The Number

NFN8 Group filed Chapter 11 in February 2026. Over 5,000 ASICs. More than 250 equipment lessors. A fire that took out half their capacity before the filing. A $2.75M DIP facility from Twelve Bridge Capital.

Total estimated asset value for the machines: under $50,000.

That's roughly $10 per unit for equipment that traded at $5,000–$10,000 each during the bull market. The collapse ratio — peak to trough — exceeds 99%.

Twenty-five years restructuring hard assets at Lehman Brothers and Lone Star Funds, and this is unlike anything in traditional distressed debt. Not because Bitcoin is exotic. Because ASICs violate every assumption that asset-based lending relies on.

What Makes Collateral Work

Traditional asset-based lending rests on a simple premise: when the business fails, the collateral retains enough independent value to repay the lender. Oil reserves still sit in the ground when the E&P company goes bankrupt. Real estate has a land floor. Even specialised industrial equipment typically recovers 20–40% of replacement cost in a forced sale.

The collateral doesn't need to hold its peak value. It needs a floor — a price below which it cannot fall because the asset has intrinsic utility independent of the borrower's business performance.

ASICs have no such floor.

An ASIC performs one function: SHA-256 hashing. It cannot be repurposed. It cannot be redeployed to a different blockchain. Its economic value is entirely a function of two external variables — hashprice and the operator's electricity cost — neither of which is controlled by the borrower or the lender.

When hashprice drops or difficulty spikes, the machine's economic value craters while it's still physically sitting in the rack, plugged in, drawing power. A running ASIC can be worth less than the electricity it consumes. That's not depreciation. That's negative residual value — something traditional ABL models don't contemplate.

The Depreciation Problem Is Worse Than You Think

Most capital equipment depreciates along a predictable curve. Industrial machinery might lose 15–20% per year. Vehicles follow well-established residual value tables. Even technology hardware — servers, networking gear — retains meaningful second-hand value because it serves general-purpose computing needs.

ASICs depreciate on two axes simultaneously.

In USD terms, they follow a technology obsolescence curve as newer, more efficient machines enter the market. Each generation improvement in joules-per-terahash makes the prior generation less competitive. But unlike general-purpose computing, there is no secondary market for "good enough" performance. A miner either operates above breakeven or it doesn't. There is no middle ground.

In BTC terms, the depreciation is even more severe. As network difficulty rises, each machine's share of block rewards shrinks. The hashrate it contributes becomes a smaller fraction of the total, generating fewer satoshis per day regardless of the BTC/USD exchange rate. This means a miner can be unprofitable in BTC terms even during a bull market if difficulty has risen sufficiently.

The interaction between these two depreciation curves creates a collapse profile that no traditional lending model captures. During the 2022 bear market, ASIC values fell 85–91% — and this was not an outlier. It was the pattern repeating.

What TradFi Misses When Valuing Mining Assets

Traditional finance analysts approaching a distressed mining asset typically evaluate it like any other industrial operation: revenue, costs, assets, liabilities. This framework misses several critical variables unique to Bitcoin mining.

Current hashprice and its trajectory matter more than any balance sheet item. Network difficulty — which hit its largest absolute increase in Bitcoin's history on February 19-20, 2026, at +14.73% — determines whether any given machine is economically viable. Machine efficiency measured in joules per terahash determines which operators survive a difficulty spike. Site-level power cost measured in dollars per kilowatt-hour is the single most important variable in a miner's survival calculus. And the interaction between all of these — not any single metric — determines whether an operation has a future.

A distressed mining asset with $0.025/kWh power and latest-generation machines is worth vastly more than one with $0.07/kWh power and the same machines. The machines are almost irrelevant. The energy position is everything.

The Lending Model That Works

Mining credit works — but only if you underwrite the right things.

The energy position is the real collateral. An operator's power purchase agreement, their access to stranded or curtailed energy, their ability to flex load up and down in response to grid conditions — these are the assets that retain value through cycles. A long-term power contract at sub-$0.03/kWh has value regardless of what machines are plugged in, because any economically viable machine can be deployed against cheap power.

Hashprice sensitivity analysis replaces static collateral valuation. Instead of asking "what are these machines worth today," the right question is "at what hashprice does this operation's cash flow cover its debt service?" If the answer is a hashprice that sits well below the historical floor, the credit is sound. If the answer requires hashprice to stay above the current level, it isn't.

Stranded gas operators running at $0.02–$0.03/kWh have breakeven hashprices so low that the operational death — the first death — can't reach them. They don't need ASIC collateral to secure their credit. Their energy position is the collateral.

And the denomination matters. A loan denominated in BTC, serviced from BTC mining revenue, doesn't break when BTC/USD drops — because both sides of the equation move together. The difficulty adjustment, which recalibrates every two weeks, partially offsets revenue declines by improving the economics for surviving operators. That built-in mechanism only helps if the capital structure is denominated in the same system.

The Implication

The Bitcoin mining industry is currently structured around a lending model borrowed from traditional asset-based finance — a model that assumes collateral holds a floor value in distress. ASICs don't. They never have. Every cycle confirms it.

The operators who survive are not the ones with the newest machines or the largest fleet. They're the ones whose energy costs are low enough that the machines are an operational detail rather than a balance sheet bet.

If you were structuring a mining loan today, the first question shouldn't be "how many machines?" It should be "what does this operator pay for power, and can they pay it for the next four years without selling a single sat?"

Ben Vincenzi is the founder of BTSF, a Bitcoin-native credit institution focused on structured lending to Bitcoin miners. He is the author of Beyond Digital Gold: Bitcoin and the Architecture of a New Monetary System.