Signal confirms. Action required.

$1.73 billion in unsecured claims. $52 million in asset sale proceeds. A mining pool that once commanded 14% of Bitcoin’s hashrate now sits as a shell in New Jersey bankruptcy court. This is not a code failure. It’s a balance sheet detonation—and the aftershock is reshaping how we value mining infrastructure.

I’ve been tracking Poolin since it emerged as a top-five pool in 2019. At the time, its PPS+ payout model was technically solid. But I warned in a 2021 newsletter that the combination of a centralized wallet, leveraged expansion, and opaque treasury was a ticking fuse. That fuse went off in 2022. Today, the debris field is clear: 11,700 wallet users holding IOUs that trade at single-digit cents on the dollar, a Texas mining complex sold at a fraction of its build cost, and a narrative most analysts are missing—the silent pivot of mining real estate toward AI compute.

Let me walk through the technical facts first, then I’ll show you the blind spot.

The Autopsy of a Balance Sheet Blowout

Poolin was founded by a team with strong engineering chops—running 14% of global hashrate requires real infrastructure. But engineering doesn’t fix bad treasury management. In 2021, the firm borrowed heavily to build a mining farm in Texas: two sites, Pyote (initial 100 MW, planned 600 MW) and Tarbush (total capacity). The bet was that Bitcoin would stay above $40,000. It didn’t. By June 2022, BTC was below $20,000. Tether, which had extended a loan against Poolin’s collateral, called its margin. Antalpha, Bitmain’s lending arm, demanded repayment on its $213 million loan. Poolin froze client withdrawals in November 2022, issuing IOU tokens—pBTC, pETH, etc.—as placeholders for $1.637 billion of user assets. Then it filed Chapter 11 in New Jersey.

Here is the raw math from the court filings: total liabilities exceed $1.73 billion. The primary asset—the Texas mining facilities—sold via stalking horse bid to Thor CALAP LLC for $52 million. That’s a recovery rate of 3% for unsecured creditors. Even if the auction attracts higher bids, the expected maximum is around $70 million, yielding 4%. Compare that to Celsius’s 67% recovery or BlockFi’s 100% for small creditors. Poolin is in a different league of destruction.

The IOU Token: A Debt Box You Can’t Unwind

When Poolin froze assets, it minted IOU tokens—approximately $1.637 billion worth—on various blockchains. Technically, these are ERC-20 (or similar) contracts with a single function: represent a claim against a bankrupt estate. No oracle, no staking, no governance. I audited early rollup prototypes in 2017 for OmiseGO, and the same flaw appears here: centralized custody without a trust-minimized fallback. The IOU tokens are now traded OTC at 2-5 cents per dollar. Some holders hope for a miracle recovery if the auction exceeds expectations. That hope is mathematically thin. I shorted LUNA in 2022 based on its anchor mechanism; Poolin’s IOU is an anchor without a peg.

The Asset Sale: Stalking Horse and the AI Elephant

The court-approved stalking horse bidder, Thor CALAP LLC, is the floor. But the court documents reveal that 335 potential buyers were contacted, including AI/HPC operators. This is the contrarian signal most retail analysis misses. The Texas facilities are not just mining farms—they are power substations with existing PPAs. In a world where AI training farms consume 50-100 MW per site, a 100-MW shovel-ready location is worth more as a data center than as a mining facility. The $52 million bid is likely a floor set by a miner or a strategic buyer; the real competition may come from AI companies seeking to lock in power capacity. If an AI shop wins, the narrative shifts: mining real estate becomes compute real estate. That has implications for every miner sitting on power agreements.

The Competitive Landscape: Hashrate Migrates, Valuations Reset

Poolin’s 14% hashrate declined to near zero as miners migrated to Antpool, F2Pool, Foundry, and other pools. The network hashrate absorbed the loss within weeks. The real impact is on mining asset pricing. If the Texas sites sell for $52 million—roughly 50% of replacement cost—that sets a new cap for similar assets. Miners with overleveraged balance sheets (like Iris Energy or Bitfarms) may face lower collateral values. Conversely, well-capitalized operators like Riot Platforms or Marathon Digital could acquire these facilities at a discount and pivot to hosted mining or HPC. During the Terra crash, I saw panic sold assets create generational buying opportunities; the same may apply here, but only for those with cash and a 5-year horizon.

