Wu Jihan's $1.3 Billion Bet: BitDeer's Transformation From Mining to AI Infrastructure

On February 20, 2026, BitDeer released a stark weekly update: 189.8 BTC mined and sold. Remaining inventory of 943.1 BTC liquidated. Bitcoin balance: zero. For a company built on mining cryptocurrency for over a decade, this move signaled something more profound than a quarterly adjustment—it announced a complete strategic pivot. Wu Jihan, BitDeer’s founder, had placed his entire company on a wager that the future of computing power would generate more revenue than digital currencies themselves.

The decision didn’t materialize overnight. Bitcoin currently trades at $67.55K, but BitDeer’s focus has shifted entirely away from price speculation. Wu Jihan had financed this transformation through convertible bonds and equity offerings totaling $1.3 billion by early 2026—a mountain of debt that now shadows every operational decision the company makes. The question haunting Wall Street analysts and industry watchers is whether this bet makes sense, or whether Wu Jihan has walked his company into a trap with no exit strategy.

From Time Arbitrage to Infrastructure Control: The Logic Behind Wu Jihan’s Pivot

Bitcoin mining, in its essence, has always been a form of time arbitrage. Purchase electricity and machinery today; hope to exchange them for bitcoins tomorrow worth more. For over a decade, this calculus worked. Wu Jihan had perfected the model—acquiring land, securing cheap power, deploying hardware—and watching time compound the investment.

Now, Wu Jihan is executing a similar playbook with one crucial difference: the target has changed. Instead of betting on cryptocurrency prices, BitDeer now bets on computing power demand under the artificial intelligence boom. The mechanism has shifted from exchanging electricity for tokens to borrowing capital to acquire land and infrastructure. The arbitrage structure remains; only the object has transformed.

This shift reveals Wu Jihan’s deeper strategic insight: infrastructure providers—not product makers, not platform operators—collect rent regardless of which technology prevails. Amazon didn’t bet on specific internet companies; it rented servers to all of them. AT&T doesn’t care what conversations traverse its phone lines; it profits from every call. Wu Jihan appears to be following this playbook: control the entrance to computing power, and collect fees regardless of whether AI or crypto dominates the computational economy.

The Scale of Wu Jihan’s Ambition: 3,000 MW of Power Infrastructure

As of early 2026, BitDeer’s global power pipeline reached 3,002 MW—equivalent to the combined electricity requirements of 10 to 30 hyperscale data centers like Google’s. This infrastructure empire consists of three main pillars:

Rockdale, Texas: Already operational with 563 MW capacity (including 179 MW expansion), primarily dedicated to bitcoin mining operations that generate stable cash flow.

Clarington, Ohio: The centerpiece of Wu Jihan’s AI transformation—a 570 MW facility with a 30-year power contract, originally scheduled for Q2 2027 completion. Designed specifically for HPC and AI workloads, this project represents the core of the entire transformation strategy. It is also the company’s largest vulnerability.

Tydal, Norway: A 175 MW mining facility under conversion to an AI data center, expected operational by end of 2026. The hydroelectric power source provides competitive energy costs, making this the fastest-progressing and least risky asset. Projected to deliver 164 MW of effective IT load.

Beyond land and power, Wu Jihan has invested in proprietary chip development through BitDeer’s SEALMINER division. The SEAL series mining chips have reached their third generation, with the SEAL03 achieving 9.7 joules per terahash—a leading efficiency metric. The SEAL04 targets 5 joules per terahash; if achieved, it would surpass all mass-produced mining chips on the market. Gross margins on self-developed chips exceed 40%, far higher than mining operations themselves—a lesson Wu Jihan learned from his earlier venture, Bitmain.

The Debt Burden: How Wu Jihan Is Financing the Impossible

To fuel this expansion, Wu Jihan orchestrated an aggressive financing campaign. By December 31, 2025, BitDeer carried over $1 billion in total debt. A February 2026 debt issuance of $325 million pushed liabilities to approximately $1.3 billion.

The financing structure reveals Wu Jihan’s understanding of capital markets. Three tranches of convertible bonds mature in 2029, 2031, and 2032—deliberately staggered to create a buffer zone rather than a cliff event:

• May 2024: Tether invested $100 million; combined with warrant options, potential total exposure reaches $150 million.
• August 2024: $150 million convertible bonds at 8.5% annual interest.
• November 2024: $360 million convertible bonds at reduced 5.25% interest rate.
• November 2025: $400 million convertible bonds plus 148.4 million share issuance.
• February 2026: $325 million convertible bonds plus 43.5 million shares, with $135 million used to refinance earlier high-interest bonds, pushing maturity from 2029 to 2032.

Convertible bonds allow Wu Jihan to outsource repayment to market forces. The 2032 bonds carry a conversion price of approximately $9.93—25% premium to the concurrent equity offering of $7.94. If BitDeer’s stock reaches that price, bondholders convert to equity automatically. The company requires no cash repayment; it only needs the stock price to appreciate.

The math is brutal. At an average 5% interest rate on $1.3 billion, annual interest burden exceeds $65 million. Yet in 2025, BitDeer’s entire AI and HPC revenue was less than six months of interest payments. Currently, all interest is being rolled forward through additional debt issuance—a strategy that works only as long as capital markets remain willing to fund Wu Jihan’s vision.

The market reaction to each new bond offering has been consistent: BitDeer’s stock price drops 10-17% per announcement, a punishing cycle for existing shareholders. Despite this, Wu Jihan has successfully raised capital each time—a testament both to investor belief in his thesis and to the appetite for AI infrastructure plays.

The Revenue Gap: Promises Versus Performance

BitDeer’s AI operations currently generate $10 million annually—less than 2% of total company revenue. For a company with ~$2 billion market value, this figure is negligible against the scale of infrastructure being built.

