Miners’ Woes: BTC for AI, Yet Selling Pressure Wanes

Key Observations

Minerly dispositions reach 2024’s nadir, despite their grand metamorphosis.
MARA, in a fit of pragmatism, parts with 15,133 BTC in a mere three weeks to finance its AI aspirations.
The cost of birthing a single BTC now stands at a staggering $79,995.
AI contracts, exceeding $70B, are announced with great fanfare across the mining realm.
Bitcoin‘s network hashrate, once a towering 1,160 EH/s, now languishes at 920 EH/s.

In this most curious of conundrums, the Miners Position Index, a barometer of sorts, reveals a 30-day moving average that has retreated to its most modest level since 2024. The great liquidation, it seems, is upon us. Yet, the immediate on-chain selling pressure has cooled to a degree most remarkable, nearing its lowest ebb in two years. Both truths coexist, and in their coexistence, they unveil a narrative far more intricate than the headlines would suggest.

The Economic Compulsion Behind the Decision

CoinShares, in their Q1 2026 mining report, declare the weighted average cash cost of producing one Bitcoin among the publicly listed miners to be a staggering $79,995 in Q4 2025. Bitcoin, ever fickle, has traded between $67,000 and $70,000. The loss per coin mined, therefore, sits between $10,000 and $14,000 at current prices. A squeeze, you say? Nay, it is a structural quandary of the most pressing sort.

Hash price, that elusive metric determining miner revenue per unit of computing power, has plummeted from $63 per petahash per day in July 2025 to a five-year low of $35-37 by November, before further collapsing into Q1 2026. James Butterfill, the esteemed Head of Research at CoinShares, notes that hash price briefly touched $28 per petahash per day in late February, only to recover to a mere $30-35 at the time of the report-an all-time post-halving low. At such a level, any miner operating below the efficiency of an S19 XP, with electricity costs above $0.06 per kilowatt-hour, finds themselves in a most precarious financial position. CoinShares estimates this to be the plight of approximately 15-20% of the global mining fleet.

Traditional mining, a costly endeavor, requires $700,000 to $1 million per megawatt to establish. AI-ready facilities, however, demand a staggering $8 million to $15 million per megawatt, owing to liquid cooling requirements and high-density power systems for current-generation GPU hardware. Tenfold the capital requirement, yet with margins above 85% and multi-year revenue visibility, it presents a tantalizing alternative to the losses faced by most miners at current Bitcoin prices.

MARA, ever the pragmatist, has seized upon this gap with the most aggressive strategic pivot in the sector.

MARA’s Strategic Pivot in Detail

Between March 4 and March 25, MARA divested itself of 15,133 BTC, a substantial 28% of its entire treasury, generating approximately $1.1 billion in proceeds. With a flourish, the company employed $1 billion of this sum to repurchase its own convertible debt at a 9% discount to par, thereby saving $88.1 million in future obligations and reducing its total convertible debt from $3.3 billion to $2.3 billion in one fell swoop.

MARA’s CEO, Fred Thiel, has confirmed that the company shall continue to part with its Bitcoin assets from time to time, all to fund its grand expansion into digital energy and AI infrastructure. After the March sales, the company retains a modest 38,689 BTC.

This calculation, it appears, has been replicated across every other major public miner with equal fervor.

The Widespread Liquidation Across the Sector

Core Scientific, in a move most bold, sold approximately 1,900 BTC in January 2026 for $175 million and has announced plans to monetize substantially all of its remaining holdings throughout 2026, all to fund its AI colocation strategy. AI revenue, it seems, already accounts for 39% of Core Scientific’s total revenue. Bitdeer, not to be outdone, reduced its Bitcoin treasury to zero-a complete and utter divestment.

Riot Platforms, ever the opportunist, sold 1,818 BTC in December 2025 for $161.6 million, only to separately liquidate 1,080 BTC to fund a 200-acre land acquisition for an AI-ready campus development. Cipher Digital, in a similar vein, divested a 49% stake in mining joint ventures for $40 million and reduced its treasury from 2,284 BTC to 1,500 BTC as it redirects capital toward high-performance computing.

