Attention to Bitdeer—an mining company that is now playing a much bigger game than just Bitcoin.

Here's the story. In February 2026, Bitdeer just cleared all their Bitcoin. Zero. The same week, they also completed a new debt issuance of $325 million. Total debt now reaches $1.3 billion. All that money is flowing into one goal: building global AI infrastructure.

What’s interesting is how Wu Jihan has shifted the fundamental business strategy. The game used to be simple—take electricity today, exchange for Bitcoin tomorrow, profit from the price difference. A 12-year-old arbitrage logic. Now he’s completely changing the game. It’s no longer about Bitcoin prices, but about long-term prices for computing power in the AI era.

Let’s look at the scale he’s building. Bitdeer’s total global power channel capacity is now 3,002 MW, with 1,658 MW operational and 1,344 MW under construction. For context, giant data centers like Microsoft and Google usually only have 100-300 MW. So Bitdeer is equivalent to packing the electricity needs of 10-30 Google-class data centers into one company. This number sounds incredible, but this infrastructure is actively being built.

The main facilities are in three locations. Rockdale, Texas with 563 MW is already operational, focused on mining, with stable cash flow. Clarington, Ohio with 570 MW is the core of the AI transition plan, scheduled to finish in Q2 2027—but here’s a big risk, there’s a lawsuit from local steel producers claiming construction will disrupt shared infrastructure. If delayed, the entire schedule could be thrown off. Then Tydal, Norway with 175 MW, scheduled to finish by the end of 2026, uses hydro power with competitive energy costs. This is currently the most advanced option with the lowest risk.

But wait, there’s one thing often overlooked— their own chips. Bitdeer is developing the third generation SEALMINER. SEAL03 has achieved 9.7 joules per terahash efficiency, and SEAL04 targets 5 joules per terahash. The gross margin from these chips exceeds 40%, far higher than typical mining. It’s similar to what Wu Jihan did at Bitmain—moving from buying third-party equipment to making their own.

Now about this large debt. The convertible bond structure is designed with staggered maturities: 2029, 2031, 2032. The idea is that each year, as bonds mature, new milestones are achieved. Tydal should be operational by the time the first batch matures. Clarington should be generating revenue by the second batch. The market will judge itself in 2032. Three points in time, three opportunities to renegotiate. But Wall Street isn’t too convinced—analysts lowered the target price from $26.50 to $14, while the current stock price is only $8.

The pressure is real. An annual interest burden averaging 5% on $1.3 billion means $650M per year. Meanwhile, AI/HPC cloud revenue throughout 2025 didn’t even reach one-tenth of six-month interest payments. Now, everything depends on ongoing debt issuance to keep the cycle going.

Their AI business is only generating $10 million a year, less than 2% of total revenue. Their GPU count increased from 584 to 1,792 in three months, but utilization rates dropped from 87% to 41%. Machines are being installed too quickly; B200 and GB200 are still in testing phases. Electricity is in place, machines are being installed, but revenue has yet to catch up.

Estimated revenue if everything goes perfectly? Roth and MKM project $850 million annually. Management is more aggressive, claiming it could reach $2 billion. But all of this depends on three unmet conditions: timely construction, long-term contracts with major hyperscalers, and GPUs running at full capacity.

The biggest risk now isn’t debt, not stock prices, but the lawsuit from Clarington. If construction is delayed by two years, the entire plan could fall apart. Bitcoin network difficulty also surged 14.7% in February, gross margins dropped from 7.4% to 4.7%. Mining operations are gradually thinning out.

But there’s another perspective here. Wu Jihan isn’t really guessing which path will win in the AI era. He’s bought a position of “whoever wins, has to pay my electricity.” Like Amazon not guessing which internet company will succeed, just renting servers to everyone. Or AT&T, which doesn’t care what you’re talking about, as long as you’re making calls.

From selling products to selling services to collecting rent—industry evolution has always followed one path. The only difference is whether you actively move closer or are pushed by others. Wu Jihan bought this window of opportunity with billions of dollars. He’s waiting for AI money to catch up with his debt acceleration. Two to three years will show whether this strategy works or everything collapses.