Bitcoin miners sell $348M BTC as power costs bite and the $7.4 billion treasury begins shrinking fast

Public Bitcoin miners collectively held 115,335 BTC as of Feb. 20, worth roughly $7.4 billion at the recent price, but that treasury dropped 4.44% month-over-month, the first sustained contraction since miners began stockpiling coins as balance-sheet assets.

The decline wasn’t an accident. Riot Platforms sold 1,818 BTC in December 2025 for $161.6 million in net proceeds. Bitdeer liquidated its entire treasury, selling 189.8 BTC it mined plus dumping 943.1 BTC from reserves to fund a pivot into AI infrastructure backed by $300 million in convertible notes.

The pattern suggests miner treasuries are shifting from strategic reserves to working capital, and the timing matters.

The market-implied hash price for the next six months sits around $28.73 per petahash per day, a level that makes older mining fleets uneconomic and forces operators to choose between selling Bitcoin, diluting equity, or raising expensive debt.

The setup compresses miner margins from multiple directions. Bitcoin’s April 2024 halving cut block subsidies to 3.125 BTC, reducing daily issuance to roughly 450 BTC. Transaction fees now contribute effectively zero to miner revenue, as CoinShares described fees as “decisively below 1%” of total miner income.

Mining difficulty rose approximately 14.73% on Feb. 19 to around 144.40 terahash, while hashprice dropped back below $30 per petahash per day.

VanEck’s mid-February 2026 analysis flagged the Antminer S19 XP as uneconomical above roughly $0.07 per kilowatt-hour under current conditions.

Riot’s third-quarter 2025 metrics illustrated the squeeze: the company’s cost to mine one Bitcoin was approximately $46,000 excluding depreciation, but $89,000 including capital equipment write-downs.

With Bitcoin trading in the mid-$60,000 range during parts of early 2026, the gap between all-in cost and spot price narrowed to the point where treasury sales became a rational form of liquidity management.

CompanyTreasury stanceBTC sold (period)Cash raised / proceedsRemaining BTC (post-action)Funding moveStated use of fundsRiot Platforms (RIOT)Selective liquidation (monetize a slice, keep most reserves)1,818 BTC (Dec 2025)$161.6M net proceeds from BTC sales18,005 BTCTreasury sale for liquidity (vs. full exit)Operational liquidity / balance-sheet management while continuing to hold a large BTC reserveBitdeer (BTDR)Full exit / zeroed treasury (treasury used as capex fuel)1,132.9 BTC total (reported Feb 2026): 189.8 BTC mined + 943.1 BTC from reserves$300M convertible notes (plus $45M option); BTC-sale proceeds not specified in the report0 BTCConvertible notes + treasury liquidationAI/data center expansion, AI cloud infrastructure, and mining hardware (AI pivot / HPC-style capex)

Treasuries as days of new issuance

At roughly 450 BTC per day in new issuance, the 115,335 BTC held by public miners represents approximately 256 days of new supply.

A 10% liquidation would release around 11,533 BTC, equivalent to 26 days of miner issuance. A 25% drawdown would amount to 28,834 BTC, or 64 days of supply.

The visible inventory pool matters because it appears on audited balance sheets and is subject to quarterly disclosure requirements.

Unlike decentralized mining operations, public miners report holdings and sales in SEC filings, making their treasuries the most transparent source of marginal supply.

Treasury concentration amplifies the dynamic. Marathon Digital holds 52,850 BTC, Riot Platforms 18,005 BTC, CleanSpark 13,513 BTC, and Hut 8 Mining 10,278 BTC.

Those four names control the bulk of disclosed reserves, meaning sell pressure is a function of how those companies fund operations when hashprice remains weak.

Bitdeer’s trajectory shows the extreme case: the company zeroed out its Bitcoin treasury while announcing $300 million in convertibles for data center expansion, AI cloud infrastructure, and mining hardware.

The pivot reframes Bitcoin holdings as capex fuel, and if hash price stays near current levels, other miners may follow suit.

The forward market is pricing sustained stress

Luxor’s hashprice forward market offers a quasi-forecast derived from market participants hedging future profitability.

As of Feb. 16, the forward curve priced the average hash rate at $28.73 per petahash per day over the next six months. That pricing suggests the market doesn’t expect a quick rebound in profitability.

