Bitcoin Miners Draw Down $7.4B Treasuries as Winter Power Costs Force 5,359 BTC Sell-Off

Publicly listed Bitcoin miners are starting to tap their once sacrosanct BTC treasuries, as winter power costs, weaker margins, and a pessimistic forward hashprice market put balance sheets under pressure.

As of Feb. 20, public miners collectively held 115,335 BTC, worth roughly $7.4 billion at recent prices. That stash fell 4.44% month-over-month — the first sustained contraction since large miners began treating Bitcoin as a core balance-sheet asset rather than just operating inventory.

The selling is not random opportunism. It reflects a structural shift in how miners fund operations and growth in an environment where block rewards are smaller, fees are negligible, and energy is expensive.

From HODL to Working Capital: What the 5,359 BTC Drawdown Signals

The recent 4.44% decline in public miner treasuries implies they collectively sold around 5,359 BTC in a single month. For a segment that has cultivated a long-term “HODL” narrative, this is a notable change in posture.

Two names illustrate the evolving playbook:

• Riot Platforms sold 1,818 BTC in December 2025, raising $161.6 million in net proceeds while retaining a sizable 18,005 BTC treasury.
• Bitdeer went further, liquidating its entire stash — a total of 1,132.9 BTC (189.8 BTC mined plus 943.1 BTC from reserves) — and pairing that with $300 million in convertible notes to fund an aggressive pivot into AI infrastructure and data centers.

These moves suggest that treasuries are now being treated less as untouchable strategic reserves and more as flexible working capital, especially when spot Bitcoin prices trade only modestly above miners’ full-in costs.

The shift is occurring under clear economic strain. The market-implied hashprice for the next six months sits near $28.73 per petahash per day, a level that makes older fleets unprofitable in many jurisdictions. Operators are being pushed toward hard choices: sell BTC, issue equity, or borrow at higher marginal costs.

The Economics Squeezing Miners: Hashprice, Halving, and Energy Costs

Miners are being hit on multiple fronts at once, compressing margins even as Bitcoin’s headline price looks healthy to spot market participants.

First, the April 2024 halving cut the block subsidy to 3.125 BTC, reducing new issuance to roughly 450 BTC per day. That instantly halved the core revenue stream for miners. Unlike some previous cycles, transaction fees have not meaningfully offset the difference. CoinShares characterizes fees as “decisively below 1%” of total miner income, effectively zero from an economic planning perspective.

Second, network difficulty and hashrate have moved against weaker operators. On Feb. 19, Bitcoin’s mining difficulty rose about 14.73% to around 144.40 terahash, while hashprice slipped back under $30 per petahash per day. In practice, more work is required to earn the same BTC, and each unit of hashrate earns less revenue when translated into dollars.

Third, energy costs are decisive for hardware economics. VanEck’s mid-February 2026 analysis flagged that an Antminer S19 XP — still widely deployed — becomes uneconomical above roughly $0.07 per kilowatt-hour under current conditions. That threshold makes location and power contracts central to survival. Miners with cheap hydro or stranded gas retain some buffer; operators paying higher tariffs face much tighter or even negative margins.

Riot Platforms’ third-quarter 2025 metrics underscore the squeeze. The company’s cost to mine one Bitcoin was approximately $46,000 excluding depreciation, but about $89,000 when capital equipment write-downs are included. With Bitcoin trading in the mid-$60,000 range at points in early 2026, the spread between all-in costs and spot narrowed enough that selling from treasury became a rational liquidity tool rather than a last resort.

Treasuries as Days of Issuance: How Much Supply Is at Stake?

To understand the potential market impact of miner selling, it is useful to view their holdings not just as an absolute BTC figure but as a multiple of daily issuance.

At current block subsidies, roughly 450 BTC are created each day. Against that backdrop, the 115,335 BTC held by public miners represents approximately 256 days of new supply.

Different liquidation scenarios highlight the sensitivity:

• A 10% drawdown — about 11,533 BTC — is roughly equivalent to 26 days of miner issuance.
• A 25% drawdown — around 28,834 BTC — would equal about 64 days of issuance.

Because public miners report their holdings and sales through audited financial statements and regulatory filings, this inventory pool is the most transparent slice of potential marginal supply. Unlike smaller or decentralized operations, public miners can’t quietly run down reserves; markets see the changes quarter by quarter.

Treasury concentration amplifies that dynamic. Four companies dominate disclosed holdings:

• Marathon Digital: 52,850 BTC
• Riot Platforms: 18,005 BTC
• CleanSpark: 13,513 BTC
• Hut 8 Mining: 10,278 BTC

These four control 94,646 BTC, or roughly 82.1% of the 115,335 BTC held across public miners as of February 2026. As a result, any meaningful wave of sell pressure is likely to be driven by decisions at a small number of large operators, rather than evenly distributed across the industry.

Bitdeer’s complete exit is the extreme case: its Bitcoin holdings have effectively become fuel for capital expenditure on AI cloud infrastructure and data center expansion. If hashprice remains at present levels, other miners may reframe their BTC similarly, as a funding source rather than a strategic reserve.

