A Historical Overview of Major Crypto Exchange Failures

Research question and scope

This article synthesizes commonly documented patterns behind historical crypto exchange failures and compares each pattern against a commonly cited mitigating practice. It is historical, educational synthesis only — not an assessment of any current platform, and not financial advice.

Key findings

• Fund commingling, disclosure gaps, and concentrated authority recur across multiple, independently documented historical exchange failures — not isolated, unrelated incidents.
• Each documented failure pattern has a commonly cited corresponding mitigating practice, though no single practice eliminates custodial risk entirely.
• These patterns are useful for evaluating custodial risk in general, independent of and without naming any specific platform operating today.

Why this history matters

Crypto's history includes several major, well-documented exchange failures. Studying the general, publicly reported patterns behind them is a useful educational exercise for understanding custodial risk, independent of any specific current platform.

Commonly documented failure patterns

• Commingling of customer funds with the exchange's own operating or trading funds, rather than keeping them clearly segregated.
• Insufficient or misleading public disclosure about the exchange's actual reserve holdings relative to customer balances.
• Concentration of decision-making authority with limited independent oversight or internal controls.

What this suggests for evaluating a custodial platform

Publicly documented proof-of-reserves practices, clear segregation of customer funds, and independent oversight are commonly cited as practices that address these historical failure patterns — though no disclosure practice alone can fully eliminate custodial risk.

Evidence base: what a large-sample academic study found

Beyond individually documented case patterns, a 2025 peer-reviewed study (Sapkota, University of Vaasa, published in the Journal of International Financial Markets, Institutions & Money) analyzed 845 cryptocurrency exchanges empirically rather than narratively, using logit and probit statistical models to predict which exchanges would fail. The study reports those models predicted exchange bankruptcies with roughly 81% accuracy -- a notable finding, since it suggests exchange fragility correlates with identifiable, measurable structural factors rather than being purely idiosyncratic or unpredictable case by case. The study specifically points to pressures from stringent regulations, high compliance costs, and advanced infrastructure that bad actors can exploit as contributing factors -- a more structurally specific framing than "concentrated authority" alone captures. Separately, governance-focused academic work (including jurisdiction-comparison research covering over 250 exchanges) finds a strong correspondence between how permissive an exchange's primary regulatory jurisdiction is and its operational risk level, reinforcing that jurisdictional choice is itself a measurable risk factor, not just a legal formality.

Independent critical analysis

The three commonly cited failure patterns in this article's own table (commingling, disclosure gaps, concentrated authority) are usually presented as roughly co-equal, independent risk factors. The academic evidence above suggests a different, more structural reading: jurisdiction choice appears to function as something closer to an upstream cause that makes the other three more likely, rather than a fourth independent item on the same list. A permissive jurisdiction with weak enforcement capacity plausibly makes commingling easier to get away with, disclosure gaps easier to leave unaddressed, and concentrated authority easier to sustain without external challenge -- meaning an evaluator who only checks for the three downstream symptoms, without also weighing the exchange's regulatory jurisdiction and its enforcement track record, may be assessing effects while missing a shared upstream cause. This is an inference drawn from the cited jurisdiction-risk correlation, not itself a tested causal claim in the underlying research.

Comparing failure patterns to mitigating practices

Limitations of this comparison

• This describes general historical patterns and commonly cited mitigations, not a complete history, and not an audit or endorsement of any current platform's actual practices.
• A mitigating practice being commonly cited doesn't mean it fully eliminates custodial risk — no disclosure practice alone has been documented to do so.
• The academic sources cited above were found and summarized via WebSearch synthesis, not independently read in full text -- the Sapkota paper's exact regression specification and the GMU jurisdiction-risk papers' precise methodology were not independently verified, and the GMU papers are noted in secondary reporting as undergraduate-level work, included here as suggestive evidence, not definitive peer-reviewed consensus.
• The jurisdiction-as-upstream-cause argument in the independent analysis above is this article's own inference from the cited correlation, not a claim the underlying studies make explicitly.
• This is general education, not financial advice about where to hold assets.

Research conclusion

Major historical crypto exchange failures share commonly documented patterns — fund commingling, disclosure gaps, and concentrated authority — and a large-sample academic study now provides empirical support that exchange fragility is measurably predictable (roughly 81% bankruptcy-prediction accuracy in one model) rather than purely idiosyncratic. General, historical education relevant to understanding custodial risk broadly, not a statement about any current platform.

References

• Primary (academic, found via WebSearch, not directly fetched in full): Sapkota, N., "Why do cryptocurrency exchanges fail?", Journal of International Financial Markets, Institutions & Money (2025), doi.org/10.1016/j.intfin.2024.102093 -- the 845-exchange sample and ~81% bankruptcy-prediction accuracy figure.
• Secondary (academic, undergraduate-level, found via WebSearch): two George Mason University student research papers on exchange governance structure and jurisdictional permissiveness as risk correlates (journals.gmu.edu/jssr), covering over 250 exchanges.
• Secondary (academic, found via WebSearch): a UCL working paper on crypto-exchange institutional architecture, custodial opacity, and maturity mismatches (discovery.ucl.ac.uk).
• For a primary-source, named-case treatment of specific historical failures with exact figures, see Kaavon's own separate research article on named-case exchange failures (FTX, Celsius, Mt. Gox, QuadrigaCX).