Recent market attention has been focused on Bitdeer's dramatic decision: to empty all their Bitcoin reserves. Last week, Wu Jihan posted an update that they managed to retain 189.8 BTC from mining operations, which they sold directly. The remaining stock of 943.1 BTC was also sold. Now, the Bitcoin balance is completely zero.

Why suddenly? The answer lies in a much more ambitious transformation strategy: from a traditional mining company to an AI infrastructure provider. And to finance this transition, Bitdeer has accumulated debt totaling $1.3 billion.

There is an interesting logic behind this. From the beginning, Bitcoin mining has always operated on a timing arbitrage: use electricity and machines today, exchange for Bitcoin tomorrow. Now, Wu Jihan is changing that arbitrage target. Instead of predicting coin prices, it’s now about the long-term value of computational power in the AI era. The method is also changing: from consuming electricity to earn coins, to borrowing money to buy land and infrastructure.

Bitdeer currently controls a global power capacity of 3,002 MW, with 1,658 MW already operational and 1,344 MW under construction or planning. For context, large data centers owned by Microsoft and Google typically range from 100 to 300 MW. So, 3,002 MW is equivalent to aggregating the power needs of 10 to 30 giant data centers within a single company. This is an extremely ambitious goal.

The main projects are in three locations. First, Rockdale, Texas, with 563 MW already operational, focused on mining, with stable cash flow. Second, Clarington, Ohio, with 570 MW—this is the most critical—leased for 30 years, targeted to be completed by Q2 2027, and will become a major HPC/AI hub. Third, Tydal, Norway, with 175 MW, transforming from a mine into an AI data center, scheduled for completion at the end of 2026, utilizing hydroelectric energy with significantly lower transformation costs.

To support this expansion, Bitdeer has issued a series of debts. In May 2024, Tether invested $100 million. Three months later, the first convertible loan of $150 million with an 8.5% interest rate. In November 2024, the second convertible loan of $360 million with interest reduced to 5.25%. In November 2025, a combined offering of $400 million in convertible bonds and additional shares. In February 2026, again, $325 million in convertible bonds plus repurchase of old debt, totaling over $1.4 billion. Each issuance causes Bitdeer’s stock price to drop 10-17%, but the company always manages to raise funds.

This debt structure is carefully designed. Three series of convertible bonds are set to mature in 2029, 2031, and 2032. This is no coincidence. When the first batch matures in 2029, Tydal and Clarington are theoretically already operational. When the second batch matures in 2031, AI revenue should be significant. When the third batch matures in 2032, the market will evaluate what this company truly is. Three time windows, three opportunities for renegotiation.

However, the annual interest burden, assuming an average of 5% on the principal of $1.3 billion, results in expenses of over $650 million per year. Meanwhile, AI/HPC cloud revenue throughout 2025 will not even reach one-tenth of the six-month interest. Currently, interest payments are entirely dependent on continuous debt issuance to keep the cycle going.

The greatest risk is not from the debt itself, but from a steel plant. At the Clarington, Ohio industrial park, there is a steel producer called American Heavy Plate Solutions, which signed a 30-year lease starting in 2018. They sued Bitdeer, claiming that building the AI data center would disrupt electricity, roads, railways, and shared communication lines. They are asking the court for a permanent injunction. Clarington accounts for 42% of the pipeline under construction. If delayed, the entire schedule must be rewritten.

Additionally, the mining side is also starting to feel pressure. In February 2026, Bitcoin network difficulty surged 14.7%, the largest increase since May 2021. With the same electricity, the amount of coins mined decreased. Gross margins in Q4 fell from 7.4% a year ago to 4.7%. The mining branch is gradually thinning out.

So, Wu Jihan is playing a very tight game with time. He is buying this window with billions of dollars. The logic he applies is very simple yet powerful: don’t guess which AI industry path will succeed, just control the entry point of electricity infrastructure. Amazon doesn’t guess which internet company will succeed, it just rents servers. AT&T doesn’t care what people talk about, as long as they use the network. From selling products to selling services to collecting rent—industry evolution always moves in one direction.

In the next two to three years, we will see whether this strategy succeeds. If Tydal is completed on time at the end of 2026, Clarington wins the lawsuit and begins construction in 2027, and two core assets operate fully by 2028-2029, then Bitdeer’s transformation narrative will be proven. Revenue will reach billions of dollars, analysts will shift the label from a discounted mining company to a premium AI infrastructure provider, and debt holders will prefer to convert bonds into shares rather than receive cash. Conversely, if lawsuits drag on, Tydal faces delays, and GPUs remain underutilized, then this story could turn out very differently. Two scenarios, both real. Now, it’s just a matter of waiting for Wu Jihan’s execution over the next few years.