Cango’s First Year Mining Bitcoin Ends With $452.8M Loss as It Sells BTC to Fund AI Pivot
Cango logged a $452.8M net loss in its first full year as a bitcoin miner, selling BTC to repay debt and bankroll a shift toward AI. Here’s what that signals for miner strategy.

Because Bitcoin
March 17, 2026
Cango closed its first full year as a bitcoin miner with a $452.8 million net loss, and it liquidated bitcoin holdings to service debt and finance a move into artificial intelligence. That single line tells you nearly everything about where miner strategy is heading: balance sheets, not belief systems, decide survival.
The core issue isn’t just the size of the loss; it’s the decision to sell BTC to keep the lights on and fund a pivot. Miners often talk about “HODL as a strategy,” but liabilities mature on calendar time, not chain time. When debt service meets volatile revenue, treasuries become shock absorbers. If capital markets aren’t open or are prohibitively expensive, selling bitcoin becomes the release valve—whether management likes it or not.
From a capital allocation lens, this is the classic trade-off: optionality vs. solvency. Holding BTC maximizes upside convexity, but only if cash burn is controlled and debt is laddered sensibly. Selling BTC reduces future upside but can extend runway and de-risk covenant pressure. Funding an AI transition by liquidating BTC is a bet that steadier, contract-driven compute revenues can outpace the forgone bitcoin beta. That is not a moral judgment; it’s a duration and volatility judgment.
Technologically, the pivot is non-trivial. ASIC mining and AI inference/training share a hunger for power and cooling, but they diverge in workload profiles, hardware cycles, and customer expectations. Miners stepping into AI need to solve for GPU supply, data center retrofits (high-density racks, liquid cooling in some cases), low-latency networking, and service-level commitments that look far more enterprise than commodity hash. Done right, existing power contracts and sites become an advantage; done loosely, retrofits turn into CapEx traps with utilization risk.
Investor psychology around BTC treasury sales is predictable: some interpret it as capitulation, others as discipline. Context matters. Selling into weakness to cover near-term liabilities can look reactive; selling into strength to derisk a pivot looks deliberate. Markets often reward clarity—codified treasury policies, hedging frameworks, and a transparent glidepath for the AI business tend to compress the “strategy uncertainty” discount on the equity.
The business calculus is straightforward: bitcoin mining offers programmatic issuance and price exposure with difficulty risk; AI infrastructure offers service revenue with utilization and customer concentration risk. Miners moving toward AI are choosing potentially smoother cash flows (if contracted) over commodity exposure. The edge comes from execution speed: securing power, GPUs, and customers on timelines that don’t outpace liquidity. Every month of delay compounds cash burn, especially after realizing a nine-figure loss.
There’s also a stewardship dimension. Funding a pivot by selling BTC asks shareholders to swap one risk regime for another. That can be sensible if management communicates the conversion math: expected IRRs on AI deployments, payback periods, utilization assumptions, and downside protection via prepayments or take-or-pay contracts. Without that, the pivot reads as narrative defense rather than strategy.
What to watch next: - Treasury discipline: explicit BTC holding/selling policy and hedge usage - Liquidity runway: cash on hand, debt maturities, and any covenant headroom - AI build economics: cost per MW deployed, rack density, GPU acquisition, and targeted utilization - Revenue visibility: customer contracts, duration, and pricing mechanics - Execution cadence: time from CapEx commit to live capacity and revenue recognition
A $452.8 million net loss paired with BTC sales to repay debt and fund an AI shift suggests urgency and a willingness to sacrifice upside for survival. That can work if the pivot converts power and real estate into contracted compute quickly and credibly. It falters if capital keeps chasing rebrands instead of unit economics. Miners don’t need perfect timing; they need consistent liquidity and a declared strategy that their balance sheet can actually support.