Collectively, the publicly listed miners have reduced their BTC treasuries by over 15,000 BTC from peak levels, according to CoinShares. Miners with secured HPC contracts now trade at 12.3 times next-twelve-month sales, compared to 5.9 times for pure-play Bitcoin miners-a valuation gap that renders every additional AI contract self-reinforcing.

This capital reallocation, however, has a most direct consequence for the network that secures Bitcoin.

Implications for the Network

Every dollar diverted toward AI data centers is a dollar withdrawn from hashrate. Bitcoin’s network hashrate, once a mighty 1,160 exahashes per second in early October 2025, has since dwindled to a mere 920 EH/s-a decline of more than 20%. This has resulted in three consecutive negative difficulty adjustments, the first such streak since July 2022.

CoinShares forecasts the network hashrate reaching 1.8 zetahashes by the end of 2026 and 2 zetahashes by the end of March 2027, contingent upon Bitcoin recovering to $100,000 by year-end. Should prices remain below $80,000, hash price will continue to fall, and more miners will exit. A sustained move below $70,000 could trigger a larger capitulation, which, in a most ironic twist, would benefit the survivors through lower difficulty.

The structural selling trend is unmistakable and accelerating. Yet, the on-chain data reveals a different tale regarding immediate selling pressure.

Insights from the MPI Data

The Miners Position Index, a most useful tool, measures the ratio of total miner outflow to its one-year moving average. An elevated reading signifies miners sending more coins to exchanges than their historical average, indicating overhead selling pressure in the near term. A suppressed reading, however, suggests miners are moving fewer coins than usual, thereby reducing immediate supply pressure on price.

According to CryptoQuant data, the MPI 30-day moving average has recently retreated to levels comparable to the 2024 lows. Despite the scale of the structural liquidations detailed above, miners are currently moving fewer coins to exchanges than at almost any point in the past two years.

The resolution lies in the timing. The large treasury liquidations-MARA’s 15,133 BTC, Core Scientific’s ongoing sales, Bitdeer’s complete divestment-were strategic one-time moves that have already cleared through the market. What remains is the day-to-day operational outflow from active mining, and this flow has cooled significantly. The miners who were destined to sell have largely done so. Immediate overhead supply from the sector has diminished as a result.

The structural pivot persists. Hashrate continues to decline. AI contracts are still being signed. The long-term selling pressure from miners converting treasuries into AI capital remains a defining feature of this cycle. The MPI at 2024 lows measures something narrower and more immediate-the coins moving to exchanges right now. By this measure, the mining sector is generating less selling pressure on Bitcoin’s price than it has in two years.

The Decisive Variable

The Bitcoin mining industry, once a group of companies securing the network and accumulating Bitcoin, is now transforming into a group that builds AI data centers and uses Bitcoin as the capital to fund them.

Bitdeer’s treasury stands at zero. Core Scientific is selling substantially all remaining holdings. MARA parted with 28% of its stack in three weeks. The structural direction is unequivocal.

What the MPI data adds to this narrative is a specific near-term nuance. The miners who were destined to sell have largely done so. Day-to-day operational outflow has cooled to 2024 lows. The immediate overhead supply pressure from the mining sector on Bitcoin’s price is near its lowest point in two years, not because the pivot has slowed, but because the largest one-time liquidations have already cleared through the market.

Two truths coexist. The structural selling is real, ongoing, and tied to economics that remain unchanged at $67,000. The near-term selling pressure from miners is at a two-year low. CoinShares’ $100,000 year-end forecast would reverse the economics that made this pivot rational in the first place, potentially slowing both trends at once.

The coins that were destined to move have already done so. Whether new ones follow depends on one number.

The information provided herein is for educational purposes only and does not constitute financial, investment, or trading advice. The author does not endorse or recommend any specific investment strategy or cryptocurrency. Always conduct your own research and consult with a licensed financial advisor before making any investment decisions.

2026-03-29 15:22