Yet the deployment curve is accelerating. BitDeer’s GPU inventory surged from 584 units to 1,792 in just three months (early 2026), but utilization plummeted from 87% to 41%. The denomination of capacity is expanding rapidly; revenue generation lags far behind. The B200 and GB200 GPU series are still in customer testing phases and haven’t generated material revenue.

Revenue projections vary wildly. Roth and MKM estimate that with full deployment of HPC capacity, annualized revenue potential reaches $850 million. Wu Jihan’s own management is more aggressive: if all 200 MW deployed to AI cloud operates at full capacity, revenue could exceed $2 billion annually—three times BitDeer’s entire 2025 mining revenue.

These projections rest on three unmet conditions: construction completes on schedule; BitDeer secures hyperscaler-level long-term contracts; and GPUs run at rated capacity. None have materialized yet. This is Wu Jihan’s core challenge: the timeline.

The Clarington Lawsuit: The Single Point of Failure

BitDeer’s most significant vulnerability is not its debt load or stock price volatility—it’s a steel manufacturer called American Heavy Plate Solutions.

In the same Ohio industrial park where Clarington would operate, American Heavy Plate leased 9.9 acres in 2018 under a 30-year contract. When BitDeer announced its 570 MW AI data center project, the steel company sued, claiming that BitDeer’s facility would interfere with shared power infrastructure, roads, railroads, and communication systems, violating restrictive covenants in their lease. They seek a permanent injunction to halt BitDeer’s construction entirely.

Clarington represents 42% of BitDeer’s entire pipeline under construction. A prolonged legal battle could delay opening by years, cascading through Wu Jihan’s entire timeline. This isn’t a financial risk; it’s an operational risk that no amount of capital can guarantee.

The mining sector isn’t dormant during this uncertainty. In February 2026, Bitcoin network difficulty surged 14.7%—the largest single-month jump since May 2021. With identical electricity costs, miners extract fewer coins. BitDeer’s Q4 gross margin on mining operations already contracted from 7.4% to 4.7% year-over-year. The legacy business is slowly eroding.

The Two Futures: Triumph or Unraveling

Wu Jihan’s strategy assumes a favorable timeline:

Scenario A (Success): By end of 2026, Tydal’s 164 MW AI data center completes operation; European contracts begin flowing. In 2027, Clarington lawsuit resolves favorably and construction begins; major U.S. hyperscalers sign long-term agreements. By 2028-2029, both core assets reach full utilization; revenue approaches $1 billion. Analysts reclassify BitDeer from “discount mining company” to “premium AI infrastructure provider.” In 2029, first bond maturity occurs; holders observe rising stock prices and convert to equity rather than demand cash repayment.

Scenario B (Unraveling): Clarington litigation extends two years; construction remains frozen. Tydal hits delays; GPU utilization stalls at 41%. The 2029 bond maturity arrives with insufficient cash reserves, forcing refinancing at unfavorable rates. Stock dilution accelerates; conversion thresholds become unreachable. The company faces a liquidity crisis.

Both pathways are mathematically possible. The outcome depends entirely on execution speed—a dimension beyond Wu Jihan’s complete control.

Why Wu Jihan Liquidated All Bitcoin Holdings

Tradition in the mining community treats bitcoin accumulation as an article of faith—a vote of confidence in cryptocurrency’s long-term value. MARA holds 53,250 BTC; Riot holds 18,000; Marathon Strategy holds 710,000. BitDeer now holds zero.

The official rationale: selling bitcoin generates liquidity for land acquisition. That’s technically accurate, but it misses the deeper shift. Competitors are following similar paths: Riot sold $200 million in bitcoin to fund AI expansion; Bitfarms abandoned its “bitcoin company” identity; MARA is also investing in HPC infrastructure.

But Wu Jihan’s move reflects something more fundamental. The mining industry has always been structured around a single bet: something in the future will be more valuable than today. Ten years ago, the bet was on cryptocurrency price appreciation. Now, the bet is on computing power demand explosion. The target changes; the underlying logic—time arbitrage—persists.

Wu Jihan has traded one form of speculation for another. His real prize is simpler: a position where “no matter who wins the AI war, they’ll have to rent my infrastructure.” He’s not betting on the outcome; he’s positioned himself to profit from the infrastructure race itself, regardless of winner.

The Verdict: Can Wu Jihan Outrun His Debt?

Wu Jihan purchased a narrow window of opportunity for $1.3 billion. BitDeer’s survival depends on whether AI infrastructure demand materializes quickly enough to generate sufficient cash flow before bond conversion thresholds become unreachable.

Keefe Bruyette slashed BitDeer’s price target from $26.50 to $14. Current share price hovers around $8. Wall Street’s message is direct: transformation narratives require revenue generation. But pressure has also provided Wu Jihan something equally necessary and cruel: time.

The next two years will clarify whether Wu Jihan’s logic was prescient or catastrophic. Whether Clarington’s lawsuit resolves favorably. Whether Norwegian hydropower can anchor European AI demand. Whether Tydal completes on schedule. Whether hyperscalers commit long-term capacity at the prices BitDeer requires.

Wu Jihan orchestrated this setup to create optionality—staggered debt maturities, convertible bond structures, strategic land acquisitions—betting that the pace of AI infrastructure demand would exceed his capital cost curve. If correct, he transforms BitDeer into an essential infrastructure monopolist and rewards early believers handsomely.

If he’s wrong, the $1.3 billion in liabilities becomes an anchor dragging the company toward restructuring, shareholder dilution, and potential dissolution.

Only the next three years will reveal which path BitDeer actually walks.