Analytics
Which Crypto Firms Went Bankrupt, and Which Became AI Powerhouses
A living tracker of every major crypto bankruptcy: who collapsed, how much creditors recovered, where each token trades now, and the miners reborn as AI datacenters.
Key Takeaways
- Thirteen firms, two waves, one benchmark. Thirteen major crypto companies filed for bankruptcy across the 2022-23 cluster and the 2026 wave. Mt.Gox (2014) is the yardstick.
- "100%+ recovery" hides a loss. FTX and Celsius repay in petition-date dollars, so creditors missed the later rally. Only Genesis paid back actual coins.
- Bankruptcy thins a token more often than it kills it. STORJ, FTT and MOVE still trade actively, while CEL and VGX cling on across just a handful of venues.
- Two bankrupt names never had a token. Any "GEN" or "BFI" trading today is an unrelated project, not Genesis Global Capital or BlockFi.
- Some bankrupt miners became AI datacenters. Core Scientific rebuilt after Chapter 11. IREN and TeraWulf did it without ever filing.
- The next trigger is the Mt.Gox repayment deadline. It falls on Oct 31, 2026, now in its third postponement.
Every Major Crypto Bankruptcy, From FTX to Movement Labs
Thirteen major crypto firms filed for bankruptcy between 2022 and 2026, from FTX's $10B-plus collapse to Movement Labs' Chapter 11 in July 2026. The table below pairs each filing with its debt, creditor recovery, capital raised, and where the token trades today. Mt.Gox (2014) anchors the comparison as the original benchmark.
| Company | Filed / Chapter | Debt at filing | Recovery / status | Raised before collapse | Token today | Company today |
|---|---|---|---|---|---|---|
| FTX | Nov 11 2022 · Ch.11 · D. Del. (22-11068) | $10-50B range | ~119% allowed (petition-date USD) | ~$1.73B-$2B / ~72 investors | FTT still trades across dozens of markets | Liquidating via FTX Recovery Trust |
| Celsius | Jul 13 2022 · Ch.11 · SDNY | >$4.7B | 35-45% liquid; 67-85% plan; ~64.9% cumulative (Aug 2025) | ~$750M Series B (2021) | CEL still trades thinly on a handful of venues | Wound down; Mashinsky convicted |
| Genesis Global Capital | Jan 19 2023 · Ch.11 · SDNY | $1.2-11B range | ~64% avg in-kind; up to 77% | None (DCG unit) | No native token | Wind-down; DCG litigation |
| BlockFi | Nov 28 2022 · Ch.11 · D.N.J. (22-19361) | $1-10B range | 100% eligible after $874.5M FTX settlement | ~$1B+ across ~9 rounds; Series E $500M (~$4.75B val) | No native token | Liquidated |
| Voyager | Jul 6 2022 · Ch.11 · SDNY | $1-10B range | ~$1.33B returned; unsecured ~72% | Alameda $75M, $60M and a $200M line | VGX still trades thinly on a handful of venues; holders repaid in USDC | Liquidated |
| Terraform Labs (Terra/LUNA) | Jan 21 2024 · Ch.11 · D. Del. (24-10070) | ~$4.47B (SEC settlement) | Wind-down per SEC plan; estate waterfall | ~$134.5M across 5 rounds | LUNC still trades across dozens of markets | Wound down; Do Kwon extradited to US |
| Three Arrows (3AC) | Jul 1 2022 · Ch.15 · SDNY + BVI | $3.4B / 154 claims | ~35.16% (Teneo, Mar 2026), down from ~46% (Dec 2023) | Hedge fund; peak ~$10B AUM | No token | Offshore liquidation (Teneo) |
| Core Scientific | Dec 2022 · Ch.11 (emerged Jan 2024) | Cut ~$400M (note to equity) | Emerged going concern | N/A (public miner) | N/A (CORZ equity) | Operating; AI/HPC pivot |
| Storj Labs | Jul 26 2026 · Ch.11 · N.D.W.Va. | Not disclosed | Restructuring; reorg ownership shared with token holders | ~$35M | STORJ still trades across dozens of markets | Operating; backed by Inveniam |
| BlockFills | Mar 15 2026 · Ch.11 · D. Del. | $100-500M liab | Assets sold to Keyrock | Not disclosed | No retail token | ~$75M losses; assets sold |
| Goliath Ventures | Mar 2026 · receivership + Ch.11 | $400M inflows / ≥$250M admitted / $328M civil | Criminal forfeiture | ~$328M from 2,000+ victims | N/A (fraud) | Fraud; CEO guilty Jul 2026 |
| Bitcoin Depot | May 17 2026 · Ch.11 | Not disclosed | Full shutdown; 9,276 kiosks offline | N/A | N/A | Wound down |
| Movement Labs | Jul 15 2026 · Ch.11 Subchapter V · D. Del. (26-11113-TMH) | assets $100-500K / liab $1-10M | Restructuring (chain + Foundation excluded) | ~$141.4M | MOVE still trades across dozens of markets | Reorg; chain still processing |
| Mt.Gox (benchmark) | 2014 · Japan civil rehab | 650k-850k BTC lost | ~19,500 creditors repaid; estate ~34,388 BTC; deadline Oct 31 2026 | N/A | N/A | Rehabilitation ongoing |
Source: court dockets and estate reports; fundraising and trading-status data, DropsTab. As of 2026-10-01.
