Generally in line with market expectations, while Cango’s transition story became clearer. Cango reported 4Q25 revenue of US$179.5M, above our US$163M estimate, while net loss from continuing operations came in at US$285.0M, below our expectation due mainly to non-recurring transformation costs, mining-machine impairment, and fair-value adjustments, rather than a deterioration in the core operating franchise. Bitcoin mining remained the core driver, contributing US$172.4M of 4Q25 revenue and US$675.5M for CY2025, with BTC production of 1,718.3 in 4Q25 and 6,594.6 for the full year. The company also completed the termination of its ADR program and transitioned to a direct NYSE listing, which in our view should gradually improve investor accessibility, lower structural holding friction, and align Cango more closely with U.S.-listed peers. On the operating side, rising network difficulty and a tougher post-halving environment continued to pressure mining economics, but Cango still delivered a comparatively resilient production profile through 2025. Over the longer term, we remain constructive on bitcoin prices, and expect Cango’s fleet optimization and lower-power-cost migration to support margin recovery, with gross profitability gradually improving into FY27.
Mining still anchors the story, but cost and balance-sheet discipline now matter more. That, in our view, also explains why the stock has recently traded less like a pure BTC beta name than some investors expected. In 4Q25, cash mining cost rose to US$84,552 per BTC and all-in cost reached US$106,251, underscoring that post-halving mining has become far less forgiving for higher-cost operators. Management’s response appears rational: phasing out older, high-energy-consuming machines, shifting capacity toward lower-power-cost regions, and accepting some short-term hashrate contraction in exchange for better fleet efficiency and stronger downside resilience. Just as importantly, the February 2026 sale of 4,451 BTC and subsequent debt repayment should be viewed less as a loss of conviction on bitcoin and more as a deliberate move to reduce leverage, recover liquidity, and create room for the next stage of execution. This is also consistent with the company’s latest messaging: rather than passively absorbing cycle volatility, Cango is proactively deleveraging and repositioning the balance sheet for a more capital-disciplined phase of growth.
AI+HPC remains early-stage, but execution is no longer just conceptual. Management made clear that EcoHash is not trying to compete head-on with hyperscale training data centers; instead, Cango is targeting AI inference and generative-AI workloads, where distributed demand, lower latency, and modular deployment can matter more. That positioning is important because it fits Cango’s existing asset base better: the company can reuse energy-linked infrastructure from its mining network and explore a lower-capex path into AI compute. Progress remains early and still in the pilot stage, but the execution framework is becoming clearer: EcoHash has been established in Texas, first-phase retrofit work has started at the Georgia site, and management has begun laying out a phased rollout plan for modular AI infrastructure. In our view, the right milestones to monitor are not aggressive near-term revenue assumptions, but proof of deployment, proof of economics, and proof of first revenue.
Cango’s current valuation still appears to underprice its asset base and transition optionality. As of December 31, 2025, Cango reported US$41.2M in cash and cash equivalents, US$663.0M in receivables for bitcoin collateral, US$248.7M in net mining machines, and US$557.6M in long-term related-party debt. Since then, it has monetized part of its BTC holdings to repay debt and secured new equity funding, including a completed US$10.5M investment and agreements for an additional US$65M. Against a current market capitalization of roughly US$249M, we think the stock is still pricing in execution skepticism far more than asset coverage, balance-sheet repair, or AI transition value. Put differently, the market is still valuing Cango closer to a stressed transition-risk miner than to a deleveraging, asset-backed compute platform with emerging AI optionality. The next rerating catalysts are relatively clear: lower mining costs through fleet optimization, continued deleveraging and treasury discipline, and tangible AI milestones that can move the story from concept to commercial validation. Cango’s shares are currently trading at US$ 0.68 per share, with a market capitalization of US$244M. We forecast revenues of US$236M for CY2026, respectively. This equates to price-to-sales (P/S) multiples of 1.0x, significantly below the peer group averages of 5.8x.
Comments