The little deer live by the water and grass

Author: Zhou, ChainCatcher

Bitdeer's Bitcoin holdings have been reduced to zero.

As one of the leading mining companies, it held over 2,400 BTC at its peak in November last year, and has since continuously increased its selling pace, completing its liquidation by mid-February, and now maintains a rhythm of selling as much as it mines.

Notably, the company's financial report shows that it achieved revenue of $224.8 million in the fourth quarter of 2025, a year-on-year increase of 226%; net profit reached $70.5 million; total hash rate reached 71.0 EH/s, a year-on-year increase of 229%, and mining efficiency significantly improved from 30.4 J/TH to 17.9 J/TH.

A company with a positive balance sheet and record hash rate chooses to liquidate its Bitcoin reserves at this time. To understand this operation, one must first return to a fundamental fact that has long been obscured by the market.

1. Hoarding coins has always been a minority

Bitdeer was never a coin-hoarding institution driven by belief.

In its early years, the company adhered to the simplest mining logic—mine, sell, and convert to cash. BTC was not an asset for it, but a product.

Only in the past two years, as MicroStrategy's hoarding model gained immense popularity in the capital markets, the valuation logic for mining companies began to be reshaped, and Bitdeer briefly switched to a hoarding narrative in line with industry trends.

This trend-following is not uncommon in the industry, but very few have truly persisted.

Analyst Tom Dunleavy from the blockchain research firm Messari published a set of data showing that from January to November 2025, ten mainstream listed mining companies, including Core Scientific, Riot, Marathon, and Hut 8, mined approximately 40,700 BTC and sold about 40,300 BTC during the same period, with a sell-through rate close to 99%.

This means these companies have never truly hoarded coins.

This figure reveals the fundamental survival rule of the industry: the essence of mining companies is energy arbitrage; Bitcoin is merely the intermediary through which they convert cheap electricity into revenue, not a long-term holding on their balance sheets.

Recently, the market's belief in the hoarding narrative was partly due to the continuous rise in BTC prices, which obscured this reality—when assets are appreciating, whether to sell becomes a matter of posture; but when prices fall below mining costs, selling coins shifts from a matter of posture to a survival instinct.

Thus, Bitdeer's liquidation this time is less a betrayal of belief and more a return to its true nature. This is certainly not a signal of bearishness towards Bitcoin; Wu Jihan himself stated on social media that a zero holding does not mean it will always be this way.

However, this time, the coins sold were not for operational cash but for the startup capital for transformation. That brief period of hoarding was merely an interlude in the industry's collective effort to tell a story to the capital market.

2. Triple pressure: How cold is the winter for mining companies?

Understanding that hoarding coins is a minority allows for a clearer view of the current situation of mining companies. What weighs on this industry is a triple tightening dilemma.

First is the cost pressure after the halving.

The halving in 2024 means that block rewards are halved, directly cutting the unit output of mining companies in half, while electricity costs, machine depreciation, and operational costs remain unchanged. Many mining machines' shutdown prices are already close to or even exceed the current BTC price, meaning that operating them incurs losses, while shutting them down wastes assets.

Dunleavy also pointed out that miners' continuous selling of newly mined BTC itself constitutes structural downward pressure on prices. The lower the price, the more mining companies need to sell coins; the more they sell, the harder it is for prices to rebound, thus forming a self-reinforcing vicious cycle.

Second are the stark numbers in financial reports.

Opening the annual reports of mining companies for 2025, almost without exception, they show a structure where revenue is rising, but losses are also increasing.

MARA Holdings' annual revenue grew from $656 million to $907 million, but net losses soared to $1.31 billion, compared to a profit of $541 million in the same period last year.

Hut 8's revenue increased from $162 million to $235 million, while net losses shifted from a profit of $331 million to a loss of $248 million. TeraWulf's annual revenue grew from $140 million to $169 million, but fourth-quarter losses per share expanded from $0.21 to $1.66.

The phenomenon of increasing revenue without increasing profit appears simultaneously in several leading companies, indicating that this is not a management issue, but a structural cyclical compression in the industry.

The fair value fluctuations of hoarded assets directly penetrate the profit and loss statement, making the numbers on the balance sheet particularly unattractive. Meanwhile, companies are still exchanging debt for transformation: Hut 8 launched a $1 billion ATM financing plan and signed a credit agreement with Coinbase for up to $400 million; Cipher Digital completed three financing rounds totaling as much as $3.73 billion.

Finally, the macro environment has changed.

Trump raised global tariffs, geopolitical uncertainties continue to escalate, and risk assets are generally under pressure, with Bitcoin falling below $65,000.

Cryptocurrency analysis firm QCP pointed out, that Bitcoin prices are significantly below the average mining cost, and prioritizing liquidity over hoarding is no longer a strategic choice but a reality constraint.

Bitdeer's liquidation and Cango's beginning to sell a small amount of BTC for operations, when taken together, outline the industry's risk reduction profile.

3. To live by dying: The gamble of transformation

In the context of simultaneous triple pressure, mining companies have only one way out: transformation, leveraging the infrastructure assets accumulated from Bitcoin mining to pry open a new source of income.

AI and high-performance computing have become the next card that the industry collectively bets on.

The logic is not hard to understand. Mining companies hold a large number of cheap electricity contracts and scalable data center land, which are precisely the most scarce resources for AI computing infrastructure. Switching from low-margin mining computing power to high-margin AI computing power leasing appears to be a profitable deal on paper.

Bitdeer is vigorously promoting its self-developed mining machine Sealminer, AI cloud services, and high-performance computing business; Cipher has changed its brand from Mining to Digital, signaling a platform transformation; multiple companies are competing to secure long-term low-price electricity contracts in hopes of establishing a structural moat on the energy cost side.

However, the reality of progress is far more conservative than the narrative.

Take TeraWulf, for example; the company's revenue from HPC in the fourth quarter was only $9.7 million, accounting for less than 30% of total revenue of $35.8 million, and significantly down from the third quarter.

Acquiring customers for AI business, landing contracts, and ramping up production capacity all take time, while debt pressure and equity dilution are immediate realities.

The outcome of this transformation gamble depends on whether the new business can truly scale before the debt matures.

Interestingly, during the same period when BTC fell nearly 17% in a month, many mining company stocks rose against the trend. For instance, TeraWulf increased by 31% over the month, Cipher rose by 8%, Hut 8 by 6%, and Core Scientific remained basically flat.

This decoupling reflects a re-evaluation by the capital market—mining companies have long been one of the sectors with the highest short-selling ratios by hedge funds, and the upward momentum may partly come from short covering.

This indicates that the market is beginning to view these companies as potential carriers of AI computing infrastructure rather than as amplifiers of Bitcoin price leverage.

In the future, their judgment criteria may no longer be based on how much BTC one holds, but rather on who has locked in the longest-term low-price electricity, whose data center assets have the greatest potential for AI transformation, and whose balance sheet can withstand the pains of transformation.

Conclusion

Mining companies have never been the most devout believers in Bitcoin; they are the most rational industry participants. When mining is profitable, they mine; when hoarding coins can support valuations, they choose to hoard; when selling coins can provide the bullets for transformation, they will sell without hesitation. This is the basic logic of business.

What is truly worth questioning is the next question: when the story of AI/HPC transformation is fully priced by the capital market, what can these companies offer to support the next round of valuation? If by then Bitcoin prices have rebounded while the transformation business is still immature, will those mining companies that liquidated today start telling the hoarding story again?

Cycles repeat, narratives remain fresh. But in every winter, surviving is always more important than belief.