Wu Jihan's Turnaround: Bitdeer Rewrites the Mining Story with $1.3 Billion in Debt

Bitdeer reported a remarkable financial statement in February 2026: it liquidated all Bitcoin holdings and raised $325 million through bond financing. Combined with the previously accumulated $1 billion loan, this makes the world’s largest publicly listed mining company’s total debt surpass $1.3 billion. Behind these numbers, Wu Jihan is undergoing an unprecedented transformation—from a miner chasing higher coin prices to an energy giant controlling global computing infrastructure. This isn’t reckless risk-taking but a profound evolution of the business model over the past decade.

The Evolution of Mining Thinking: How Wu Jihan Views the Next Era

Bitdeer’s story began in 2018 with a mining machine sharing platform. By early 2026, it had become the most centralized mining enterprise globally, with 63.2 EH/s of self-owned mining capacity, accounting for about 6% of the Bitcoin network’s total hash rate. But while everyone else was competing for higher Bitcoin mining yields, Wu Jihan was pondering a different question: if mining is essentially exchanging today’s costs for tomorrow’s rewards, could a different bet yield greater success?

This is Wu Jihan’s core shift. Over the past decade, the mining industry bet on rising Bitcoin prices. Today, Bitdeer is betting on the long-term global demand for computing power—especially with the advent of AI. Its approach has shifted from simply exchanging electricity and equipment to acquiring land and energy assets worldwide. This logic may seem radical, but it actually continues the oldest arbitrage tradition in mining: turning from buyer to seller.

Today, Bitdeer controls a global grid capacity of 3,002 MW, with 1,658 MW in operation and 1,344 MW under construction. To put this into perspective, a single hyperscale data center from Microsoft or Google typically ranges from 100 to 300 MW. In other words, Bitdeer’s energy assets amount to the power needs of 10 to 30 Google-scale data centers. On paper, this is indeed impressive.

The True Face of $1.3 Billion Debt: Convertible Bonds and a Time Gamble

In May 2024, Tether invested $100 million, becoming Bitdeer’s second-largest shareholder. Since then, financing has become a relay race. In just two years, Bitdeer completed multiple bond issuances: in August 2024, $150 million in convertible bonds (8.5% annual interest); in November 2024, $360 million bonds (interest rate down to 5.25%); in November 2025, $548 million financing; and in February 2026, an additional $368 million. With each round, Wu Jihan successfully raised funds, but at the cost of a 10-17% drop in stock price. Market reactions have become almost reflexive.

The core tool supporting all this is the convertible bond. These bonds set a conversion price—new bonds due in 2032, with a conversion price of about $9.93, 25% higher than the $7.94 at the time of equity financing. When the stock price reaches this level, bondholders will choose to convert into shares rather than demand cash repayment. In other words, Bitdeer essentially doesn’t need to repay—if the stock price rises, the bonds will automatically convert.

The brilliance of this structure lies in its nature as a market narrative gamble. With $1.3 billion in debt and an average interest rate of 5%, Bitdeer’s annual interest expense exceeds $65 million. Yet, the entire AI/HPC cloud service revenue in 2025 was less than six months’ interest. Currently, this interest is entirely sustained by continuous financing—there’s no room for complacency.

Critical Windows for Execution: Two Completely Different Future Paths

Bitdeer’s funds are directed toward assets in three locations. First is 563 MW in Rockdale, Texas, already operational, mainly for traditional mining with stable cash flow. Second is the more ambitious Kingston project in Ontario, Canada—570 MW, with a 30-year lease agreement signed, power contracts secured, scheduled for completion in Q2 2027, positioning as a core hub for AI/HPC. Third is 175 MW in Tvedestrand, Norway, transforming from mining facilities into an AI data center, expected to be completed by late 2026, with 164 MW of available IT load, leveraging low-cost hydropower.

Wu Jihan has carefully designed bond maturities—2029, 2031, and 2032. On the surface, this disperses risk; deeper down, it provides multiple opportunities for renegotiation. Ideally, by 2029, the first bonds mature, Tvedestrand is operational, Kingston is launched, AI revenue has scaled, and the stock price is rising, prompting bondholders to convert rather than claim.

But a major threat exists in reality: within the industrial park where Kingston is located, there is a steel manufacturer, U.S. Heavy Steel Plate, which has filed a lawsuit against Bitdeer. They claim that building AI data centers interferes with shared power, roads, railways, and communication lines, violating lease restrictions, and are seeking a court injunction to halt construction. This lawsuit is significant—Kingston accounts for 42% of the ongoing pipeline. If stalled, the entire timeline must be rewritten.

Currently, Bitdeer’s biggest risk isn’t debt size or stock volatility but this lawsuit. Meanwhile, mining difficulty is surging. In February 2026, Bitcoin network difficulty jumped 14.7% in a single month—the largest single increase since May 2021. At the same electricity cost, fewer coins are mined. Gross margins in Q4 fell from 7.4% last year to 4.7%. Mining is becoming less profitable.

The worst-case scenario is also imaginable: Kingston’s lawsuit delays two years, construction stalls; Tvedestrand’s progress is delayed, GPU utilization remains at 41%; by 2029, the first bonds mature, cash reserves are insufficient, forcing refinancing, diluting shares further, and making conversion even harder. Both paths are real possibilities.

Clearing Bitcoin: Not Abandonment, But a Game-Changing Shift

Bitdeer has liquidated all its Bitcoin holdings, causing a stir in the mining industry. Traditional miners like Marathon (MARA) hold 53,000 BTC, Riot 18,000 BTC, and Strategy even controls 710,000 BTC. Holding coins has been seen as a form of faith—market participants believe management’s confidence in Bitcoin’s long-term value. Wu Jihan has broken this tradition.

The official explanation is that selling cryptocurrencies provides liquidity for global land and energy acquisitions. This isn’t inherently problematic. Competitors are also shifting: Riot sold $200 million worth of BTC for AI expansion, Bitfarms is moving away from being a “Bitcoin company,” and Marathon is starting to focus on HPC.

But there’s a deeper strategic shift. Since its inception, the mining industry has been engaged in the same gamble: betting that something in the future will be worth more than its current cost. Ten years ago, it was betting on rising coin prices; today, it’s betting on explosive demand for computing power. The bet has changed, but the essence of arbitrage over time remains. Wu Jihan is truly buying into a position—“whoever wins, I can profit from electricity bills.” It’s very much like Amazon’s logic: not betting on which internet company will win, but renting servers to everyone. AT&T doesn’t care what you talk about on the phone, as long as you pay the bill.

From selling products to selling services, and now to collecting rent, industry evolution has always followed this path. The only difference is whether it’s proactive evolution or forced adaptation. Wu Jihan has spent billions acquiring the window for this evolution. He is waiting for AI’s money growth to catch up with debt growth.

Whether this gamble succeeds depends on the next two to three years of execution. Time is both Wu Jihan’s weapon and his greatest opponent.