According to DropsTab fundraising data, Storj Labs raised about $35M across five rounds (a ~$30M STORJ token sale in 2017 plus ~$8M in equity) before filing. The token still trades across dozens of markets. The same pattern repeats across the table, where raising big tells you nothing about whether a token survives a filing.
How to read the timeline below:
- Top to bottom is time, drawn to scale from 2014 to 2026, so the empty years between failures stay empty.
- Each dot sits on its actual filing date; the tight 2022 knot is six collapses in under six months, the contagion cluster.
- Colour shows what became of the company: red is liquidated or gone, green is survived and still operating (Core Scientific, Storj), blue is still working through the courts (Mt.Gox, Movement).
What FTX, Celsius and Genesis Creditors Got Back
FTX and Celsius creditors are repaid in petition-date dollars, so the headline recovery figures exclude the later crypto rally. FTX's roughly 119% allowed recovery still left many holders short. Genesis is the counter-example: it paid creditors back in actual coins, capturing the price appreciation that the dollar-based estates froze out.
FTX pays through the FTX Recovery Trust. Its plan became effective on Jan 3, 2025, and claims were locked at the Nov 11, 2022 petition date, when BTC sat near $16,871. Five distribution rounds have run so far.
| Round | Date | Amount |
|---|---|---|
| 1 | Feb 18 2025 | ~$1.2B (convenience, 120%) |
| 2 | May 30 2025 | >$5B |
| 3 | Sep 30 2025 | ~$1.6B |
| 4 | Mar 31 2026 | ~$2.2B (classes 5B/6A/6B to 100%) |
| 5 | Jul 31 2026 | ~$900M (classes 5A/5B to 105%) |
Source: FTX Recovery Trust distribution announcements, as of 2026-10-01.
Cumulative payouts reach roughly $11B. The estate is valued near $16-17B as of 2026, and no sixth-round date is confirmed.
Celsius tells the same story in smaller numbers. Through its Celsius claims process, claims were locked at BTC $19,881 and ETH $1,088.17, with the plan effective Jan 31, 2024. At the petition date, 1 BTC plus 1 ETH was worth $20,969. A creditor at 57.9% recovery received $12,141 in value, against $45,550 for the same coins on the effective date.
Genesis went the other way. It distributed about $4B in crypto and cash from Aug 2, 2024, after the court rejected a petition-date price cap. Filing-date BTC near $24,000 later passed $66,000, so creditors received actual coins plus the appreciation. DCG received nothing, and live 2026 clawback suits seek over $1.2B from DCG and Barry Silbert.
| Estate | Valuation basis | Headline recovery | Upside captured? |
|---|---|---|---|
| FTX | USD at petition | 105% (5A/5B), 120% convenience | No |
| Celsius | USD at petition | 35-45% liquid; 67-85% plan | No |
| Genesis | In-kind coin | ~64% avg; up to 77% | Yes |
Source: estate distribution reports and court filings, as of 2026-10-01.
How to read the chart below:
- Each dot is a creditor recovery rate, ranked high to low; the bright vertical line marks 100%, fully made whole.
- Only FTX and BlockFi reach it. The faint track running from each dot to the line is the shortfall, the share creditors never got back.
- Above 100% still is not whole: FTX's roughly 119% is measured in petition-date dollars (hourglass marker), frozen before the later crypto rally, so holders lost the upside. Genesis (coins marker) paid in-kind and captured that rally.
Independent analyst Aakash Gupta framed the paradox on X:
FTX claims were fixed in dollars at the Nov 11, 2022 petition date, so the roughly 119% recovery excluded the later crypto rally.
One number frames the loss at the estate level. FTX's own reconstructed portfolio of early bets (Anthropic, Solana, SpaceX), worth about $4.7B, could be near $52.5B today, implying roughly $47.8B in theoretical estate upside forgone as of 2026. That figure belongs to the estate, not individual creditor claims.