CoinShares floated the possibility that global hashrate could reach 1.5 zettahash per second by mid-2026 if aggressive capacity expansion continues. A rising hashrate without a proportional increase in Bitcoin’s price would compress the hashprice further.

The difficulty adjustment mechanism creates timing risk. Difficulty increases lag hashrate surges, meaning miners can experience temporary profitability improvements when hashrate drops, only to see difficulty adjust upward and erase those gains weeks later.

A Feb. 22 analysis framed recent difficulty swings as a “difficulty up, hashprice down, fees thin” environment that arrived precisely when miners needed relief. The mismatch between when revenue improves and when difficulty recalibrates creates cash flow volatility that pushes operators toward preemptive treasury sales.

Selective liquidation vs full exit

Riot’s December 2025 sales offer one playbook.

The company sold 1,818 BTC for $161.6 million, reducing holdings to 18,005 BTC while retaining the majority of its treasury. The approach signals confidence that Bitcoin’s long-term trajectory justifies holding most reserves, even if short-term liquidity needs require partial monetization.

Riot’s cost structure, with mining costs around $46,000 per BTC excluding depreciation, suggests the company can generate positive cash flow if Bitcoin stays above that threshold.

Bitdeer represents the opposite extreme. The company liquidated its entire Bitcoin treasury, converting reserves into capital for AI and data center expansion. The move reframes mining as one revenue line within a diversified infrastructure business.

Bitdeer’s $300 million convertible notes financing shows the company betting it can generate better returns by deploying capital into AI cloud services than holding Bitcoin.

If other miners conclude that AI infrastructure or power monetization offers higher risk-adjusted returns, similar treasury drawdowns could follow.

The BTC runway calculation

The real question isn’t whether miners will sell, but which miners must sell and how much.

A simplified liquidity analysis ranks miners by their “BTC runway,” which is the number of months they can cover operating costs, interest, and capital commitments using cash, undrawn credit facilities, and convertible debt issuance, before needing to liquidate Bitcoin.

Miners with robust liquidity cushions can wait out low hash rate environments, while operators with thin cash buffers face pressure to monetize their treasuries.

Offsets complicate the picture. Hosting revenue from third-party miners, HPC contracts, power curtailment payments, and equipment sales can generate cash flow independent of Bitcoin mining.

Hedging strategies using futures or options can lock in forward prices. Miners with diversified revenue streams face different funding pressures than pure-play Bitcoin miners, who depend solely on block rewards and treasury appreciation.

Sell pressure won’t arrive uniformly, it will concentrate among operators with the shortest runways and the fewest alternative funding sources.

The market is already signaling stress

Glassnode’s Puell Multiple, a metric that compares daily miner revenue to its 365-day moving average, stood at 0.673 as of Feb. 23.

Readings below 1.0 indicate that miner revenue sits below its one-year average, a condition that historically precedes either industry consolidation or forced asset sales.

VanEck’s analysis of the S19 XP threshold being uneconomical above roughly $0.07 per kilowatt-hour matters because electricity costs across the industry aren’t uniform.

Miners operating in jurisdictions with cheap hydro or stranded gas enjoy margins that persist even as the hash price weakens. Operators in higher-cost regions face binary outcomes: relocate, upgrade to more efficient hardware, or shut down.

Treasuries become funding variables

The shift from HODL narrative to working capital tool reframes how the market should interpret miner balance sheets.

Public miners built treasuries during periods when the hash rate supported profitable operations, and Bitcoin’s price appreciated faster than returns from alternative investments. That environment is reversed.

Hash price forward curves signal sustained weakness, transaction fees contribute negligibly, and equipment obsolescence accelerates as difficulty rises.

The visible inventory pool of 115,335 BTC across public miners represents 256 days of new supply at current issuance rates, making even modest liquidation percentages meaningful in the daily market context.

Riot and Bitdeer demonstrated the range of responses: selective treasury sales to preserve optionality versus full liquidation to fund diversification.

The differences lie in capital access, revenue diversification, and management’s view of Bitcoin’s risk-adjusted returns. As long as forward hashprice expectations remain near $28.73 per petahash per day and older fleets turn uneconomic above $0.07 per kilowatt-hour, miner treasuries will function as a funding variable, not a HODL signal.

The market’s job is tracking which miners sell, how much, and whether the sales represent tactical liquidity management or systematic de-risking.