Little Deer of Bitcoin live by water and grass.

Author: Zhou, ChainCatcher

Bitdeer’s Bitcoin holdings have hit zero.

As one of the leading mining companies, it held over 2,400 BTC at its peak in November last year. Since then, it has continuously increased sales, completing a full liquidation by mid-February. Now, it maintains a rhythm of mining as much as it sells.

Notably, the company’s financial report shows that in Q4 2025, revenue reached $224.8 million, a 226% year-over-year increase; net profit was $70.5 million; total hash rate hit 71.0 EH/s, up 229%, and mining efficiency improved significantly from 30.4 J/TH to 17.9 J/TH.

A company with strong financials and record-high hash rate choosing to clear its Bitcoin reserves at this time. To understand this move, we need to revisit a fundamental fact long obscured by the market.

1. HODLing has always been a minority practice

Bitdeer is not a faith-driven HODLer.

In its early years, the company followed the simplest mining logic—mine, sell, convert to cash. For them, BTC is not an asset but a product.

Until recent years, MicroStrategy’s HODL strategy gained massive popularity in capital markets, reshaping the valuation logic of mining companies. Only then did Bitdeer briefly switch to a HODL narrative, following industry trends.

This kind of bandwagoning is common across the industry, but few persist long-term.

Analyst Tom Dunleavy from Messari shared data showing that from January to November 2025, 10 major publicly listed miners—Core Scientific, Riot, Marathon, Hut 8, and others—mined about 40,700 BTC but sold approximately 40,300 BTC in the same period, with a sale rate approaching 99%.

In other words, these companies have never truly been HODLers.

This figure reveals a fundamental survival principle in the industry: Mining companies are essentially energy arbitrageurs. Bitcoin is a means to convert cheap electricity into revenue, not a long-term asset on their balance sheets.

Recently, the market’s belief in the HODL story was partly due to the sustained rise in BTC prices, which masked this reality—when assets appreciate, whether to sell or not becomes a matter of stance; but when prices fall below mining costs, selling shifts from a stance to a survival instinct.

Therefore, Bitdeer’s liquidation isn’t a betrayal of faith but a return to fundamentals. This isn’t a bearish signal for Bitcoin. Wu Jihan himself stated on social media that holding zero doesn’t mean it will stay that way forever.

This time, the coins sold weren’t for operational cash but for transition capital. That brief HODL phase was just a collective industry story-telling episode with capital markets.

2. Triple pressure: how cold is the winter for miners?

Understanding that HODLing is a minority practice helps clarify the current predicament of miners. But what weighs on the industry is a triple synchronized tightening.

First, the cost pressure after the halving.

The 2024 halving cut block rewards in half, directly halving the output per unit. Meanwhile, electricity, depreciation, and operational costs remain unchanged. Many miners’ shutdown prices are approaching or even exceeding current BTC prices, meaning turning on machines results in losses, while shutting down wastes assets.

Dunleavy also pointed out that miners continuously selling newly mined BTC create a structural downward pressure on prices. The lower the price, the more miners need to sell; the more they sell, the harder it is for prices to rebound, forming a self-reinforcing vicious cycle.

Second, the stark numbers in financial reports.

Looking at the 2025 annual reports, almost all show rising revenue alongside rising losses.

MARA Holdings’ revenue grew from $656 million to $907 million, but net loss soared to $1.31 billion, compared to a profit of $541 million the previous year.

Hut 8’s revenue increased from $162 million to $235 million, but net loss shifted from a profit of $331 million to a loss of $248 million. TeraWulf’s revenue rose from $140 million to $169 million, with quarterly per-share loss expanding from $0.21 to $1.66.

The phenomenon of rising revenue but not profit is common among leading companies, indicating this isn’t management’s problem but a structural industry cycle.

The fair value fluctuations of HODL assets directly impact earnings, making the books look ugly. Meanwhile, companies are still borrowing to fund transitions: 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 financings totaling $3.73 billion.

Finally, the macro environment has shifted.

Trump’s tariffs increased global trade tensions, geopolitical uncertainties rose, and risk assets came under pressure, with Bitcoin dropping below $65,000.

Crypto analytics firm QCP pointed out that Bitcoin’s price is significantly below average mining costs. Liquidity prioritization over HODLing is no longer a strategic choice but a practical constraint.

Bitdeer’s liquidation and Cango’s small BTC sales for operations, along with other moves, outline a trend of industry de-risking.

3. Facing death to be reborn: the gamble of transformation

Under the combined pressures, the only way out for miners is transformation—using the infrastructure assets accumulated from Bitcoin mining to unlock new revenue streams.

AI and high-performance computing have become the next strategic focus for the industry.

The logic is straightforward. Miners hold large amounts of cheap electricity contracts and scalable data center land—resources that are scarce for AI compute infrastructure. Switching from low-margin mining to high-margin AI compute leasing appears to be a profitable move on paper.

Bitdeer is pushing hard on self-developed miners Sealminer, AI cloud services, and high-performance computing; Cipher is rebranding from Mining to Digital, signaling platform-based transformation; many companies are locking in long-term low-cost power contracts to build a structural moat around energy costs.

However, progress is more conservative than the narrative suggests.

Take TeraWulf as an example: in Q4, revenue from HPC was only $9.7 million out of $35.8 million total, a significant drop from Q3.

Customer acquisition, contract signing, and capacity ramp-up for AI take time, while debt pressures and equity dilution are immediate.

The success of this transformation gamble depends on whether new businesses can scale before debt maturities.

Interestingly, during the same month Bitcoin fell nearly 17%, several mining stocks actually rose against the trend. TeraWulf gained 31%, Cipher 8%, Hut 8 6%, and Core Scientific remained flat.

This decoupling reflects a market revaluation—mining companies have long been among the most shorted sectors by hedge funds, and their rally may partly be due to short covering.

It indicates that the market is starting to see these companies as potential AI compute infrastructure providers rather than just Bitcoin price leverage amplifiers.

In the future, valuation criteria may shift from “how much BTC do they hold” to “who has secured the longest-term, lowest-cost power,” “whose data centers are most AI-ready,” and “whose balance sheet can withstand the pain of transformation.”

Conclusion

Mining companies have never been the most faithful believers in Bitcoin; they are the most rational industry players. When mining is profitable, they mine; when HODLing can support valuation, they HODL; when selling coins funds transformation, they sell without hesitation. That’s basic business logic.

The real question is: after the market fully prices in the AI/HPC transformation story, what can these companies still offer to support their next valuation? If Bitcoin prices recover but their new businesses are not yet mature, will those who liquidated today start telling the HODL story again?

Cycle repeats, narratives evolve. But in every winter, survival always matters more than faith.