The 2026 Bankruptcy Wave vs the 2022-23 Cluster
The 2022-23 cluster was a contagion chain: FTX, Celsius, Genesis, BlockFi, Voyager, 3AC and Core Scientific collapsed into each other. The 2026 wave is scattered by cause, from Storj's restructuring to Goliath's fraud. And most 2026 crypto closures were not bankruptcies at all.
The trigger was Terra. Its UST stablecoin depegged in May 2022 and drove LUNA to near-zero, cratering 3AC and rippling into Celsius and Voyager. Terraform Labs itself filed Chapter 11 in January 2024 and settled with the SEC for $4.47B. FTX was BlockFi's second-largest unsecured creditor at about $275M. Genesis Asia Pacific was 3AC's largest creditor at $2.3B. One failure pulled in the next.
The 2026 wave looks different. Storj (July 26) is a restructuring with a live network. BlockFills (March 15) was a trading firm with ~$75M in losses, sold to Keyrock. Goliath Ventures (March) was fraud, with CEO Delgado found guilty in July 2026. Movement Labs (July 15) filed Chapter 11 Subchapter V after a market-maker deal let a counterparty control its float, with roughly 66M MOVE dumped the day after the December 2024 token launch. Co-founder Rushi Manche left in May 2025. Others from the year include Archblock, BitRiver, AscendEX and Poolin.
RootData counts about 99 crypto projects closing in 2026, rising past 100 through August, but that figure blends bankruptcies with voluntary shutdowns. It is not 99 bankruptcies. Theft is a separate loss track entirely; large-scale thefts are tracked in our living crypto hacks hub. Most of the "failure wave" was solvent wind-down, settled by governance votes, not a court process with creditors. CoinEx is the clean example, announcing an orderly closure with reserves above 100%.
CoinEx says it is winding down with reserves above 100% and withdrawals open until Dec 22, 2026. That is a solvent closure, not a court bankruptcy.
Bitcoin Depot is a real bankruptcy with a regulatory cause. The Bitcoin-ATM operator filed on May 17, 2026 and took all 9,276 kiosks offline. Q1 2026 revenue was about $83.5M, down 49% year over year, with a $9.5M net loss. CEO Alex Holmes blamed a changed regulatory environment as several US states move to ban crypto ATMs. The machines charged fees up to about 20%, versus 0.4-1% online. The FBI's IC3 logged roughly 13,500 crypto-ATM complaints and about $388M in losses in 2025, and over half of those ATM-fraud losses came from victims aged 50 and older.
Mt.Gox remains the benchmark. A decade on, the estate still holds about 34,388 BTC, and roughly 19,500 creditors have been repaid. The repayment deadline, now in its third postponement, falls on Oct 31, 2026. On-chain analytics firm Arkham tracks the wallets:
Arkham's labelled wallets show the Mt.Gox estate still holding about 34,388 BTC, with the repayment deadline set for Oct 31, 2026.
How Bankrupt Miners Rebuilt as AI Datacenters
Only one bankrupt miner, Core Scientific, actually rebuilt itself as an AI datacenter after Chapter 11. IREN and TeraWulf reached the same AI pivot without ever filing for bankruptcy. No verified figure claims a set share of miners made this move. The pivot is documented in corporate filings, not a trend statistic.
Core Scientific (NASDAQ: CORZ) is the only one that filed. It entered Chapter 11 in December 2022 and emerged on Jan 23, 2024, cutting about $400M in debt by converting notes to equity, and relisted on Nasdaq the next day. A small CoreWeave lease scaled into roughly 590 MW across six sites and about $10.2B in contracted revenue over 12-year terms (SEC 8-K, Feb 26, 2025). Colocation revenue jumped from $10.6M in Q2 2025 to $136.7M in Q2 2026, about 83% of the total. Its Q2 2026 net loss of roughly $1.2B was driven by non-cash items, a ~$1.05B warrant and CVR remeasurement plus a ~$266.5M mining-asset impairment, not an operating loss. Adjusted EBITDA was positive at $41.1M. Shareholders rejected CoreWeave's ~$9B all-stock buyout on Oct 30, 2025. Core Scientific now trades as a Nasdaq stock rather than a token; traders who want on-chain exposure to equities like it can look at how to trade tokenized stocks.
IREN (formerly Iris Energy) never filed for bankruptcy. It IPO'd on Nasdaq in November 2021 and, in late 2022, let ring-fenced subsidiaries default on equipment loans without a parent-level filing. It has since signed a roughly $9.7B, five-year Microsoft AI cloud contract using NVIDIA GB300 hardware, and plans to grow its GPU fleet from 23,000 toward 140,000 by 2026.