Forward Markets and Difficulty Swings Point to Sustained Stress

The underlying conditions that are pushing miners toward treasury sales look unlikely to resolve quickly, based on current forward pricing and network behavior.

Luxor’s hashprice forward market provides a real-money gauge of expectations, as miners and counterparties hedge future profitability. As of Feb. 16, the forward curve priced the average hashprice at about $28.73 per petahash per day over the next six months. That level implies the market is not anticipating a near-term recovery in mining economics.

CoinShares has floated a scenario where global hashrate could reach 1.5 zettahash per second by mid-2026 if aggressive capacity expansion continues. If that materializes without a proportional rise in Bitcoin’s price, hashprice would come under further pressure as more computing power competes for the same fixed block rewards.

The difficulty adjustment mechanism adds timing risk. Difficulty responds with a lag to changes in hashrate. When hashrate falls, miners may see a brief uplift in profitability before the protocol recalibrates and partially erases those gains; when hashrate rises, difficulty may not catch up immediately, creating short-lived windows of better economics. A Feb. 22 analysis described the current environment as “difficulty up, hashprice down, fees thin” — an unfavorable combination arriving just as miners most need relief.

This uneven revenue profile contributes to cash flow volatility. To avoid being forced sellers at precisely the wrong time, some operators are opting for preemptive treasury sales to build liquidity cushions ahead of potential further difficulty increases or fee stagnation.

Riot vs. Bitdeer: Two Opposite Treasury Strategies

Recent actions by Riot Platforms and Bitdeer outline two ends of the strategic spectrum for miner treasuries under stress: selective liquidation versus full exit.

Riot Platforms chose a calibrated approach. By selling 1,818 BTC for $161.6 million in December 2025, the company raised meaningful liquidity while still retaining 18,005 BTC on its balance sheet. This signals that management still views long-term Bitcoin exposure as valuable, but is willing to trim holdings to manage operational risk and near-term obligations.

Riot’s underlying economics support this posture. With a reported cost of around $46,000 per BTC mined (excluding depreciation), the firm can remain cash-flow positive as long as Bitcoin trades comfortably above that threshold, using its treasury as a buffer rather than an existential lifeline.

Bitdeer, by contrast, opted for a decisive pivot. The firm liquidated its entire Bitcoin treasury — selling both newly mined coins and reserves — while simultaneously raising $300 million via convertible notes (with an additional $45 million option) to fund AI and high-performance computing (HPC)-style capital expenditures. Its stated use of funds includes data center builds, AI cloud infrastructure, and mining hardware.

In this model, Bitcoin mining becomes one revenue stream within a broader infrastructure business, and BTC holdings are treated as redeployable capital rather than long-term store-of-value assets. If other miners conclude that AI or alternative data center uses of power deliver better risk-adjusted returns than Bitcoin mining at current hashprice levels, similar treasury restructurings could follow.

Runway, Stress Metrics, and the New Role of BTC Treasuries

For investors and analysts, the immediate question is less whether miners will sell, and more which miners must sell, and on what timeline.

One way to frame this is via a simplified “BTC runway” concept — the number of months a miner can cover operating costs, interest, and committed capex using available cash, undrawn credit facilities, and potential debt issuance before needing to tap its Bitcoin holdings. Companies with ample liquidity and diversified revenues (such as hosting, HPC contracts, power curtailment payments, or equipment sales) can afford to wait out weak hashprice conditions. Pure-play miners with limited cash buffers and higher power costs cannot.

Glassnode’s Puell Multiple provides another lens on sector-wide stress. As of Feb. 23, the metric stood at 0.673, meaning daily miner revenue is just 67.3% of its 365-day moving average. Historically, readings below 1.0 have preceded either periods of consolidation — where weaker miners shut down or are acquired — or forced asset sales, including BTC treasuries.

Electricity price dispersion compounds the divergence in outcomes. VanEck’s finding that S19 XP units turn uneconomic above roughly $0.07 per kilowatt-hour underscores how location-specific energy deals shape survivability. Miners with access to very cheap power can continue operations with modest or no treasury drawdowns, while those at the higher end of the cost curve must either relocate, upgrade fleets, or exit.

Against this backdrop, Bitcoin treasuries are no longer straightforward “HODL” indicators. The same 115,335 BTC that once signaled long-term bullish conviction now represents a flexible funding pool equivalent to 256 days of new supply. Riot and Bitdeer illustrate the two poles of how that pool can be used: selective monetization to maintain optionality versus complete liquidation to finance diversification.

As long as forward hashprice expectations hover near $28.73 per petahash per day and older rigs remain marginal or unprofitable above $0.07 per kilowatt-hour, market participants should treat miner treasuries as dynamic funding variables. Tracking who sells, in what size, and with what stated strategic intent will be central to understanding both mining-sector risk and an increasingly visible source of marginal Bitcoin supply.