OTA narrative

Observe. The observational signals bearing on FTX were industry-available and largely visible to the exchange's operators. Alameda's concentration in an affiliated exchange token (FTT), the existence of a bespoke software exemption permitting Alameda to carry a negative balance on FTX untethered from collateral requirements, the scale of intercompany lending, and the deterioration of Alameda's liquidity during the May–June 2022 crypto drawdown were all observed inside the firm — the SEC complaint alleges the code creating Alameda's privileged line of credit was written in or around August 2019 and updated in or around May 2020, and Caroline Ellison testified she altered balance sheets in mid-2022 precisely because she could see the position. The CoinDesk balance-sheet leak of 2 November 2022 then made the essential signal public. Observe was not a root cause of the collapse; the apparatus produced the signals it should have produced, and insiders could see them. Observe functioned as a transmission step: the signals existed, were read internally, and reached external counterparties in time to trigger the run — but the failure lay downstream of seeing.

Think. The reasoning failure is the root cause of this episode. The interpretive and decision-making apparatus at FTX and Alameda treated customer deposits as available working capital for the affiliated trading book, and treated basic fiduciary and accounting norms — segregation of customer assets, arm's-length intercompany terms, truthful investor disclosure, functioning board oversight, reconciled books — as optional rather than binding. The correct framework was not esoteric: it is the standard-of-care framework for any custodial financial platform and was explicitly represented to customers and equity investors in FTX's own marketing materials and terms of service. The reasoning failure was therefore an Easy-Wrong Think at the easy end of the task-difficulty axis: the framework existed, was accessible, was represented to outsiders as the framework in use, and was deliberately not applied internally. This is not a case of a hard interpretive problem beyond the peer group's reach; regulated exchanges and custodians across the conventional finance peer group operate the very controls whose absence John J. Ray III flagged as a "complete failure of corporate controls." The reasoning step that converted observation into action selected misappropriation over segregation, and did so repeatedly from inception.

Act. Execution was technically competent in the narrow sense that what the leadership decided to do, the organisation was able to carry out: the privileged code path for Alameda's account was built, customer funds were moved, balance sheets were restated, equity rounds were closed, marketing campaigns were run, and political and venture disbursements were made. Act was not the root cause of the collapse; the operational capability executed faithfully on the reasoning above it. Where execution failed in the final days — the inability to meet the withdrawal run, the abandoned Binance rescue, the chaotic handover — it failed because the reasoning had already committed the firm to a position from which no competent execution could recover once the liquidity mismatch became public. Act was the transmission step between a corrupted decision layer and the external outcome; it was neither the locus of the failure nor a candidate for rescuing the episode through better performance.

4. Modality evidence

Direction. The foundational directional choice in this episode was made at the moment Bankman-Fried instructed Gary Wang to write the "allow_negative" code on 31 July 2019 — within weeks of FTX opening for business in May 2019 — enabling Alameda's FTX account to carry an unlimited negative balance (Wang trial testimony). This was not an operational improvisation: Wang testified that Bankman-Fried ordered the change specifically so that Alameda could access customer funds to cover FTX-token launch expenses, and that the initial credit limit of $1 billion was set deliberately high to "never be hit" before eventually rising to $65 billion (Wang trial testimony; SEC complaint). The choice encoded a direction: FTX would operate not as a segregated custodial exchange but as a funding vehicle for an affiliated trading firm, with the two entities treated as financially unified at Bankman-Fried's discretion.

A second directional decision reinforced this trajectory: in September 2021, Bankman-Fried relocated FTX's entire senior staff from Hong Kong to Nassau, Bahamas, citing the regulatory environment. Caroline Ellison wrote at the time that the move was "due primarily to the friendly regulatory environment" (Faux, Number Go Up; FT archive). The Bahamas move positioned the enterprise in a jurisdiction where domestic regulators lacked the reach and resources of the SEC or CFTC, and where no functioning corporate governance requirements would force the separation of customer assets that U.S. or EU regulation would have demanded. Both choices — the July 2019 code and the September 2021 relocation — meet the Direction Evidence Rule's specificity, timing, and attribution tests: each is a discrete, datable decision attributable to Bankman-Fried by named witnesses.

Structure. The governance architecture of FTX was, by design, one that provided no independent check on Bankman-Fried's control. The company's board consisted of Bankman-Fried and two FTX employees; it held no board meetings; investor requests for independent board seats were repeatedly rejected by Bankman-Fried on the grounds that their ownership was too small to warrant representation (Ray III declaration; FT archive). The Ray III declaration, filed 17 November 2022, stated that FTX operated with "virtually no internal controls and no separateness whatsoever" between FTX and Alameda, and that the absence of an independent board was a root feature, not an oversight gap (Ray III declaration).

Audit arrangements were structurally deficient in a way that compounded the board absence. FTX Trading Ltd. engaged Prager Metis CPAs as its auditor; the SEC's 2024 enforcement action found that Prager Metis lacked the competence and independence to perform those audits and that the firm did not investigate adequately after FTX disclosed intercompany loans to Alameda (SEC Press Release 2024-133). Accounting records for a multi-billion-dollar enterprise were maintained in QuickBooks, personal messages, and Slack channels; no daily reconciliation of blockchain positions occurred (Ray III declaration). The Fraud Case Structure-Culture Rule requires these governance and structural mechanisms to be treated as a distinct modality contribution from the normative and behavioural mechanisms covered under Culture. The structural contribution here is specific: the absent board, the captured audit function, and the QuickBooks-based accounting created formal channels incapable of surfacing the Alameda intermingling to any independent party — which is the "can't communicate" side of the Structure/Culture boundary test.