TeraWulf (NASDAQ: WULF) also never filed. It went public via reverse merger in 2021, then signed deals with Core42, Fluidstack and an Abernathy, Texas joint venture, bringing roughly 510 MW under contract. Google really does stand behind TeraWulf, though only in a narrow role. It provides a $3.2B backstop and warrants on the Fluidstack deal, rather than buying hosting as a customer.
Does Bankruptcy Mean the Token Is Dead
Bankruptcy rarely kills a token outright. It thins liquidity across a spectrum, from STORJ and MOVE trading actively to CEL and VGX clinging to a few venues. FTT is the hollow case: liquid but backed by nothing. A separate loss mode hits creditors, whose frozen petition-date payouts forfeit the rally even when the headline recovery looks whole.
Think of survival as a spectrum of exchange breadth. At the active end, STORJ and MOVE both trade across dozens of markets. Storj also still runs a live, revenue-generating storage network whose plan proposes sharing reorg ownership with token holders.
All five sit side by side in a DropsTab custom tab, "Bankrupt but Still Trading." It lines up each token's price, market cap, venue count and 24h volume on one screen, and you can build a tab like it for any watchlist.
FTT sits in a stranger spot. It stays liquid across dozens of markets with no exchange, no utility and no backer behind it, a warning rather than an endorsement. At the thin end, CEL and VGX both cling to just a handful of venues each, with VGX holders repaid in USDC. The entities are gone, but the tickers hang on. Bankruptcy degrades a token's liquidity far more often than it kills it outright. Financial collapse is only one way a token can fade; a longer-horizon risk to even healthy chains is the arrival of quantum computing, covered in our guide to quantum-resistant crypto.
A separate loss mode hits creditors directly. FTX and Celsius holders were "paid back" in frozen petition-date dollars and forfeited the appreciation. Genesis showed this is avoidable with in-kind payouts.
Bankruptcy Myths and False Alarms
The most-cited denial of bankruptcy comes from inside FTX. Convicted FTX founder Sam Bankman-Fried has publicly claimed FTX "was never bankrupt" and that lawyers filed a "bogus bankruptcy" to loot the company. The claim is false. FTX filed Chapter 11 on Nov 11, 2022 (D. Del., No. 22-11068), and Bankman-Fried was sentenced to 25 years for fraud. A separate post of his claiming FTX was solvent carries a Community Note stating FTX was not always solvent.
One guardrail matters for anyone searching tickers. Bankrupt Genesis Global Capital and BlockFi never issued a retail token. The "GEN" and "BlockFi-AI (BFI)" tokens trading today are unrelated namesake projects, not claims on either estate.
Two fear-queries deserve a flat answer. Coinbase (NASDAQ: COIN) has not filed for bankruptcy. The fear traces to a Form 10-Q disclosure (SAB 121) that custodial crypto could be treated as estate property in a hypothetical bankruptcy. That is a risk disclosure, not a filing. An August 2026 New Jersey ruling let an investor suit over that disclosure proceed, but it is still not a bankruptcy. Coinbase also cut about 14% of staff (~700 roles) in May 2026, citing the downturn and AI automation, with $50-60M in charges, and it remains solvent. CEO Brian Armstrong framed the cut on X:
Coinbase's CEO framed the roughly 14% staff cut in May 2026 as an AI-era restructuring by a company he describes as well-capitalized. It is not a bankruptcy filing.
MicroStrategy, now Strategy (NASDAQ: MSTR), has not filed either. A viral "Chapter 11 in 2026" claim was an AI-generated hypothetical. Its holdings of roughly 712,647 BTC far exceed about $8.2B in debt, where the debt is fixed but the BTC value is volatile. Prediction markets agree. On Polymarket, the odds of a MicroStrategy bankruptcy before 2027 fell from about 40% at the end of 2025 to roughly 3% as of 2026-10-01.
| Event | Yes | No | Volume | Expires |
|---|---|---|---|---|
| MicroStrategy announces bankruptcy before 2027 | ~3% | ~97% | ~$205K | Jan 1, 2027 |
Source: Polymarket, 2026-10-01. Market-implied probability, not a forecast.
How to Track Every Crypto Bankruptcy and Where Its Token Stands
Bankruptcy filings, creditor payouts and token fate each move on their own timeline, and the headlines rarely keep up. DropsTab pairs every collapsed project's fundraising history with its live coin page, so you can see what a company raised and whether its token still trades in one place. For the open questions, like whether MicroStrategy ever files, Drops Bot streams the prediction-market odds straight to Telegram so you can watch the probability move: t.me/Drops?start=pm_microstrategy-bankruptcy.