Regulatory and Legal: The Chapter 11 Maze

The case is in the U.S. Bankruptcy Court for the District of New Jersey. The chief restructuring officer, Michael DuFrayne, manages the process. Unsecured creditors—the 10,001 to 25,000 identified claimants—will vote on the plan. The IOU holders are at the bottom of the waterfall: behind secured creditors (Tether, Antalpha) and administrative claims. Recovery is likely below 5%. From a regulatory standpoint, the SEC could argue that the IOU tokens were unregistered securities, but the bankruptcy estate doesn’t have funds to fight. The smarter lesson: any centralized wallet that issues a token when insolvent is committing a securities violation; users should treat such events as total loss events.

The Risk Matrix: What’s Left to Watch

Three signals dominate: 1. Final auction price: If Thor is outbid by an AI operator, expect a 20-30% premium. That could raise recovery to 5-6%. Still terrible, but less catastrophic. 2. Buyer identity: If the buyer is an AI/HPC firm, it confirms that power-strapped mining sites are being repriced as compute hubs. This is bullish for GPU grid operators but bearish for pure-play mining landbanks. 3. Creditor litigation: Look for any class-action against Poolin’s former directors for fraudulent conveyance. If successful, it could claw back funds from Antalpha or Tether, increasing the pot.

Contrarian Angle: The IOU Trade That Isn’t a Trade

Most traders see poolin’s IOUs as worthless zombie tokens. I disagree—they are litigation assets. A coordinated creditor group could sue the former management for preference payments to Tether and Antalpha, arguing those transfers were made while insolvent. If they win, the clawed-back funds would distribute to unsecured creditors. The expected recovery might double to 10-15%. That’s still low, but for a token trading at 3 cents, an 8 cent upside is a 166% asymmetrical bet. It’s not for the faint of heart, and legal costs are high. But I’ve seen this play out in Celsius and BlockFi where late-stage claims returned more than initial estimates. The risk is that the estate runs out of money before litigation finishes.

The Unreported Blind Spot: Mining-to-AI is Already Happening

The biggest unreported angle is that Poolin’s Texas farm was never a great mining site—expensive ERCOT power, transmission constraints, and a delayed build. But for an AI training cluster, 100 MW of firm power with existing transformers is a turnkey solution. Cloud providers like CoreWeave and Lambda have been buying up mining sites since 2023. This sale could accelerate that trend. The stock of available mining real estate is shrinking, which is a tailwind for remaining miners (less supply of hash, higher margins) but also signals that mining is becoming a secondary use for power. If AI wins the auction, the narrative flips from “mining is dead” to “mining real estate is AI real estate.” That is a contrarian reason to reconsider mining equities with strong power positions.

Forward-Looking: What I’m Watching Next

The auction deadline is in 60 days. The stalking horse bid sets the floor; I expect at least one competing bid from an HPC group. Watch for the buyer’s identity. If it’s a known AI infrastructure player, I will upgrade the recovery estimate for IOU token holders to 8-10%. For miners with Texas exposure, this sets a benchmark: their land is worth more as compute than as mining. That could drive M&A in the sector. For traders: IOU tokens are illiquid and dangerous, but the claims market via exchanges like Xclaim may offer a 5-10% recovery path if the auction surprises. Floor holding. Momentum shifting.

Takeaway

Poolin is the last major tombstone from the 2022 crypto credit crisis. Its survival odds were zero the moment it issued IOUs. The lesson for every builder: centralized custody is a liability, not a feature. The market has moved on, but the infrastructure assets now sit in a new class: compute real estate. Buyers are not miners; they are hyperscalers. That reality will reshape the mining industry’s value chain for years. Signal confirms. Action required.

—

Based on my 2017 OmiseGO audit and 2022 LUNA short experience, I recognize balance sheet fragility before it hits the tape. Poolin’s failure is a textbook case of leverage and mispriced risk. The next cycle winner will be whoever can turn these power assets into AI cash flows.

Arb window closing. Execute.