Processes. The operational machinery that should have connected FTX's custodial obligations to its execution failed at every standard control point. No customer-asset segregation process existed: Alameda's FTX account operated under the "allow_negative" exemption with no collateral requirement, no margin call, and no auto-liquidation trigger that applied to any other customer (Ray III declaration; Wang trial testimony). No intercompany reconciliation process existed to track the growing Alameda deficit: Singh testified that by June 2022, Alameda had built up a $2.7 billion deficit on the FTX platform, and that a "buggy accounting system" contributed to the deficit not being visible to FTX's own engineering team until September 2022, at which point the hole was $8 billion (Singh trial testimony).

The processes used to manage external counterparty relationships were deliberately deceptive rather than merely absent. Caroline Ellison prepared seven alternative versions of Alameda's balance sheet in mid-2022, selecting version seven as the one to present to lenders because it moved the FTX intercompany borrowing into long-term liability categories and omitted the scale of the position; Bankman-Fried reviewed and approved the approach (Ellison trial testimony). The absence of a process distinguishing normal treasury management from misappropriation — and the active substitution of manipulated balance sheets for accurate reporting — is a Processes-level failure in the specific sense that the machinery for truthful financial disclosure did not exist and was replaced by a fabrication routine.

Scoring note (zero-modality rationale): the Processes contribution described in this subsection is classified at the boundary with Capability per the methodology §3 Processes / Capability replacement test ("if the current operating staff were replaced by new hires of comparable background, would the operational pattern survive?"). The §4 evidence applies the test explicitly and concludes that the strategic weight sits on the Capability side — the operational edge depends on the specific individuals and tacit judgement carrying it, not on documented routine. The Processes component is acknowledged in narrative but does not carry standalone weight; both modalities are evidenced and the boundary call is recorded in the audit trail. Categorisation under METHODOLOGY-ota-scoring-v4.md §5: classification boundary with an adjacent modality.

Capability. FTX and its senior technical staff demonstrated substantial engineering and product capability: the exchange was built quickly, scaled to handle significant derivatives volume, and executed a rapid sequence of equity raises at escalating valuations that required credible investor presentations and product demonstrations (DOJ indictment; Faux, Number Go Up). Gary Wang and Nishad Singh possessed the software engineering capability to build the privileged code path for Alameda and to maintain a platform that competitive market participants used in preference to alternatives (Wang trial testimony; Singh trial testimony). The capability the organisation lacked was not technical: it was the regulatory, accounting, and compliance competence required to operate a custodial platform at the standards that FTX publicly represented to customers and investors. The SEC's enforcement against Prager Metis confirmed that even FTX's external audit engagement failed to supply the independent accounting competence that the structural absence of internal controls made necessary (SEC Press Release 2024-133). Whether this capability gap constitutes a genuine organisational incapacity or a deliberate choice not to build compliance infrastructure is a question the evidence does not resolve cleanly; Caroline Ellison's desire to step down — which she suppressed out of fear of triggering a bank run — suggests that some individuals within the firm understood the liability they faced (Ellison trial testimony).

Scoring note (zero-modality rationale): the capability described in this subsection is recorded at zero per cent in the modality weights on the rationale of insufficient causal weight — the §4 evidence establishes that FTX / Sam Bankman-Fried possessed the technical and operational capability the situation required; the failure mechanism was located in Direction, Structure, Culture rather than in a capability gap. The capability is acknowledged as present in the narrative but does not carry standalone weight in the failure attribution. Categorisation under METHODOLOGY-ota-scoring-v4.md §5 "Zero-modality rationale rule": insufficient causal weight.

Culture. The Fraud Case Structure-Culture Rule requires Culture to be scored on the normative and behavioural mechanisms that motivated and sustained the fraud, distinct from the structural governance mechanisms above. The cultural evidence in this case runs in multiple directions. Bankman-Fried's public persona was built on a specific normative register — Effective Altruism, utilitarian ethics, responsible crypto stewardship — that served as both a marketing instrument and an internal authority claim; it created a cultural context in which questioning the leader's ethical reasoning required arguing against his framework on its own terms (Lewis, Going Infinite; Faux, Number Go Up). When Singh confronted Bankman-Fried in September 2022 on the balcony of their Nassau penthouse and raised the $13 billion liability, Bankman-Fried dismissed the concern, telling Singh he planned to cut expenses and that FTX.US expansion would replenish the shortfall; Singh testified that he was "blindsided and horrified" (Singh trial testimony). Ellison testified that she had wanted to leave Alameda before the collapse but feared that her departure would trigger a bank run, and that she prepared the seven alternative balance sheets because both she and Bankman-Fried "thought the things being concealed were bad" — a statement of shared normative awareness, not inadvertent error (Ellison trial testimony).

The inner circle's eventual cooperation with prosecutors — Wang, Singh, and Ellison all pleaded guilty and testified against Bankman-Fried — suggests that the behavioural defaults within the group did not require active coercion to sustain the fraud; the individuals understood what they were doing and participated. This is characteristic of a cultural pattern in which a dominant founder's authority and the group's shared project (building a successful exchange, pursuing Effective Altruism goals) created sufficient legitimation for individual compliance over an extended period (DOJ indictment; Lewis, Going Infinite; Faux, Number Go Up).


Cite this case: OTA-200 Study, Case F-009 (FTX / Sam Bankman-Fried — collapse of a crypto exchange), methodology v4. Read and cite with attribution; no redistribution or commercial reuse — License & Terms.

Spotted an error? Report a correction for F-009. Implemented corrections are published and credited in the Corrections Log.