Key Highlights
- Full-year 2025 brought a staggering $452.8 million net loss for Cango (CANG), despite generating $688.1 million in total revenue
- Fourth quarter losses reached $285 million, primarily due to an $81.4M impairment write-down on mining equipment and a $171.4M fair-value decline on Bitcoin-backed receivables
- In February 2026, the firm liquidated approximately $305 million in Bitcoin holdings to reduce outstanding debt
- The company is abandoning Bitcoin mining operations in favor of AI infrastructure development, operating under the new brand EcoHash
- Shares have plummeted over 84% across the last six months, currently hovering around $0.68
Cango (CANG) endured a devastating inaugural year in the Bitcoin mining sector. The firm disclosed a full-year 2025 net loss totaling $452.8 million, despite recording $688.1 million in overall revenue — with $675.5 million generated from mining operations alone. Operational expenses dramatically exceeded income.
The fourth quarter of 2025 mirrored this troubling pattern. While quarterly revenue reached $179.5 million, total operating costs and expenses surged to $456.0 million. This imbalance resulted in a quarterly net loss of $285 million.
The primary culprits were non-cash accounting adjustments. An $81.4 million write-down on mining equipment combined with a $171.4 million loss stemming from fair-value fluctuations in Bitcoin-backed receivables accounted for the majority of losses. Additionally, comprehensive mining costs escalated to $106,251 per BTC during Q4.
CFO Michael Zhang attributed the losses primarily to one-time transformation expenses and market-driven fair-value recalculations.
Throughout 2025, Cango successfully mined 6,594.6 Bitcoin — averaging approximately 18.07 BTC daily. However, total operating expenses reached $1.1 billion, which included $338.3 million in impairment charges on mining hardware.
Transition to Artificial Intelligence Infrastructure
The company has been strategically repositioning its business model. In April 2025, Cango divested its traditional China-based automotive financing operations for $352 million to Ursalpha Digital Limited, a Bitmain-affiliated entity. This transaction included a transfer of 32 exahashes per second in mining capacity, effectively transforming Cango into a pure-play Bitcoin miner.
Another major shift followed. During February 2026, Cango secured $75.5 million through equity financing and liquidated 4,451 BTC for approximately $305 million to reduce debt obligations. CEO Paul Yu announced the company is “progressing our transformation into an AI infrastructure provider.”
This strategic realignment includes a corporate rebrand to EcoHash. The strategy involves repurposing existing computational and energy infrastructure for AI inference applications.
Cango is walking a path already traveled by industry peers. Following Bitcoin’s April 2024 halving event that reduced block rewards by 50%, mining operations globally began reassessing their energy-intensive infrastructure. Artificial intelligence workloads presented an alternative revenue opportunity.
Bitfarms, Hut 8, Riot Platforms, and Core Scientific have all pursued similar strategies. Core Scientific’s acquisition by CoreWeave in a $9 billion transaction last year provided compelling evidence that AI companies value miners’ existing energy agreements as strategic assets.
Dramatic Share Price Collapse
Broader market conditions have compounded the challenges. Bitcoin dropped beneath $90,000 in November 2025, declining nearly 30% from its October high above $126,000. By March 2026, the cryptocurrency was trading near $73,700.
CANG shares have mirrored this downward trajectory. The stock declined from approximately $4.50 on October 1, 2025 to roughly $1.50 by year-end. Currently trading at $0.68 — representing a collapse exceeding 84% over six months.
The company’s 6,594.6 Bitcoin mined throughout 2025 came at an all-in cost of $106,251 per BTC in Q4, a threshold that provided minimal profit margins even before accounting for substantial impairment charges.





